;^^^''^ ^^h- tr > 'i^■^m ,v ^"v* j^^^fS* :;^'r> •c^-. THE UNIVERSITY OF ILLINOIS LIBRARY 5^0.5 AC V.I6 AT URPANA-CHAMPAiGN BCOKSTACKS "SSS t,-y^ I AMERICAN '^r Chemical Journal IRA REMSEN Propbssok of Chemistry in the Johns Hopkins Unh BRSITY. Vol. XVL 1894. BALTIMORE : THE EDITOR. Thb Fribdenvvald Co., Pki » '^ 1 M -J "rRSr. CONTENTS VOL. XVi. No. I. On the Combination of Sulphuric Acid with Water in the Presence of Acetic Acid. By Harry C. Jones, . The Atomic Weight of Palladium. By Edward H. Keiser and Mary B. Breed, Contributions from the Chemical Labokatory of Harvard College: LXXXI. — On the Action of Water upon Tribromtrinitro- benzol and Tribromdinitrohenzol. By C. I.oring Jackson and W. H. Warren, .... LXXXn. — Trianilidodinitrobenzol and certain related Corn- founds. By C. Loring Jackson and H. N. Herman, ....... The Polymeric Modifications of Acetic Aldehyde, Paralde- hyde AND Metaldehyde. By W. R. Orndorff and John White, Contribution from the Kent Chemical Laboratory of the University of Chicago : The Action of Phosphorus Pentachloride on Ureihanes, By Felix Lengfeld and Julius vStieglitz 35 43 No. 2. Upon the Separation of Thorium from the Rare Earths of THE Cerium and Yttrium Groups by Means of Potassium Hydronitride. By L. M, Dennis and F. L. Kortright, . 79 On the Composition of Certain Petroleum Oils and of Re- fining-Residues : n. — The Sulphur Compounds in Ohio Petroleum. By Chas. F. Mabery and Albert W. Smith, .... 83 HI — Preliminary Examination of tf-e Canadian Sulphur- Petroleum. By Charles F. Mabery 89 1)1*^ C.0 Contents. Solubility of Metallic Oxides in Normal Potassium Salts of Tartaric and other Organic Acids. By Louis Kahlen berg and H. W. Hillyer Contribution krom the Laboratory of Wabash College : On two Stcreo-Isomeric Hydrazones of Benzoin. By Alex Smith and J. H. Kansoni, On the Action of Benzenesulphonic Acid upon PoifAssiuM Iodide, A New Class of Organic Periodides. By J. H Kastle and Herbert H. Hill Contribution from the Kent Chemical Laboratory of the University of Chicago : On Diamidoorthophosphoric and Diamidolrihydroxylphos- phoric Acids. By H. N. Stokes, Note on Monamidophosphoric Acid. By H. N. Stokes, 94 1 08 123 1 54 REVIEWS ANM) REPORTS. Principles and Practice of Agricultural Analysis, . . 156 Introduction A LA MECANiQUE cHiMiQUE (J. E. T.), . . . 157 Ostwald's Klassiker der exakten Wissknschaften (J. E. T.), 157 Jahrbuch der Chemie (I. R.) 158 A Guide TO Stereochemistry, 159 Address of Emil Fischer before the Berliner Akademie der Wissenschaften, 159 No. 3. Researches upon the Phenomena of Oxidation and Chemical Properties of Gases. By Francis C. Phillips, Contributions from the Chemical Laboratory of Harvard College : LXXXIII. — On Certain Substituted Crotonolactones and Mucobromic Acid. By Henry B. Hill and Robert W. Cornelison, .... An Isothermal Curve of Solubility of Mercuric and Sodium Chlorides in Acetic Ether. By C. E. Linebarger, . The Benzoyl Halogen Amides. By C. E. Linebarger, . Pentosans in Plants. By G. de Chalmot, .... Note on Pentosans in Soils. By G. de Chalmot, . Contribution from the Chemical Laboratory of Lehigh Uni VERSITY : On Phospho-Hydrocyanic Acid. By W. B. Shober and F W. Spanutius, . . ...... 163 214 216 218. 229 229 Contents. v REVIEW. Tabellarische Uebersicht DER Naphtalinderivate(I. R.), . 233 NOTE. On a New Class of Organic Bases, containing Iodine but no Nitrogen (I.R.) 233 No. 4. On the Decomposition of Diazo-Compounds: VIII. — A Study of the Action of the Salts of Diazo-Benzene on Methyl and Ethyl Alcohols under diff'erent Condi- tions. By J. L. Beeson, 235 Researches upon the Phenomena of Oxidation and Chemical Properties of Gases. By Francis C. Phillips, . . . 255 Contributions from the Chemical Laboratory of Harvard College: LXXXIII. — On Certain Substituted Crotonolactones and Mucobromic Acid. By Henry B. Hill and Robert W. Cornelison 277 Contributions from the Chemical Laboratory of the Rose Polytechnic Institute : lY..— Camphoric Acid. By W. A. Noyes 307 REVIEW. Lecture-Notes on Theoretical Chemistry (I. R.), . . . 312 No. 5. Contribution from the Kent Chemical Laboratory of the University of Chicago : The Affinity- Constants of Weak Acids and the Hydrolysis of Salts. By R. W. Wood, 313 The Color of Salts in Solution. By J. H. Kastle, . . . 326 Researches upon the Phenomena of Oxidation and Chemical Properties of Gases. By Francis C. Phillips, .- . . 340 Contribution from the Chemical Laboratory of Purdue University. XI. — A Reduction- Product of Orthosulpkobenzoic Chloride. By Walter Jones, 366 vi Contents. Contribution from the Kent Chemical Laboratory of the University of Chicago : On Nitrogen Halogen Compounds. By Felix Lengfeld and Julius Stieglitz, 370 On the Addition- Prodticts of the Aromatic Isocyanides. By Warren R. Smith 372 REVIEW. A Detailed Course of Qualitative Chemical Analysis, with Explanatory Notes (Thomas M. Drown) 393 No. 6. The Menthol Group. By Leo C. Urban and Edward Kramers, . 395 Ketones from Pinene Derivatives. By Leo C. Urban and Edw. Kremers, 404 Researches upon the Phenomena of Oxidation and Chemical Properties of Gases. By Francis C. Phillips, . . . 406 The Condensation-products of Aromatic Hydrazides of Acetacetic Ether. Indol and Pyrazol Derivatives. By C. Walker, 430 A Systematic Study of the Action of definitely related Chemical Compounds upon Animals. By Wolcott Gibbs and Edward T. Reichert, 443 Contributions from the Chemical Laboratory of the Rose Polytechnic Institute : X. — The Nitrites of some Amines. By W. A. Noyes and H. H. Ballard, 449 A Study of the Constituents of the Nodes and Internodes of the Sugar Cane. By J. L. Beeson, 457 Notes of Work from the Chemical Laboratory of the Uni- versity OF Virginia : 160. —On the Solubility of Cream of Tartar in Alcohol of various Strengths and of various Temperatures. By J. A. Roelofsen 464 161. — The Iodine-Absorption of some of the rarer Fatty Oils. By J. A. Roelofsen, ...... 467 REVIEWS AND REPORTS. Recent Progress in Physical Chemistry (J. E. Trevor), . . 470 Handbuch DER Stereochemie (Arnold Eiloart) 475 A System of Instruction in Qualitative Chemical Analysts (E-R-). 476 Centenary Commemoration of Antoine-Laurent Lavoisier, 1794 — May 8— 1894 47^ Contents. vii No. 7. Instruments for the Graduation and Calibration of Volu- metric Apparatus. By H. N. Morse and T. L. Blalock, . 479 Contribution from the Chemical Laboratory of the Costa RiCAN Government : Note on the Influence of certain Metals on the Stability of the Amalgam of Ammonium. By Gustave Michaud, . . 488 Mixed Double Halides of Antimony and Potassium. By Charles H. Herty, 490 Contributions from the Chemical Laboratory of the Rose Polytechnic Institute : XI. — Camphoric Acid. By W. A. Noyes 500 The Electrolytic Reduction of Paranitkobenzoic Acid in Sulphuric-Acid Solution. By A. A. Noyes and A. A. Clement, 511 Investigations on the Sulphon-PhthaleIns : III — Phthale'ins of Ortho-sulpho-para-toluic Acid. By James A. Lyman, ......... 513 IV. — Orcin-sulpho7i-phthalein. By J. E. Gilpin, . . .528 On Para-chlor-meta-sulpho-benzoic Acid and some of its Derivatives. By H. M. Ullmann, 530 * Contributions from the Chemical Laboratory of the Case School of Applied Science : IV. — On the Determination of Sulphur in Volatile Organic Compounds. By Charles F. Mabery, . . . 544 REVIEWS AND REPORTS. Recent Progress in Physical Chemistry (J. E. Trevor), . . 551 Die Lehre von der ElektrizitAt (J. E. Trevor), .... 563 Physikalisch-chemische Tabellen (J. E. Trevor), . . 565 Die wissenschaftlichen Grundlagen der analytischen Chemie (J. E. Trevor), 564 The Theory of Heat (J. E. Trevor) 565 Elektrochemie, ihre Geschichte und Lehre (J. E. Trevor), . 565 Josiah Parsons Cooke (Theodore William Richards), . . . 566 No. 8. On Electrosyntheses by the Direct Union of Anions of Weak Organic Acids. By J. B. Weems, .... 569 Pentosans in Plants. By G. de Chalmot, 589 On Chemical Equilibria as Temperature-Functions. By J, E. Trevor and F. L. Kortright, 611 viii Contents. A New Formula for Specific and Molecular Refraction. By W. F. Edwards 625 Contribution from the Chemical Laboratory of the Uni- versity OF Kansas : On Paraisobutylsalicyl Aldehyde and some of its Derivatives. By F. B, Dains and I. K. Rothrock, 634 Contributions from the Chemical Laboratory of Cornell University : The Polymeric Modifications of Propionic Aldehyde: Parapro- pionic and Metapropionic Aldehydes. By W. R. OrndorfE and Miss L. L. Balcom, 645 REVIEWS AND REPORTS. Select Methods in Chemical Analysis (H. N. M.), . . . 650 Quantitative Chemical Analysts and Electrolysis (H. N. M.), 652 Lessons in Qualitative and Volumetric Chemical Analysis, FOR the Use of Physicians, Pharmacists and Students (E. R.) 652 A Brief Introduction to Qualitative Analysis, for Use in Instruction in Chemical Laboratories (E. R.), . . 653 A Manual of Microchemical Analysis (E. R.), .... 653 Manual of Physico-Chemical Measurements (H. C. J.), . . 654 Laboratory Manual and Principles of Chemistry, for Begin- ners (W. W. R.) 654 An Elementary Manual of Chemistry (W. W. R.), . . . 655 EiNFOHRUNG in das STUDIUM DER QUALITATIVEN CHEMISCHEN Analyse (E. R.) 656 Inorganic Chemistry for Beginners (W. W. R.), . . , 656 Index TO Volume XVI, 657 Vol. XVI. [January, 1894.] "" No. AMERICAN CHEMICAL JOURNAL, ON THE COMBINATION OF SULPHURIC ACID WITH WATER IN THE PRESENCE OF ACETIC ACID. By Harry C. Jones. The strong attraction of sulphuric acid for water is utilized daily for the purpose of removing water from other substances. Not only is the qualitative fact established, but certain definite com- pounds of water with the acid have been isolated. The ordinary acid is but the first hydrate, containing one molecule of water to one of sulphur trioxide, OSS.H2O. A second hydrate containing two molecules of water is also known, OS(OH)4 = 03S.2H20. Other hydrates containing a much larger number of molecules of water have been thought to be probable from a study of cer- tain physical properties of the solutions of sulphuric acid in water. Quantitative measurements of the amounts of heat evolved when sulphuric acid is added to water in varying proportions, have been made by Hess, Favre and Silbermann and others, but probably the most accurate work has been done by Julius Thomsen.' He has found that the amount of heat liberated when the two are brought together in the proportion of one molecule of sulphuric acid to one molecule of water, is about one- third of the total heat set free when almost an unlimited amount of water (1599 molecules) is added to the sulphuric acid ; further, that about one-half of the total heat is liberated when the two are mixed in the proportion of one molecule of acid to two of water. 1 Ber. d. chem. Ges. 3, 498 ; and Thermochem. Untersuch., Bd. 3, 44. Vol. XVI.-i. 2 Jones. These results show at least that the amount of energy which appears as heat on the addition of a molecule of water is much greater for the first than for the second molecule, and this amount decreases rapidly with increase in the quantity of water. Mendeleeff' has studied Thomsen's data for the heat of solution, and Marignac's for the specific heat of the solutions, and has pointed out that the maximum amount of heat for say lOO volumes of the solution, and the maximum increase in temperature when water and sulphuric acid are mixed, correspond very closely to the proportions which would form the trihydrate H2S04.2H20 = S(0H)6. He also states^ that the maximum contraction nearly corresponds to these same proportions. He has also studied' the specific-gravity determinations of solu- tions of sulphuric acid in water, measurements of which have been made by Lunge, Marignac, Mendel6eff, Ostwald, Schertel, Wink- ler and others. From these it appears that the changes in the dif- ferential quotient -4i indicate the hydrates SOs.H^O, SO3.2H2O dp and SO3.3H2O (/> = percentage composition of HaSO*). When the percentage composition of solutions of sulphuric acid and the value -^ are plotted as a curve, this is stated to indicate, in addi- dp tion to the above, the hydrates H2SO4.6H2O and H2SO4.150H2O. It can be said in advance that none of these higher hydrates has been isolated, and their existence as definite chemical com- pounds in solution is far from proven. While certain definite compounds have been isolated, but little is known of the way in which sulphuric acid combines with water when, e. g., an excess of water is present, or which of the different compounds exists in solution under such conditions. Again, when just one or two equivalents of water are present, is all the water in combination with all the sulphuric acid in these proportions, or does a part of each remain free ? It would be of special interest if positive evidence could be furnished that the supposed hydrates, containing a hundred or more molecules of water, were definite chemical compounds existing as such in the solutions. The question of the stability of the different hydrates in solution also arises, as to whether they are capable of existence independent of the solvent, or are dissociated or decomposed by it. 1 Ber. d. chem. Ges. 19, 400. ^Ibid. 19, 387. *[bid. 19, 379 ; Ztschr. phys. Chem. 1, 273; Grundlagen d. Chemie, 923-925. The Hydrates of Sulphuric Acid. 3 A satisfactory quantitative study of the combination of sulphuric acid with water seems not to have been made up to the present time, and the above points remain unsettled. To solve this problem, some method must be employed which would enable one to determine the total www^^r of molecules present under the varying quantitative conditions in which water and sulphuric acid are brought together. Such a method has been found in the low- ering of the freezing-point of the solvent by the dissolved sub- stance. But for this purpose some solvent must be employed on which neither water nor sulphuric acid would act chemically. The only reaction which could then take place would be between the water and the sulphuric acid, and this could be determined by the changes in the freezing-point of the solvent. Such a solvent is pure acetic acid. This has also the special advantage that its freezing-point is in a very convenient position to determine. The method of work consisted in preparing pure acetic acid and determining its freezing-point, then adding a known amount of water, and determining the lowering produced by this water. A weighed amount of sulphuric acid was then added to another quantity of the acetic acid, and the lowering of the freezing-point produced by the sulphuric acid determined. Finally a known quantity of sulphuric acid was added to acetic acid to which a weighed amount of water had been added, and the lowering pro- duced by both together determined. The same amount of acetic acid was used in every case. We should then know the lowering produced by the water alone, by the sulphuric acid alone, and by both when brought together in the acetic acid, which are the necessary data for determining the state of combination of the sulphuric acid and water, as will be seen. Two series of determinations, in one of which a very small amount of water was used, and different amounts of sulphuric acid, and in the other, much water and again varying amounts of sul- phuric acid, might have been sufficient to show what compounds were formed, but it was also desired to test the stability of these compounds at very different concentrations in the acetic acid, and in the presence of varying amounts of their constituents in excess. Therefore several series of measurements were "necessary, using at first very little water, and gradually increasing this from one series to the next, as far as was possible with the thermometer employed. Different amounts of sulphuric acid were added, and 4 Jones. the freezing-points determined. Thus for any given series of measurements, the amount of water remained constant and the quantity of sulphuric acid was varied. Riidorff' has pointed out that when sulphuric acid is added to acetic acid containing water, the freezing-point is raised. Thus, when lo parts of water are added to loo parts of acetic acid the freezing-point is 4.3°. If 20 parts of sulphuric acid are then added, the freezing-point is 10.7°. Pickering^ has more recently utilized the same fact to determine, as he thought, whether dissociation takes place in solution. He has stated the problem clearly, and the method at first sight would seem capable of throwing light upon it. The results which he obtained show that the sulphuric acid and water combine to form a smaller number of parts, rather than dissociate into a larger number. From this Pickering argues against the dissociation theory. But when we consider that most of the solutions which he employed were so concentrated that their conductivity was almost zero, his argument against the theory, based on these results, is without force, since, to state a truism, the theory of electrolytic dissociation applies only to those solutions in which electrolytic dis- sociation takes place, and this as shown by the conductivity or some other method. He has also made a few determinations employing much less sulphuric acid in proportion to the water and, without determining the conductivity of his solutions, has con- cluded that the theory must at least apply to these. Had it even been shown that such solutions conduct electricity to some extent, the meagreness of his experimental data from the freezing-point method should have prohibited final conclusions being drawn from them. Method of Work. The ordinary Beckmann apparatus and thermometer divided into hundredths of a degree, were employed. It was not practi- cable to use the large thermometer divided into thousandths of a degree, because the lowerings produced were far too great. Since the acetic acid used as a solvent would take up moisture from the air, the arrangement which has been devised by Beckmann^ for protecting the solvent from this moisture was employed at first. 1 Ber. d. chem. Ges. 8,393. "^ ^bid. 24, 1579. ' Ztschr. phys. Chem. 7, 323. The Hydrates of Sulphuric Acid. 5 This was not, however, found to be entirely satisfactory, for when only a small amount of sulphuric acid was added to the bulb and a very gentle stream of air passed through the sulphuric acid for a long time, the acetic acid in the apparatus was found to contain a trace of sulphuric acid. The amount of sulphuric acid which was thus carried over mechanically was very small, yet was always found when the stream of air was allowed to flow for a consider- able time. The apparatus was, however, found to work perfectly satisfactorily when the air was first passed through a tube containing sulphuric acid and then allowed to enter the apparatus through the side Beckmann bulb which was kept dry. When this device was employed, no sulphuric acid was found in the acetic acid. The freezing-point of 25 cc. (26.335 grams) of acetic acid was determined by the thermometer, a weighed amount of water was then added and the freezing-point again determined. The differ- ence gave the lowering produced by the water alone. Then weighed amounts of sulphuric acid were added and the freezing- points ascertained. Duplicate determinations were made and the mean taken as the true value. Preparation of Pure Acetic Acid. In order that the lowering of the freezing-point produced by sulphuric acid alone could be determined, it was necessary to have acetic acid as free as possible from water. Some difficulty was experienced in preparing such a specimen of acetic acid. The glacial acetic acid of commerce was fractionally recrystallized thirty-five times. The product from the last crystallization froze at 16.5° C. and had a specific gravity of 1.0534. This acid, which was used in the following determinations, seemed to be entirely free from any appreciable amount of water: when a drop of sulphuric acid was added it produced a lowering of the freezing- point. During the last three or four crystallizations the freezing- point of the acetic acid did not change to an observable extent. Five or six liters of the commercial acid were used in the first crystallization. From this about 200 cc. of the pure acid were obtained after thirty-five crystallizations. In order to prepare a quantity of acid sufficient for this work the entire process was repeated twice. 6 Jones. Resrdts. The lowerings of the freezing-point of acetic acid produced by the different amounts of water were determined and plotted as Curve I, in which the ordinates represent concentrations and the abscissas lowerings of the freezing-point. The unit on the ordi- nate 1^0.04 gram, on the abscissa =0.2°. The lowering produced by a given amount of water, say a milligram, gradually decreased with increase in the total amount of water present. This is seen in the gradual bending of the curve frpm the abscissa, and also from the decrease in the molecular lowering (Column 3, Table A). This decrease is far greater than would be caused by the change in volume due to the addition of more and more water. The corresponding results from sulphuric acid have been plotted in Curve 2, in which the ordinate unit = 0.250 gram, and the abscissal unit = 0.2°. The molecular lowering, as calculated, decreases to a minimum and then increases, yet the total volume of the solution is ever increased by the addition of more and more sulphuric acid. From these curves the lowerings of the freezing-point of the acetic acid produced by the different amounts of sulphuric acid alone and water alone, can be read off at once. Each must then be corrected for the increase in volume due to the presence of the other. Table A gives the gram-molecular lowering for water, B that for sulphuric acid. Table A. Lowerings of the Freezing-point of Pure Acetic Acid by Small Quantities of Water. Wt. Water added to .335 grams CH3COOH. Lowering of the Freezing-point. Gr.-mol. Low. 18 grams H,0 to 1000 grams CH3COOH. 0.062 gram 0.480° 3.67 0.130 0.985 3-59 0.204 1.497 348 0.275 1.970 3-39 0.349 2.444 3-32 0.417 2.858 3-25 0-531 3-507 3-13 0.637 4.076 3-03 0.716 4-500 2.98 0.807 4-970 2.92 v^ :^>-^ dSh \t -^v^ t t-|M ...^ '1 The Hydrates of Sulphuric Acid. Table B. Lowerings of the Freezing-point of Pure Acetic Acid by Small Quantities of Szilphtiric Acid. Wt. Sulphuric Acid to 26.335 grams CH3COOH. Lowering of the Freezing-point. Gr.-mol. Low. 98 grams H2SO4 to 1000 grams CH3COOH. 0.146 gram 0.245° 4-33 0.474 0.630 3-43 0.893 I.OI5 2.93 I-57I I.52I 2.50 2.254 2.005 2.29 2.9 rg 2.505 2.21 3-599 3.075 2.21 4.293 3.700 2.22 4.956 4.385 2.28 5.662 5065 2.31 Six series of measurements have been made in which both water and sulphuric acid were employed. The amount of water used in the different series varied from 0.068 gram to 0.753 gram, ^ and varying amounts of sulphuric acid were added in each series. The results are given in the following Tables I to VI. Just 25 cc. acetic acid =: 26.335 grams were used in each series of measure- ments. The amount of water used in each series is given in the first line above the table. Column i gives the amounts of sulphuric acid which were added to the acetic acid containing the amount of water stated; Column 2, the lowerings of the freezing-point pro- duced by the different amounts of sulphuric acid, as read from the curve and corrected for the increase in volume caused by the presence of the water; Column 3, the lowerings produced by the water as read from the curve and corrected for the increase in volume produced by the sulphuric acid. This value decreases as the increase in sulphuric acid present increases the volume. Column 4 gives the sum of the lowerings of the sulphuric acid and water as thus corrected ; Column 5, the lowerings of the freezing- point actually found when the sulphuric acid and water were brought together in the acetic acid; Column 6, the difference between the calculated lowering, if the two did not combine with each other, and the lowering found. 8 Jones. Lowering of the Free zmg-point of Acetic Acid by Sulphuric Acid and Water. Table I. 26.335 grams acetic acid, 0.068 gram water. Amounts of H,SO, in grams. Lowering by H5SO4. Lowering by Sum of the last two. Lowering found. Difference 0.088 0.160° 0.540° 0.700° 0.630° 0.070'' 0.178 0.290 0.540 0.830 0.760 0.070 0.294 0.430 0.535 0.965 0.820 0.145 0.410 0.550 0-535 1.085 0.870 0.215 0.642 0.800 0.530 1-330 0.950 0380 0.883 1. 000 0.530 1-530 1.060 0.470 1-347 1.360 0.525 1.885 1.280 0.605 2.050 1.870 0.520 2.390 1.740 0.650 3-312 2.830 0.505 3-335 2.658 0.677 4.898 4-340 0.490 4.830 4.IIO 0.720 Table II. 26.335 grams acetic acid. 0.126 gram water. Amounts of H,S04 in grams. Lowering by H,SO,. Lowering by Sum of the last two. Lowering found. Difference. 0.092 0,165° 0.950° 1.115° 1.002° 0.113° 0.760 0.895 0.935 1.830 1. 106 0724 1.422 1-415 0.920 2-335 1.392 0.943 2,098 1.900 0.910 2.810 1.792 I.OI8 2.824 2.440 0.895 3-335 2.300 1-035 3-498 3.000 0.885 3-885 2.832 1053 4.187 3.600 0.875 4-475 3-447 1.028 4.860 4.280 0.865 5-145 4-074 1.071 5-544 4.960 0.855 5-815 4-749 1.066 Table III. 26.335 grams acetic acid. 0,255 gram water. Amounts of H2SO4 Lowering by in grams. H2SO4. 0.108 0.185° 0.321 0.455 0.565 0.705 Lowering by H5O. 1.860° 1.845 I 840 Sum of the last two. 2-045" 2.300 2-545 Lowering found. Difference. 1.862° I-750 1-655 0.183° 0.550 0.890 The Hydrates of Sulphuric Acid. Amounts of H5SO4 in grams. Lowering by Lowering by H,0 Sum of the last two. Lowering found. Difference. 1.029 1.110° 1.815° 2-925° 1.580" 1-345° I.717 1.620 1.800 3.420 1.690 1.730 2.406 2.100 1.770 3.870 2.013 1.857 3-II3 2.655 1-750 4-405 2.470 1-935 3-794 3.230 1.720 4.950 2.955 1.995 4.460 3.830 1-705 5-535 3565 1.970 5-133 4.515 1.685 6.200 4.230 1.970 5-9II 5-230 1.660 6.890 5.080 1.810 Table IV. 26.335 grams acetic acid. 0.496 gram water Amounts o) H2SO4 in gramsv Lowering by H,S04. Lowering by H,0. Sum of the last two. Lowering found. Difference . 0.201 0.305° 3-310° 3.615° 2.980° 0.635° 0.651 0.780 3-270 4.050 2.439 1. 611 I-33I I-315 3.220 4-535 2.060 2.475 2.055 1.835 3-175 5.010 2.052 2.958 2.744 2.335 3-135 5.470 2.280 3.190 3403 2.860 3.085 5-945 2.655 3.290 4-095 3.460 3.055 6.515 3.170 3-345 4.808 4.160 3.010 7.170 3-790 3.380 5-497 4.820 2.975 7-795 4.410 3.385 6.190 5.470 2-945 8.415 5-030 3-385 Table V. 26.335 grams acetic acid. 0.728 gram water Amounts of HjSO, in grams. Lowering by H,S04 Lowering by H,0. Sum of the last two. Lowering found. Difference. 0.659 0.775° 4.520° 5.295° 3-340° 1.955° 1.367 1.330 4.450 5.780 2.680 3.100 2.056 1.820 4.380 6.200 2.457 3-743 2.752 2-335 4.310 6.645 2.510 4-135 3676 3.055 4.250 7-305 2.900 4405 4.354 3.660 4.200 7.860 3.350 4.510 5-050 4.360 4.140 8.500 3.870 4.630 5.728 4.980 4.080 9.060 4.560 4.500 6.187 5.410 4-065 9-475 5.020 4.455 Jones. Tabl E VI. 26.335 grams acetic acid. 0.753 gram water . Amounts of H5SO4 Lowering by in grams. H^SO^. Loweringby Sum of the last two. Lowering found. Difference 0.109 0.185° 4.690° 4-875° 4-474° 0.401° 0.224 0.330 4.670 5.000 4.263 0-737 0.447 0.585 4-645 5-230 3-850 1.380 0.679 0.795 4.615 5.410 3-540 1.870 1.209 1. 215 4-565 5-780 3.002 2.778 1.654 1.540 4-545 6.085 2-734 3-351 2-349 2.020 4-475 6-495 2.607 3.888 3. 1 19 2.600 4.410 7.010 2.752 4-258 3.829 3.185 4-345 7-530 3-037 4-493 4.520 3.840 4-295 8.135 3-495 4.640 5.218 4.510 4-235 8-745 4.032 4-713 5-927 5.160 4.190 9-350 4-713 4-637 6.632 5-825 4.130 9-955 5-383 4-572 The results cannot be regarded as very exact, since some small sources of error are unavoidably present in them. Thus, when the lowering of the freezing-point for sulphuric acid as read from the curve was corrected for the increase in volume caused by the presence of the water, this correction was based on the assumption that the water simply played the role of so much more acetic acid. Again, the lowerings produced by the water were corrected for the increase in volume due to the sulphuric acid, on the same assumption for the sulphuric acid. The assumption was exactly this, that a given amount of water would produce the same lower- ing of the freezing-point, say with 28 cc. acetic acid as with 25 cc. acetic acid and 3 cc. sulphuric acid, if it did not combine with the sulphuric acid. The contraction in volume when the sulphuric acid or water was added to the acetic acid was also taken into account in these corrections. Doubtless the above assumption does not hold rigidly, and slight errors are the consequence, but it is fairly certain that their magnitude cannot be great. However, the main features in this combination of sulphuric acid with water appear so clearly from the results, that they would be entirely unmasked by the small errors to which reference is above made. The Hydrates of Sulphuric Acid. 1 1 In the series of determinations given in Table VI the largest amount of water was used. The quantity of sulphuric acid added in the first determination was 0.109 gram. There were then pres- ent about thirty-eight equivalents of water to one of sulphuric acid. Yet from the column of " difference " it will be seen that the "difference" is very nearly twice the lowering produced by the sulphuric acid alone. This means that each molecule of the sulphuric acid has combined with two, and only two, molecules of water, which have disappeared, so far as the lowering of the freez- ing-point is concerned. The " difference " is slightly greater than twice the lowering produced by the sulphuric acid alone, but this is doubtless due to the fact that an equivalent of water produces a greater lowering of the freezing-point than an equivalent of sul- phuric acid. This is seen from the gram-molecular lowerings of the two in Tables A and B. It is, however, impossible to say whether the molecule H2SO4.2H2O would give exactly the same lowering as the molecule H2SO4 or as the molecule H2O. As the quantity of sulphuric acid approaches one half-equiva- lent to the water, the " difference " becomes less than the lowering produced by two molecules of water to each of sulphuric acid. At one half-equivalent of sulphuric acid this "difference" is consid- erably less, showing that at this point some of the water remains uncombined, and some of the sulphuric acid has probably not formed a hydrate higher than H2SO4.H2O. As the quantity of sulphuric acid is increased beyond one half-equivalent, more and more of the hydrate H2SO4.H2O seems to be formed, and when an equivalent of sulphuric acid is present the water is practically all combined with the acid, probably as this hydrate. The fact that when one half-equivalent of sulphuric acid is present a part of the water is not combined to form this hydrate H2SO4.2H2O, is doubtless due to the action of the acetic acid, dissociating this compound. That the acetic acid can probably cause such disso- ciation will be seen later when we consider the more dilute solu- tions of the hydrates in the acid and when a smaller excess of water is present. The same general facts brought out by the results in Table VI are also established by the results in Tables IV and V, where slightly smaller quantities of water were used. In the series of determinations, the results of which are given in Tables III, II and I, the amounts of water used were very much 12 Jones. smaller, and the solutions of the hydrates in the acetic acid, when all the water was combined, were very much more dilute. Under these conditions some entirely new facts have appeared. Let us consider, e. g., the second determination in Table III, and the " difference " is far less than would correspond to the hydrate HsS04.2HiO, yet more than four equivalents of water are present to one of sulphuric acid. The same fact is seen even more clearly in the first determination of Table II, where more than seven equivalents of water are present to one of sulphuric acid, and where the entire amount of water is small. Here the " difference " is less than it would be if each molecule of sulphuric acid were combined as H2SO4.H2O. Under these conditions of dilution in the acetic acid there must be free sulphuric acid, and this in the presence of seven equivalents of water. Yet we have seen from the results in Table VI that when sulphuric acid and water are present in these proportions and in quantities sufficient to give a much more concentrated solution of the hydrates in the acetic acid, every molecule of the sulphuric acid is in combination with two molecules of water. To test this question of the dissociation of the hydrates still further, a very small amount of water (0.068 gram) was added to the acetic acid and then small amounts of sulphuric acid added, and the freezing-points very carefully determined. The results of these measurements are given in Table I. In the first deter- mination more than four equivalents of water are present, yet the " difference " shows that less than one-half of the sulphuric acid is in combination with even one molecule of water. In the fourth determination somewhat more than an equivalent of sulphuric acid is present, yet less than half of the water is combined. For this dilution it is only when the amount of sulphuric acid has somewhat exceeded two equivalents that all the water is com- bined, yet for the concentrations represented in Table VI all the water was combined when one equivalent of acid was present. There was just a possibility that the explanation of the above results was not to be found in the dissociation of the hydrates of sulphuric acid, but in an electrolytic dissociation of the sulphuric acid itself in the acetic acid or in the acetic acid and water. It was improbable that any large amount of electrolytic dissociation should take place at these concentrations ; yet to obtain light on this point the " conductivity " of the solutions must be determined. The Hydrates of Sulphuric Acid. 13 The conductivity of the pure acetic acid was at first determined, then the conductivity of the acid to which small amounts of water had been added ; also the conductivity of the acetic acid to which small amounts of sulphuric acid had been added, and finally the conductivity of the acetic-acid solutions containing both water and sulphuric acid. The conductivity was calculated from the well-known formula w.b //=: molecular conductivity; A'l^the "capacity" of the vessel used; z' = the volume of the solution in liters which contains a gram-molecular weight of the electrolyte ; a =: reading on the left side of the bridge; ze/ = the resistance in ohms introduced into the circuit; <^=: reading on the right side of the bridge. The results of these determinations are given below. Conductivity Results. Conductivity of pure acetic acid, /-t =1:0.00001. Conductivity of acetic acid with small amounts of water: Water added to lOCC. CH3COOH. V. M- 0.071 gram 2-5 0.0002 0.287 0.65 0.0004 Conductivity of acetic acid with small amounts of sulphuric acid : Sulphuric AcW added tOIOCC. CH3COOH. V. M. O.I 21 gram 0.371 0.918 1.798 8.1 2.7 1,12 0.6 0.098 0.465 1-559 2.506 3-095 4-403 0.37 0.27 2.313 1.989 5-499 0.23 1-973 Conductivity of acetic acid to which small amounts of water and sulphuric acid have been added : H Jones. Series I. Amount of Water added to 10 CC. CH3COOH. Amounts of Sulphuric Acid added to , lOCC. CH3COOH+H2O. V. ft. 0.071 gram 0.042 gram 23.35 0.331 0.085 11.64 0.632 0.198 5-05 0.687 0.565 . 1.80 1.524 1. 175 0.88 2.256 1.723 0.62 2.800 Series II. Amount of Water added to 10 CC. CH3COOH. Amounts of Sulphuric Acid added to lOCC. CH3COOH + H5O. V. M- 0.287 gram 0.108 gram 9.34 2.219 0.635 1.63 2.770 1.712 0.64 2.888 2.990 0.39 2.679 4.367 0.28 2.347 The conductivity of the pure acetic acid was nearly zero, as would have been expected. The conductivity of acetic acid to which small amounts of water were added was also very slight. This agrees well with what is already known, since Ostwald' has shown that the dissociation of a tenth-normal solution of acetic acid in water is only about one per cent., and the same fact has been established by my own work on the lowering of the freezing- point. The results show, however, that when sulphuric acid is added to the acetic acid there is a slight conductivity. It is probable that it is the sulphuric and not the acetic acid which is dissociated, for the dilution of the sulphuric acid in the acetic is much greater than that of the acetic in the sulphuric, and, for the same concen- trations, sulphuric acid is much more readily dissociated in water than is acetic acid. The " volume," v, is calculated for the sulphuric acid in the acetic acid as a solvent. The " molecular conductivity," /x, increases with increase in concentration to a maximum and then diminishes slightly. Kablukoff ' found that the '.' molecular con- > Ztschr. phys. Chem. 3, 174. The Hydrates of Sulphtiric Acid. 15 ductivity" of hydrochloric acid in ether and isoamyl alcohol decreased with increase in the dilution of the solutions. When both water and sulphuric acid were added to the acetic acid in quantities as small as are given above, the solutions show also a slight conductivity, but only a little greater than when the sulphuric acid was added alone. It seems, therefore, that the small amount of water present exerts but little effect on the elec- trolytic dissociation. Since this is true, the " volume" v was here also calculated for the sulphuric acid in the acetic acid. It is at present impossible to calculate the exact amount of dis- sociation of these solutions in acetic acid from the known values of ixv, because the value of /-/.oo for sulphuric acid in acetic acid is unknown. This would probably involve a determination of the velocity of migration of the ions of sulphuric acid in acetic acid. But the very small values found for iiv render it very probable that the electrolytic dissociation of the sulphuric acid under the above conditions is small, and far too small to account for the results obtained. The only satisfactory explanation of the results, then, seems to be the one offered,— that the hydrates of sulphuric acid are un- stable in very dilute solutions in acetic acid when the excess of water present is not very considerable. Under the same conditions an excess of sulphuric acid would probably have a similar effect in increasing the stability of these hydrates. The Action of Water and Ethyl Alcohol on the Freezing-point of Acetic Acid. The lowerings of the freezing-point of acetic acid produced by different amounts of alcohol alone were first determined and plotted as Curve 3. The ordinate-unit=|o.in. and the abscissa- unit =0.2''. The curve deviates only slightly from a straight line. With the exception of the first determination, the molecular lowering also changes only slightly. The object in studying the action of the alcohol and water together on the lowering of the freezing-point of acetic acid, was to determine if any combination between the two took place. The results show that the lowering found is very nearly the sum of the lowerings produced by each separately in the presence of the other. The small "difference," which increases slightly with the 1 6 Jones. amount of alcohol, is possibly caused by the fact that the correc- tions introduced for changes in volume due to the addition of water and alcohol are not exact. Lowering of the Freezing-point of Acetic Acid by Small Quantities of Ethyl Alcohol. Table VII. Vt. Alcohol added to 26.335 grams CH3COOH. Lowering of the Freezing-point. Gr.-mol. Low. 46 grams C,H.O 1000 grams CH3COOH. 0,087 gram 0.305" 4-25 0.261 0.850 3-95 0443 1.450 3-96 0.617 2.000 3-91 0.791 2.510 3.84 0.976 3.038 3-77 1. 168 3-565 3.70 1-343 4-059 3-66 1.518 4-555 3-63 1.696 5-045 3-60 Lowering of the Freezing-point of Acetic Acid by Ethyl Alcohol and Water. Table VIII. 26 .335 grams acetic acid. 0.245 gram water. Amounts of CaH^OH. Lowering by C2H5OH. Lowering by H.,0. Sum of the Lowering last two. found. Difference. 0.085 gram 0.305° 1.760° 2.065° 2.050° 0.015° 0.273 0.880 I -740 2.620 2.575 0.045 0.452 1.465 1.720 3-185 3-105 0.080 0.621 2.000 1-705 3-705 3-557 0.148 0.780 2-455 1,690 4.145 3,968 0.177 I-I55 3-505 1.665 5,170 4,975 0.195 These determinations were the more desirable because Men- del6eff' has stated that the specific gravities of solutions of alcohol in water indicate the following compounds: 3OH6O + H2O, C2H6O + 3H2O, OH60+i2H=0. • Ztschr. phys. Chem. 1, 284. The Hydrates of Sulphuric Acid. 17 My results furnish no evidence whatsoever in favor of the exist- ence of such compounds. It may be said, of course, that such compounds are unstable in the acetic acid. Yet it seems probable that if there were any strong attraction between the water and the alcohol, it would manifest itself at least to some extent in the first determination, where some eight equivalents of water are present to one of alcohol, and where the total excess of water present is not inconsiderable. Pickering' has also studied the action of the same compounds on the lowering of the freezing-point of acetic acid, and found that no combination between the two took place. Action of Water and Dry Sodium Acetate on the Freezing-point of Acetic Acid. The neutral acetate of sodium combines with three molecules of water, and many of the acid acetates also contain water. The following have been described by Vielliers '?■ OH30=Na.OH402, OH302Na.2C2H40=.H.O, 5C2H3O2Na.4OH4O2.6H2O, 4C2H302Na.C2H402. 1 1 H2O. 5C2H302Na.2C2H402. 1 3 H2O. Which is formed depends, according to Vielliers, not simply on the relation between the amounts of acetate and acid, but also on the amount of water present. If the acetate which existed in the solution of acetic acid combined with water, this would be shown by the lowering of the freezing-point of the acetic acid. Sodium acetate was therefore carefully dried, and the lowering of the freezing-point produced by this alone determined. The results are plotted in Curve 4, in which the ordinate-unit^o.in, and the abscissa-unit rr 0.2°. The molecular lowering increases slightly with the amount of acetate. This may be due to the dis- appearance of a small amount of acid forming the acid acetate. The results in Table X seem to show that a very small amount of water was combined with the salt. Those in Table XI show a much greater "difference," indicating the combination of more water, but even here the amount is far less than would correspond to one molecule of water for each molecule of the acetate. 1 Ber. d. chem. Ges. 34, 1584. sCompt. rend. 8*, 774; 85, 755. Vol. XVI.— 2. 8 Jones. Lowering of the Freezing-point of Pure Acetic Acid by Small Amounts of Sodium Acetate. Table IX. Wt. CHjCOONa added to 26.335 grains CHjCOOH. Lowering of the Freezing-point. Gr.-mol. Lowering 82 grams CHsCOONa 1000 grams CH3COOH 0.109 gram 0.210° 4.16 0.288 0.600 4-50 0.568 I-I45 4-35 0.857 1.690 4.26 1.246 2.488 4-31 1.586 3.180 4-33 1-945 3-940 4-37 Lowering of the Freezing-point of Pure Acetic Acid by Sodium Acetate arid Water. Table X. 26.335 grams acetic acid. 0.606 gram sodium acetate. Amounts of H2O Lowering by in grams. H^O. Lowering by CHsCOONa. Sum of the last two. Lowering found. Difference 0.145 i-o8o° 1.220° 2.300° 2.360° —0.060° 0.279 2.010 1.205 3-215 3.180 +0.035 0.409 2.810 1.200 4.010 3905 0.105 0-553 3-625 I-I95 4.820 4.640 0.180 Table XI. 26.335 grams acetic acid. 1.645 grams sodium acetate. Amounts of H2O Lowering by in grams. H.^O. Lowering by CHaCOONa. Sum of the last two. Lowering found. Difference 0.085 0-645° 3.310° 3-955° 3-585° 0.370° 0.194 1.430 3-285 4-715 4.210 0-505 0.315 2.230 3.270 5-500 4.870 0.630 Coyiclusion. The chief interest connected with the results obtained in this work seems to be in connection with the facts brought to light in reference to the hydrates of sulphuric acid. When the quantity of water in proportion to the acetic acid is as great as in Table VI, and an equivalent of sulphuric acid is added, the water is all The Hydrates of Sulphuric Acid. 19 in combination, and probably as the hydrate H2SO4.H2O. If under these conditions, say a tenth of an equivalent of sulphuric acid be added to the acetic acid and water, every molecule of sulphuric acid is in combination with two molecules of water as HiSO^.sHaO =S(OH)6, which is the normal hydrate. When as much as thirty-seven equivalents of water are present to one of sulphuric acid, there is no evidence of the formation of hydrates with more water than S(OH)6. Evidence is furnished of the existence of the compounds H2S04.HiO and H2SO4.2H2O. The higher hydrates containing a much greater number of molecules of water to the sul- phuric acid, which have been thought by Mendel^eff to be probable, find no support from my results. On the contrary, if such did exist one would expect to obtain some indication of their forma- tion under some of the conditions which have been here employed. When, however, much less water is present in proportion to the acetic acid, as in Table I, more than two equivalents of sulphuric acid must be present before all the water is in combination with the acid. If under these conditions, say a fifth of an equivalent of sulphuric acid be present, much of this remains free, and the remainder probably combined only as H2SO4.H2O. These hy- drates are somewhat unstable in the acetic acid when their solu- tions are very dilute, and when the excess of water present is not very great. They can be regarded as dissociated under these con- ditions by the acetic acid into sulphuric acid and water. The hydrates of alcohol, which Mendeleeff supposed to have been indicated by the specific gravity of the solutions of alcohol in water, find no support from the results obtained in connection with this work. In conclusion, I wish to express my sincere thanks to Dr. Svante Arrhenius, from whom valuable suggestions were received, and in whose well-equipped laboratory in Stockholm this piece of work was carried out. Physical Institute, Stockholm, September, 1893. 20 Keiser and Breed. THE ATOMIC WEIGHT OF PALLADIUM. By Edward H. Keiskr and Mary B. Breed. In an investigation by one of us, published in 1889,' upon the atomic weight of palladium, it was found that palladium diammo- nium chloride, Pd(NH3Cl)2, is a much more suitable compound for the determination of the atomic weight of this metal than the double chloride of palladium and potassium that had previously been used by Berzelius for this purpose. The palladammonium chloride is a compound that can be prepared in a state of great purity ; it can be thoroughly dried without undergoing decompo- sition ; it is not hygroscopic, and the amount of palladium in it can be accurately determined by heating it in a current of hydrogen, whereby ammonium chloride is volatilized and metallic palladium remains behind. As the result of nineteen determinations made in this way upon different preparations of the compound, the pumber 106.35 was obtained for the atomic weight. This number was calculated directly from the weights of the substances in air. If the weights be reduced to a vacuum, then the value of the atomic weight becomes 106.27. Recently several investigators have worked upon this subject and have published results, some of which are considerably higher and others decidedly lower than the number 106.3. Thus Bayley and Lamb,^ after examining a number of palla- dium compounds for the purpose of devising a new method of determining the atomic weight, have adopted the method described above, namely, reduction of palladammonium chloride in a cur- rent of hydrogen, and find as the mean of ten determinations the value 105,46. From a series of chlorine determinations in the compound they obtained the number 106.37, but this they regarded as less accurate than the other. Keller and Smith' have also published the results of their determinations of the atomic weight of palladium. They too have determined the ratio of palladium to palladammonium chloride. Weighed quantities of the compound were dissolved in dilute ammonia, and the solution subjected to electrolysis in platinum dishes containing a deposit of metallic silver. The "This Journal 11, 398. -J. Chem. Soc. (1892) 61, 745. 3ThisJourn.il 14,423. The Atomic Weight of Palladium. 21 increase in weight of the dish was taken as the weight of the palla- dium deposited. Nine determinations were made, from which the number 106.91 was obtained for the atomic weight. Joly and Leidie' have made some preliminary experiments on the atomic weight of palladium. Weighed quantities of the double chloride of potassium and palladium were dissolved in water, and the metal precipitated by electrolysis. From the ratio of palladium to potassium palladium chloride they obtain the number 105.44 for the atomic weight. The different determinations, then, stand as follows : Berzelius, 1829, 106.2. Keiser, 1889, 106.3. Bayley and Lamb, 1892, i05-5' Keller and Smith, 1892, 106.9. Joly and Leidie, 1893, 105.4. As these recent determinations differ from one another by as much as one and a half units, it seemed desirable to take up the work again and to endeavor to establish this constant with a greater degree of certainty than has yet been attained. We have endeavored in the first place to find some compound of palladium that could be vaporized, and therefore subjected to fractional distillation. For, as Stas has shown, it is only the substances that can be vaporized that can be obtained in a high state of purity. Distillation of Palladium Chloride in a Current of Chlorine. We have found that palladium dichloride, PdCla, can be distilled at a low red heat in a current of chlorine. For the purpose of obtaining palladium chloride and purifying it by distillation, spongy palladium, obtained by the reduction of palladium diam- monium chloride — the metal remaining from the atomic weight determinations of 1889 — was placed in a thin layer in one end of a combustion- tube. In the middle of the tube there was a constric- tion which prevented liquid from flowing from one end to the other when the tube was in a horizontal position. The end of the tube containing the palladium was connected with an apparatus furnishing a stream of dry chlorine. The tube was placed in a combustion-furnace and so arranged that that portion of the tube ' Compt. rend. (1893) 116, 146. 22 Keiser and Breed. containing the palladium could be heated to a bright red heat. The chlorine was prepared from hydrochloric acid and manganese dioxide, and after washing with water was led through two bottles containing concentrated sulphuric acid. The gas after passing through the tube containing the palladium and before escaping into the air, was conducted through water contained in two wash- bottles to remove palladium chloride from it. After the chlorine had displaced all the air in the apparatus, that portion of the tube containing the palladium was heated in the furnace. When the temperature began to approach a low red heat the metal combined rapidly with the chlorine. Soon it began to glow, and in a few moments the entire quantity of metal had been converted into chloride. The heating was continued and a slow current of chlorine passed through the tube. The palladium chloride melted and became a dark liquid, and soon a red sublimate began to appear in the cold portion of the tube. The heating was continued, and after a time the entire cross-section of the tube for a length of several inches became filled with minute dark-red, needle-shaped crystals. The molten chloride in the front portion of the tube appeared to be boiling, as gas-bubbles were slowly evolved from it. After a considerable quantity of the crystalline sublimate had been prepared in this way the tube was allowed to cool and the crystals removed. An analysis showed that they were palladium dichloride, PdCh. 0.20II gram gave on reduction in a current of hydrogen 0.0804 gram CI. Calculated for PdCIj. Found. CI 40.01 39-98 When the crystals are heated in a stream of chlorine they melt and form when cold a dark-red crystalline mass consisting of dichloride mixed with monochloride. The crystals are deliques- cent and absorb moisture in the air. Several distillations of palladium chloride in a current of chlorine were now made, and a quantity of the distilled chloride weighing a little more than ten grams was obtained. This was redistilled in the same manner, and of the distillate the more vola- tile and the least volatile portions were rejected and only the middle portion used for the subsequent work. We hoped at one time to use the dichloride directly for the determination of the ratio of palladium to chlorine, but owing to its The Atomic Weight of Palladium. 23 deliquescent nature and the fact that when melted the dichloride is partially converted into monochloride, we abandoned experiments in this direction and decided to convert the dichloride into palla- dium diammonium chloride. For this purpose the redistilled chloride was dissolved in distilled water to which a few drops of hydrochloric acid had been added. Ammonia was added gradu- ally and the solution warmed upon the water-bath until the flesh- colored precipitate had dissolved. The filtered solution was diluted with a large volume of water, and into the dilute solution a current of pure hydrochloric acid gas was passed. The pallad- ammonium chloride thus formed was washed repeatedly by decantation, and finally upon the filter. It was then dissolved in dilute ammonia and reprecipitated with hydrochloric acid gas, thoroughly washed and then dried in desiccators. Analysis of the Palladium Diavimoyiium Chloride. The ratio of palladium to palladammonium chloride was deter- mined by reducing weighed quantities of the dried compound in a current of hydrogen. The palladammonium chloride was weighed in platinum boats. It was dried in an air-bath at a tem- perature of iio°-i20° until constant weight was obtained. The boat was then introduced into a combustion-tube through which a current of pure hydrogen was passed. The hydrogen was made by the action of pure dilute hydrochloric acid upon pure zinc, and was washed with alkaline lead solution and potassium permanga- nate. It was then dried by passing over solid caustic potash, after which it passed over red-hot metallic copper, and, finally, before entering the reduction-tube it was dried with phosphorus pent- oxide. After the hydrogen had been passing through the reduc- tion-tube for some minutes and all air had been expelled, the portion of the tube containing the boat was heated very gently. The palladammonium chloride unites with the hydrogen very gradually under these circumstances. There is no decrepitation whatsoever, and by gently raising the temperature the ammonium chloride slowly sublimes. The sublimate in all cases was found to be perfectly snow-white in appearance. There was absolutely no loss of palladium by volatilization nor by decrepitatiofi,^ as was 1 Keller and Smith: This Journal 14, 423, found that palladium was volatilized. This was no doubt due to a rapid reduction of the imperfectly dried compound in a strong current of hydrogen. The metal used by one of us (K.) in the earlier determinations was not, as they suppose, obtained from Eimer & Amend in New York. 24 Keiser and Breed. shown by a careful examination of the aqueous and acid washings of the tube after the reduction had been finished. In neither solu- tion could a trace of palladium be found when tested with potas- sium iodide or with hydrogen sulphide. It is only when an imperfectly dried specimen of palladammonium chloride is heated very rapidly in a current of hydrogen that there is loss by spatter- ing and the mechanical carrying away of minute particles of palladium. The results of the analysis of the palladammonium chloride prepared from distilled palladium chloride are as follows : Weight of Weight of Number. PdNaH^Clj taken. Pd obtained. Atomic Weight. 1 1.60842 0.80997 106.271 2 2.08295 1.04920 106.325 3 2.02440 1-01975 106.334 4 2.54810 1.28360 106.342 5 1-75505 0.88410 106.341 Total, 10.01892 5.04662 [106.325] Reducing now the weights in air to the weights in a vacuum, by adding the weight of air displaced, we obtain the ratio : PdN2H6Cl2: Pd:: 10.02373:5.04717 or Pd= 106.246. In this calculation the specific gravity of palladium is taken as II, and that of the palladium diammonium chloride as 2.5, values found as the result of duplicate determinations. The following atomic weights have been used: H:=i, N=: 14.01, = 35.37. The weight of a cubic centimeter of air is taken as 0.0012 gram. Purification of Palladitini in the Wet Way. In order to test the accuracy of the determinations made in 1889, we decided to purify some palladium by the method then used, with additional precautions. A piece of palladium foil, weighing about fifteen grams, was therefore dissolved in aqua regia, and the solution of dichloride thus obtained evaporated to dryness. The residue was then dissolved in water, to which a little hydrochloric acid was added, and the filtered solution was treated with ammonia and warmed upon the water-bath until the precipitate first formed was dissolved. The insoluble residue. The Atomic Weight of Palladium. 25 containing iron and other impurities, was filtered off; but the filtrate was pale bluish-green in color, owing to the presence of copper. Hydrochloric acid gas, made by allowing pure concen- trated sulphuric acid to act on pure, strong hydrochloric acid, and washed through a concentrated solution of hydrochloric acid, was passed into the solution. The palladium diammonium chlo- ride thus precipitated was washed by decantation and then upon the filter with the aid of a suction-pump, and afterwards redis- solved in very dilute ammonia. This process of precipitating and dissolving the substance was repeated five times. All rhodium and the last trace of iron were thus removed ; the bluish color gradually disappeared from the ammoniacal solution, which became a clear straw color, showing that all copper had been removed. The salt was then dried in desiccators, and the metal obtained from it by reduction in hydrogen, as before described. The spongy metal was then again dissolved in aqua regia, evaporated repeatedly with hydrochloric acid to expel nitric acid, and the chloride dissolved in very dilute hydrochloric acid. In order to make sure that no gold was present, sulphur dioxide was passed into the solution. The gas was prepared from pure sulphuric acid and copper, was washed through water, and run into the solution of palladium until the color changed from dark red to yellow. The solution was protected from dust and allowed to stand for a week, but no gold was found to be precipitated, and the solution was therefore evaporated to dryness to expel sulphur dioxide. The residue, after being dissolved in water containing hydrochloric acid, and treated with ammonia, was converted into palladammonium chloride in the same manner as before. At this point, however, it became evident that not only had the intro- duction of sulphur dioxide been unnecessary for the purification of the substance, but had become itself a source of difficulty ; for the palladammonium chloride was now dull brownish yellow, instead of pure orange-yellow as before. Dissolving and repre- cipitating did not alter the appearance of the salt, and when dried it turned dark brown on all the surfaces exposed to the air. We decided that it would be necessary to reduce to the metallic form in order to expel sulphur. This was done in the same way as before, and the reduced metal was again converted into pallad- ammonium chloride. Though the freshly precipitated salt was now of the proper color, yet it changed as soon as it dried, becom- 26 Keiser and Breed. ing dull lemon-yellow throughout, and turning brown on the exposed surfaces. That the source of difficulty lay in some foreign substance present in the salt itself was shown by the fact that the change in color took place when the most careful precautions were taken to exclude all laboratory fumes from the workroom, and when the salt was placed in the desiccator as soon as the wash- ing was completed. After precipitating and redissolving the salt three times, but without removing the impurity, we reduced it again in a current of hydrogen, and then, placing the reduced metal in a hard-glass tube, through which hydrogen was passing, heated it in the combustion-furnace and kept the metal at a bright-red heat for five minutes. After cooling it was dissolved again in aqua regia, evaporated repeatedly with hydrochloric acid, dis- solved in water and a few drops of the acid, and the solution was then treated with mercuric cyanide in slight excess. The palla- dium cyanide thus precipitated was allowed to settle, and was washed by decantation repeatedly. It was then washed on the filter, dried in desiccators, and reduced to metallic form by heating in the air. Cyanogen gas was given off, and after the reaction was complete, the heating was continued for about an hour, to expel any mercury that might have been retained mechanically by the very gelatinous precipitate of palladium cyanide. The metal, which was partly oxidized, was reduced in a Rose crucible in a current of hydrogen, dissolved in aqua regia, the solution evaporated repeatedly with hydrochloric acid and then digested with ammonia. Just as the precipitate thus formed was about all dissolved, a white crystalline precipitate separated in small quantity, which, when filtered off, proved to be the so-called "white precipitate" formed by the action of ammonia on mer- curic chloride. The small quantity of mercury that could not be removed by heating was thus separated by filtration, and the palladium was then precipitated as palladammonium chloride by running in hydrochloric acid gas. The precipitate, after washing, was dissolved and reprecipitated, to prove that all mercury had been removed, washed and dried for several months in a desic- cator over solid caustic potash. Analysis of Palladium Diammonium Chloride piirified by Precipitation. The palladium in the salt was determined in the same way as in that purified by distillation — i. (N02)2.CHCl3, which crystallized in blackish-red, well-formed prisms, but lost its chloroform partially ' Ber. d, chem. Ges. 33, ii. Trianilidodinitrobenzol and Certain Related Compounds. 37 even at ordinary temperatures. A similar addition-product was obtained with the corresponding para- but not with the ortho- toluido compound; an attempt will be made in this laboratory to trace the limits of this reaction. Numerous attempts to prepare a second modification of trianili- dotrinitrobenzol or of anilidotrinitrophenylmalonic ester have without exception led to negative results, as no change in the full yellow color or the crystalline form of either of these substances could be observed; but this result seems so strange in view of the occurrence of two modifications of the trianilidodinitrobenzol on the one hand, and of the anilidotrinitrophenyltartronic ester on the other, that a more careful study of these substances will be made hereafter, and the work extended to other substances, in the hope of collecting in time enough observations to determine the cause of the isomerism of the substituted tartronic ester. Experiments with Trianilidodinitrobenzol. In searching for isomeric forms of the trianilidodinitrobenzol we studied more carefully than before the crystallization of the sub- stance from various solvents, and found that, when a mixture of benzol and alcohol was used as the solvent, it appeared in crystals of two sorts : one, the nearly square prisms of an orange color like that of potassic dichromate already described; the other, as yellow as potassic chromate, in bladed crystals, or plates looking like flattened monoclinic prisms terminated by two planes, or less commonly with square ends, which, when the cooling took place rapidly, appeared in circular groups of little needles. These two modifications differed entirely in crystalline habit and color, and resembled strongly the yellow and red forms of anilidotrinitrophe- nyltartronic ester, the discovery of which had led us to undertake this work; but whereas the two esters showed a difference of 21° in their melting-points (red 143°, yellow 122°), the yellow and red forms of the trianilidodinitrobenzol melted at the same tempera- ture, 179°. To be sure, the yellow form turned red at about 140**, but there were no signs of melting, and we do not feel that such a change from yellow to orange-red is definite enough to have much significance. The yellow form was also much less stable than the corresponding form of the anilidotrinitrophenyltartronic ester, so that we have not obtained it absolutely free from the 38 Jacksoji and Herman. orange modification. The best way for preparing it that we found was to crystaHize the orange form from a mixture of benzol and alcohol containing much benzol, when a portion of the substance usually appeared in the yellow form. On the other hand, a single crystallization of the yellow crystals from a mixture of benzol and alcohol containing but little benzol, was sufficient to convert them completely into the orange prisms. As it was possible that the yellow crystals might be a compound containing alcohol or benzol of crystallization, instead of an isomer, we heated to ioo° 0.3190 gram of the best we could obtain, which had been air-dried, but found that the loss was only 0.0004 g^'^""'. showing that this is not the explanation of the occurrence of this form. The presence of a small amount of impurity seemed to be favor- able to the existence of the modification crystallizing in plates, as, if a little tribromdinitrobenzol was present, crystals were obtained of this form, although usually a little more orange in color than that made from pure material, and these would undergo several crystallizations before they were converted into the orange form. A sample, which we obtained accidentally, was even more stable, and owed this stability to an oily impurity, the presence of which was indicated by the low melting-point, and which we finally suc- ceeded in separating, but only after a great number of crystalliza- tions, when the substance passed into the orange form ; the amount of this impurity, however, was very small, as shown by the following analyses: I. 0.2522 gram substance gave on combustion 0.6004 gram CO2 and 0.1086 gram H2O. II. 0.2197 gram substance gave 31.6 cc. N at 25° and 755.4 mm. Calculated for Found. C,H(C,H4NH)3(NO,)j. I. II Carbon 65-31 64.94 Hydrogen 4-30 4-79 Nitrogen 15.88 I5-S The substance gave no test for bromine with a copper wire, nor even when treated according to the method of Carius. As the amount of impurity is so small that we thought it could have no eflfect, we have used this preparation for the determination of the molecular weight of the yellow modification, which, with that of the orange prisms, was made by the method of Raoult, using benzol as the solvent, since preliminary experiments had shown Trianilidodinitrobenzol aiid Certain Related Compounds. 39 that these compounds were not sufficiently soluble in glacial acetic acid. Yellow Plates. Substance, 0.2290 gram; benzol, 10.299 grams; depression, 0.27°. Orange Prisms. Substance, 0.2561 gram; benzol, 10.468 grams; depression, 0.29°. From these results the following molecular weights are ob- tained : Molecular Weight. Yellow Plates 404 Orange Prisms 413 Calculated for C6H(C6H5NH)3(NO.> 441 There can be no doubt, therefore, that the substances are not polymeric. The benzol-solutions obtained in the determinations were mixed with a little alcohol, and allowed to evaporate sponta- neously (the substance is deposited as a varnish from benzol alone), when each yielded as the principal product the modifica- tion which had been originally dissolved in it, although in each case this was mixed with an insignificant amount of the other form. From the observations given above we should infer that these two modifications of the trianilidodinitrobenzol are not true chemical isomers, but physical isomers, or perhaps rather that the substance is dimorphous. We have made a great many experiments to get other isomeric forms of the trianilidodinitrobenzol, both by varying the method of preparation in every way we could devise, and by treating the orange prisms with reagents, which usually convert one stereo- isomeric form into another, but have met with no forms except the two already described. Compound of Trianilidodinitrobenzol and Chloroform. If in crystallizing the trianilidodinitrobenzol a mixture of chloro- form and alcohol was used instead of benzol and alcohol, dark red well-formed short prisms were obtained entirely different from 40 Jackson and Herman. either of the forms just described. The whole of the substance could be converted into these prisms, if the solution in chloroform and a little alcohol was allowed to evaporate at temperatures from 50° to 70°. This substance, however, proved to be, not an isomer, but a compound with chloroform, since on heating some of the dry crystals with sodic hydrate and aniline a strong odor of phenyl isocyanide was perceived. Our first attempts to analyze it showed that the chloroform was given up partially at ordinary temperatures, but that to drive oflf the remainder it was necessary to heat to 100°. Accordingly we proceeded as follows. A quantity of the orange trianilidodinitro- benzol melting at 179°, was dissolved in warm chloroform, and after the addition of a little alcohol poured into a large watch- glass to crystallize ; when nearly all the chloroform had evapor- ated, the crystals were pressed as quickly as possible between filter-papers till free from adhering chloroform, and then transferred at once to a stoppered glass tube, in which they were weighed, I. 1.5012 grams compound lost at 100° 0.3226 gram CHCls. II. 1.4455 grams substance lost at 100° 0.3059 gram CHCh. Calculated for Found. C6H(CeH5NH)3(N02)2.CHCl3. I. II. Chloroform 21.32 21.49 21.16 The residue was trianilidodinitrobenzol melting at 179°. Properties of the Addition- compound of Triayiilidodinitrobenzol and Chloroform. — This substance crystallizes in short thick prisms with both terminations well developed, apparently of the mono- clinic system, which have a very dark brownish-red color not unlike that of well-crystallized potassic ferricyanide, and show a blue reflex. The chloroform is not securely held, part of it being given up even at ordinary temperatures with great rapidity, whereas heating to 100° is necessary to drive out the last traces. On this account its presence has only a slight effect on the melt- ing-point, lowering it by a variable amount not exceeding 3° or 4°. Standing with alcohol if the crystals are small, or grinding them with it if they are large, also drives out the chloroform, leaving the usual orange prismatic form, and the same effect is produced by one crystallization from alcohol. Its action with other solvents was not studied. Triparaioluidodinitrobe7izol,C6y{(Ci\li^\l)i(HO-i)-.tt X q resulting from the action of the ' Geuther: Ztschr. Chem. 1865, 32. '■ Kekule and Zincke ; Lieben : Ann. Chem. (Liebig) 106, 336. > Ber. d. chem. Ges. 3, 468; Ann. Chem. (Liebig) 168, 150. ' Loc. ctt. ^Loc. cit. 46 Orndorff and White. various polymerizing agents on the simple aldehyde, which leads to the following structural formula : ">c<^"- CHs ---^r)^ ^CHs This formula is in strict accord with all the facts known, and has, moreover, received experimental confirmation at the hands ofBriihl,' who showed that the molecular refraction of paralde- hyde was Ma = 52.48, while that required by the above formula is Ma = 52.77. Paraldehyde does not react with phenylhydrazine nor with hydroxylamine, facts which are also in accord with the above formula. The solid polymer, metaldehyde, is formed from aldehyde by the action of the same polymerizing agents that bring about the formation of paraldehyde, a low temperature being most favor- able for its production. Only a small part of the aldehyde is converted into metaldehyde. Paraldehyde is always formed simultaneously with the metaldehyde and in much larger quan- tity. Metaldehyde differs from paraldehyde only in its physical properties, such as melting-point, crystalline form, solubility, volatility, etc. In its chemical conduct towards various reagents it acts exactly like paraldehyde ; thus it is partly converted into aldehyde on heating, and undergoes complete dissociation when heated in the presence of substances which polymerize the simple aldehyde. Like paraldehyde, it is not acted on by caustic alkalies or by oxidizing agents. It does not reduce an ammoniacal solu- tion of silver nitrate, unites with neither ammonia nor sodium bisulphite, nor does it react with hydroxylamine or phenylhydra- H I zine. Hence it does not contain the aldehyde group — C = 0. When it does react chemically it forms the same products that result from the action of the same reagents on aldehyde, probably owing to the re-formation of aldehyde in the presence of these substances ; thus with phosphorus pentachloride it forms ethyl- idene chloride, and with chlorine, chloral. Its conduct towards small amounts of hydrochloric acid gas is especially noteworthy. 'Ann. Chem. (Liebig) 803, 44. See also in this connection Ber. d. chem. Ges. 24, 657. Acetic Aldehyde, Paraldehyde and Metaldehyde. 47 According to Kekul6 and Zincke,' hydrochloric acid gas in con- tact with metaldehyde changes it gradually into paraldehyde containing some aldehyde. Troeger'' has noticed this same con- version of metaldehyde into paraldehyde, but in the case cited by him it took place apparently spontaneously. Regarding the molecular weight and structural formula of metaldehyde comparatively little is known. The substance undergoes partial dissociation into ordinary aldehyde when heated, and hence the vapor-density determinations made by Liebig,' Hofmann,* and Kekul6 and Zincke' gave unsatisfactory results. In this connection Kekul6 and Zincke state "that since the vapor-density of metaldehyde is not known and the molecular weight cannot be determined by any of the facts now in our pos- session, nothing definite can be said concerning its constitution. The formation of ethylidene chloride and the easy transformation into the simple aldehyde make it probable that several aldehyde molecules (perhaps two) are here joined by oxygen to form a more complex molecule." The formula of metaldehyde is there- fore generally given in treatises on chemistry as (C2H40)jr, indi- cating that the molecular weight is unknown. The vapor-density and consequently thie molecular weight ot metaldehyde have, however, been determined by Hanriot and Q^conomedes,^ both by the Dumas and the Hofmann methods. This they accomplish by ingeniously introducing a correction for the amount of metaldehyde-vapor converted into aldehyde and thus determining the true vapor-density of the undissociated metal- dehyde. Their results by both methods lead to the same con- clusion, viz. that the vapor-density and the molecular weight of metaldehyde are the same as those of paraldehyde. According to them, therefore, the formula for metaldehyde is the same as that of paraldehyde : (C2H40)3 or CeHi^Os. Hanriot and CEcon- omedes offer no explanation of this peculiar isomerism, which, if their results be correct, exists between paraldehyde and metal- dehyde, but content themselves simply with the publication of the results. Meldola and Streatfield" endeavored to determine the molecular weight of metaldehyde from the depression of the freezing-point 'Ann. Chem. (Liebig) 163, 149. 2 ggr. d. chem. Ges. 35, 3316. 3Ann. Chem. (Liebig) 14. 141. « Ber. d, chem. Ges. 4, 590. 'Ann. chim. phys. [5] 25, 22S. 'Chem. News 60, 66 and 67 (1889). 48 Orndorff and White. of solutions of metaldehyde in glacial acetic acid, but with nega- tive results — owing probably to the slight solubility of the metal- dehyde in this solvent. Quite recently — while we were engaged on this investigation, indeed — Zecchini' published a paper on the molecular weight of metaldehyde, in which he criticises the results of Hanriot and QEconomedes, and gives the results which he obtained using the boiling-point and freezing-point methods. With ether as a solvent he obtained very discordant results, which he thinks were due to the very slight solubility of the metaldehyde in this liquid. From the results obtained from the rise of the boiling- point of alcoholic solutions, he states it would be impossible to conclude precisely what is the true formula of metaldehyde, but asserts, nevertheless, that in alcoholic solutions its molecular weight is considerably higher than that required by the formula (C2H40)3. The formula would indeed seem to be, he says, not less than (C2H40)e. In solution in chloroform his results point to the same formula (C2H40)6, while in solution in phenol they lead to a formula intermediate between (C2H40)3 and (C2H40)4. Zecchini concludes his paper by saying that it is probable that the solid compound called metaldehyde has a high molecular complexity, and that, under certain circumstances, its molecules are separated into simpler molecules, perhaps corresponding to other polymers not yet isolated. Perhaps it would be advan- tageous, he continues, to repeat the experiments of Hanriot and CEconomedes, varying the quantity of the substance which is converted into vapor, and the conditions of temperature and pressure ; probably formulas different from those proposed by them would be found. It will be seen from this review of the literature of the subject that while the molecular weight and structural formula of paral- dehyde have been accurately determined, and all the methods of finding the molecular weight which have been used in the case of paraldehyde give concordant results, the molecular weight of metaldehyde is still in doubt, and, of course, the structural formula of the substance is unknown. In order to determine the molecular weight of metaldehyde and, if possible, the rela- tions existing between it and paraldehyde and the simple alde- hyde from which both are derived, the present investigation was undertaken. 1 Gazz. chim. ital. 33, 586. Acetic Aldehyde, Paraldehyde and Metaldehyde. 49 Experimental. To determine the molecular weight of metaldehyde we have made use of the cryoscopic method of Raoult, as well as of the vapor-density method of Hofmann. Preliminary experiments soon showed that phenol and thymol were the best solvents for metaldehyde, and that the metaldehyde underwent absolutely no dissociation in these solvents. This was proved by dissolving a weighed quantity of the metaldehyde in the liquid solvent, deter- mining the depression of the freezing-point several times, then treating the solution with ether and dissolving out the solvent, phenol or thymol, and recovering the metaldehyde. It was found that the metaldehyde could thus be recovered quantita- tively. Any dissociation of the metaldehyde would have been immediately shown by a change in the freezing-point of the solution. In every case the freezing-point was determined a number of times and no change was noted, so that it may be assumed that no dissociation of the metaldehyde takes place in either phenol or thymol. The metaldehyde was made according to the method given by Kekulg and Zincke. Pure paraldehyde was mixed with a few drops of concentrated sulphuric acid and distilled, the vapors passing through a Hempel distilling-tube one meter long filled with beads, to condense and return the undecom posed paralde- hyde. The aldehyde thus formed was condensed by passing it through a long Liebig's condenser and into a receiver connected with the condenser by means of a bent adapter. The receiver was surrounded with a freezing-mixture of ice and concentrated hydrochloric acid. A few bubbles of hydrochloric-acid gas were then passed into the aldehyde still surrounded by the freezing- mixture. Crystals of the metaldehyde began to separate out at once. After allowing the mixture to stand for half an hour these were filtered off (using a Witt plate and suction to drain thoroughly), and washed with alcohol and ether and dried on filter-paper. As the filtrate still had a strong odor of aldehyde, it was again treated with hydrochloric acid gas and another crop of metaldehyde crystals obtained. The filtrate, consisting now of paraldehyde alone, was distilled with sulphuric acid and the aldehyde regained was again subjected to the above treatment. Vol. XVI.— 4. 50 Orndorff arid U'htie. An examination of the metaldehyde crystals under the micro- scope showed that they were filled with inclusions of liquid, either paraldehyde or aldehyde, and when they were ground to powder the odor of paraldehyde was distinctly perceptible. The metaldehyde used for the molecular-weight determinations was therefore recrystallized several times from boiling chloroform, in which it is soluble to the extent of about four per cent. The metaldehyde undergoes dissociation into the simple aldehyde very rapidly in warm chloroform solution, so that it is expedient in recrystallizing metaldehyde to dissolve the metaldehyde in the boiling chloroform, filter quickly and cool the clear filtrate in ice- water. The crystals of metaldehyde formed in this way were small and needle-shaped. They contained no liquid inclusions, and when ground gave no odor of either paraldehyde or aldehyde. When separated from the chloroform solution and spread out on drying-paper, the solvent evaporates very rapidly and leaves the metaldehyde crystals in a very desirable form for further work. The crystals of the recrystallized metaldehyde were then ground to powder in an agate mortar and made up into the form of compressed tablets in the usual manner. Only the freshly crys- tallized metaldehyde was used in the following molecular-weight determinations, as it was found that metaldehyde undergoes decomposition on standing, being partially converted into paral- dehyde, recognizable by its odor. We shall refer to this change later in this paper, Beckman's method, apparatus and thermom- eter were used in the freezing-point determinations, and in working with phenol solutions care was taken to exclude all moisture. This was done by means of the simple device used by us in the determination of the molecular weight of benzoyl peroxide' in glacial acetic acid. The results of the freezing-point determina- tions with the freshly crystallized metaldehyde follow. Metaldehyde (C2H40)3= 132. Solvent, phenol ;" molecular depression, 75 ; depression-coef- ficient, 0.5682. ' This Journal 15, 353, and Ztschr. phys. Chem. 12, 63. *The phenol and thymol used in these determinations were prepared from the C. P. articles by redistillaiion. The purified products had the melting-points 4i°-4a° C. (uncorr.) and 49°-So° t)- (uncorr.), respectively. Acetic Aldehyde, Paraldehyde and Metaldehyde. 51 Wt. Solvent. Wt. Sub. Concentr. Depress. Dep.-coef. Mol. Dep. Mol. Wt. found. 20.45 0.0712 0.3481 0.206 0.5918 78.1 127 20.45 O.1615 0.7897 0.446 0.5647 74-5 133 20.45 0.2891 I.4I36 0.799 0.5652 746 133 20.45 0.4187 2.0474 1. 119 0.5465 72.1 137 21.86 0.1407 0.6434 0.332 0.5160 68.1 145 22.40 0.0578 0.2580 0.153 0.5930 78.3 126 22.40 O.15I9 0.6781 0.363 0.5353 70.7 140 22.40 0.2686 1.2000 0.622 0.5183 68.4 145 25.24 0.1004 0.3978 0.219 0.5505 72.7 136 Solvent, thymol ; molecular depression, 73.9 ; depression-co( ficient, 0.5598. Wt. Solvent . Wt. Sub. Concentr. Depress. Dep.-coef. Mol. Dep. Mol. Wt. found. 17.00 0.0780 0.4588 0.266 0.5800 76.5 128 17.00 0.1696 0.9976 0.558 0.5593 73.8 132 17.00 0.2892 I.7OII 0.869 0.5108 67.4 145 21.12 0.1469 0.6955 0.375 0.5391 71.2 137 These results certainly show that the molecular weight of the freshly prepared metaldehyde is 132, or three times that of acetic aldehyde (44), and that consequently the formula of metal- dehyde is the same as that of paraldehyde, (C2H40)3 or CcHiaOs. Determt7iaiio7i of the Vapor-density of Metaldehyde by the Hofmann Method. Having thus shown that in solution in phenol and in thymol the freshly prepared metaldehyde is a trimolecular compound, we next repeated the work of Hanriot and CEconomedes, and endeavored to determine the density of metaldehyde-vapor by the Hofmann method. As has already been stated, metalde- hyde undergoes partial dissociation into ordinary aldehyde when vaporized, so that to determine the vapor-density of metaldehyde by this method it becomes necessary to determine the weight and volume of the aldehyde formed by the dissociation of some of the metaldehyde in order to find the volume and weight of the metal- dehyde-vapor remaining undissociated. To do this we proceeded as follows : a weighed quantity of the freshly crystallized metalde- hyde was introduced into the inner tube of the Hofmann apparatus, which was then heated to the temperature of the vapor of boiling 52 Orndorff and White. aniline. The total volume occupied by the vapors of metaldehyde and aldehyde (formed by the dissociation of some of the metalde- hyde) was then read off. The outer jacket of the Hofmann apparatus was then removed, and the inner tube cooled as quickly as possible to the temperature of the room to prevent a recombi- nation of the aldehyde to form metaldehyde. The volume was again read off; this is the volume occupied by the aldehyde alone (which is, of course, a gas at this temperature and pressure), as all the metaldehyde remaining undissociated condenses at this temperature to a solid. Subtracting the second volume from the first — after having previously reduced both to the standard con- ditions of temperature and pressure, o° C. and 760 mm. — we obtain the volume occupied by the undissociated metaldehyde- vapor at 0° C. and 760 mm. pressure. From the volume of the aldehyde we can easily deduce the weight by finding the volume at 0° C. and 760 mm., and multiplying this by the weight of one cubic centimeter of aldehyde-vapor. Subtracting this weight from the weight of the metaldehyde originally taken we get the weight of the metaldehyde remaining undissociated. We thus have the weight and \.\i^ volume oi \\\^ metaldehyde-vapor remaining undis- sociated, and can readily calculate from these the specific gravity of the metaldehyde-vapor referred to either air or hydrogen. Practically, this determination of the density of the vapor of met- aldehyde is not quite so simple. Owing to the fact that it is almost impossible to get all the air and moisture out of the inner tube, even though the mercury be heated to 300° C. previous to using it, it was found necessary to heat the inner tube to the temperature of the experiment, and take readings before introducing the metal- dehyde. This enabled us to make a correction for the very small amount of gas present at the beginning of the experiment. While still at this temperature the metaldehyde is introduced,' and when this is entirely vaporized and the temperature has become constant, the necessary readings are again made. The outer jacket is then removed and similar readings taken when the gas has cooled to the temperature of the room. Corrections were introduced for the tension of mercury-vapor at the boiling-point of aniline as well as for that of the solid metaldehyde at the tem- perature of the room. The barometric height was taken by means 1 By means of a small glass capsule made from a short piece of wide glass tubing closed at one end. The powdered metaldehyde was packed into this by means of a glass rod with rounded edges. Acetic Aldehyde, Paraldehyde and Metaldehyde. 53 of a siphon-barometer located in the same room, and all mercury readings were corrected for temperature-changes. The inner tube of the Hofmann apparatus was graduated in cubic centime- ters and carefully calibrated. Readings on this tube were taken by means of a pendulum-cathetometer in the usual manner. We endeavored to determine the density of metaldehyde- vapor at temperatures different from those used by Hanriot and CEconomedes, but did not succeed on account of the fact that metaldehyde is not entirely converted into vapor until the temper- ature of the vapor of boiling aniline is reached, and higher tem- peratures cause its complete dissociation into the simple aldehyde. The formulas used in reducing the observed volumes to stand- ard conditions, and containing all the corrections above enumer- ated, are the following : v.: r p J P" , p' , v\ \_\ -f .00018/ Vi-t-.oooi8/'' "* I 4- .00018/^ *" /J 760(1 -f-a/') V P _( p , II. TT_'' |_iH-.oooi8/ ^+•00018^^'" V,: 760(1 -|- at) Formula I was used to calculate the volume z'J (very small), due to the small amount of air and moisture contained in the mercury at the beginning of the experiment, and the total volume v^l, after the introduction of the metaldehyde; formula II, to calculate the volume, z/'o", of the aldehyde resulting from the partial dissociation of the metaldehyde. v\ v" and v'" are the observed volume at the beginning, the total volume, and that of the aldehyde respectively. /=the temperature of the room. /'=ithe temperature of the aniline- vapor in the inner tube. /"=:the mean temperature = -^^-—. P=zthe barometric pressure. ^'=zthe height of the mercury column within the vapor-jacket at the temperature t'. p"=zthe height of the mercury column below the vapor-jacket at the temperature i". />=zthe height of the mercury column in the inner tube at the temperature /. j:=the tension of mercury- vapor at the temperature i'. e = the tension of metaldehyde— about 1.2 mm. between 20°- 25° C. 54 Ortidorff and White. (vH' — v'o) = the volume occupied by the aldehyde-vapor at o° C. and 760 mm. (i/J' — z^'o) X 0.00197076 = the weight of the aldehyde-vapor =zw. (v'o' — v'o")z=X, the volume occupied by the undissociated metaldehyde-vapor at 0° C. and 760 mm. lV=the weight of the metaldehyde taken. ^ = the weight of the metaldehyde remaining undissociated = JV—w. «= 0.00367. 0.00018:= the coefficient of expansion of mercury. The following are the readings' : I. II. w 0.0357 gram 0.0325 gram Pr' 739.3 mm. 745.6 mm. P. 570.5 mm. 554 mm. PI 154.5 mm. 191 mm. v' 41.3 cc. 38.6 cc. U, 19.6° 22.2° il 180.9° 180.7° n, 100.3° 101.4° s II mm. 1 1 mm. P., 739.3 mm. 745.6 mm PKn 302.1 mm. 336.6 mm. Pi, 183.0 mm. 187.9 mm. v" 85 cc. 78.2 cc. Un 19.6° 22.2° t'.„ 180.9° 180.7° €, 100.3° 101.4° P.n, 739.3 mm. 745.6 mm. P.n 555.3 mm. 591.0 mm. v'" 72.3 cc. 66.1 cc. u 19.6° 22.2° iu 21.5^ 20.6° e 1.2 mm. 1.2 mm. X 0.00535 gram. 0.00873 gram. X 0.93 cc. 1.39 cc. ' The subscripts v' v" and v"i indicate the readings for the observed volum and v'". Acetic Aldehyde, Paraldehyde and Meialdehyde. 55 Whence we deduce the vapor-density : Found. Calculated Density I. II. forCC^H^Oa. Referred to air 4.45 4.85 4.57 Referred to hydrogen 64.22 70.12 66.00 Hanriot and CEconomedes give the following figures as the results of their determinations: By the Dumas method : I. II. Calculated. Density referred to hydrogen 72.2 63.7 66.00 By the Hofmann method : I. II. III. IV. Density referred to hydrogen 59.1 63.45 59-^5 67.45 We have recalculated the results given by Hanriot and CEcon- omedes from their data and find that unless they have introduced corrections of which they make no mention, these are not quite so accurate as they appear to be. After eliminating some mis- takes, plainly typographical errors, we have deduced the following values : Dumas method: Density referred to hydrogen Hofmann method : I. II. X (grams) 0.01538 0.0085 X(cc.) 3.1 1.39 Density, Hz=i 55.38 68.20 The results of the vapor-density determinations made by Han- riot and CEconomedes as well as those made by us, it will be seen, agree with the molecular weights deduced from the depression of the freezing-point of solutions in phenol and thymol, so that we must conclude that the freshly prepared metaldehyde has the same molecular weight and the same formula as paraldehyde. Zecchini^ made an attempt to determine the molecular weight of metaldehyde by the use of the boiling-point method. He used as solvents alcohol and chloroform, and gives the following results : 1 It is impossible to recalculate the density from the figures given by H. and CE. in experi- ment No. 2, by this method, owing to the fact that the authors have plainly inserted the wrong data. ' Gazz. chim. ital. 83, 586. I. 69.6 II. III. 0.0109 2.13 57.06 , IV. 0.0075 1-55 54.00 56 Orndorff and White. I . Metaldehyde in alcohol ; molecular rise, 11.5' Concentr. Rise. Coef. of Rise. Mol. Rise. 1.2742 0.03° 0.02355 11.40 for {Ci\i,0)xx 1.3296 0.075 0.04348 11.48 " (GH.O)6 2.8035 0.06 0.03316 11.70 " (OH40)8 1-9725 0.053 0.02656 11.81 " (OH40).o //.' Metaldehyde in chloroform , • molecular rise, 36.6°. 0.9100 0.14 0.1648 36.25 for (GH40)8 1.2779 0.22 0.1721 36.86 " (O.H4 0)3 1-2957 0.16 0.1389 36.67 " (OH40> 1.7716 0.24 0.1355 36.00 " (OH40)6 From the foregoing results, Zecchini concludes that metalde- hyde has a molecular weight larger than that represented by the formula (C2H40)3, and that under certain conditions these com- plex molecules break down into simpler ones. He nowhere mentions the conditions of pressure under which his experiments were conducted. This, together with the fact that metaldehyde undoubtedly undergoes dissociation when its solutions are heated, would, we think, be sufficient to account for the discordant results which he obtained. It is difficult to understand how he could have drawn any conclusion from such results. Earlier in this paper mention was made of the fact that in mak- ing molecular-weight determinations of metaldehyde, it was necessary to use freshly crystallized metaldehyde, as metaldehyde which had stood for any length of time was found to have under- gone some sort of decomposition, forming paraldehyde as one of the products. We have investigated this decomposition very carefully and have proved that metaldehyde, far from being a very stable compound as most investigators have stated, is an exceedingly unstable substance, undergoing decomposition in some cases within twenty-four hours after it is prepared. In order to determine the course of this decomposition as well as the products formed, we have made molecular-weight deter- minations of the decomposition-product both by the cryoscopic and the vapor-density methods. Several samples of material were used for this purpose. One was a sample of metaldehyde which had been prepared two years 1 At our request, Mr. B. H. Hite, at the Johns Hopkins University, repeated this work, but could get no constant results. Acetic Aldehyde, Paraldehyde and Metaldehyde. 57 before by one of us, and had been kept in a tightly closed bottle ever since ; the other was prepared by recrystallizing some of the metaldehyde recently made by us, from chloroform, and allowing the crystallized product to stand for about two weeks. One or two individual determinations were made with recrystallized metaldehyde, which had stood from one to three days after crys- tallizing. In making these determinations we proceeded as follows : the solid metaldehyde was pulverized in an agate mortar and allowed to stand, spread out on drying-paper for a short time until the odor of paraldehyde had completely disappeared. This odor of paraldehyde was found to be characteristic of metalde- hyde which had been kept for any length of time, when it was powdered. After the product had lost the odor of paraldehyde it was made up into compressed tablets in the usual manner, and the molecular weight determined in solution in phenol and in thymol by the Beckman method. In working with phenol solu- tions care was taken to exclude all moisture in the manner already described. The results were almost identical in the several cases and were as follows : Sample of metaldehyde which had stood two years : (C2H40>=i76. Solvent, phenol ; molecular depression, 75 ; depression-coeffi- cient, 0.4261. ^ Mol.Wt. t. Solvent. Wt. Sub. Concent. Depress. Dep.-coef. Mol. Dep. found. 19.78 0.1966 0.9939 0.458 0.4608 81. 1 163 19.78 0.3160 1.5975 0.731 0.4576 80.5 164 18.61 0.0946 0.5083 0.228 0.4483 78.9 167 18.61 0.2032 I.O919 0.490 0.4487 78.9 167 18.61 0.31 15 1.6738 0.769 0.4594 80.8 163 19.65 0.0987 0.5023 0.229 0.4557 80.2 165 1965 0.2174 1. 1063 0.522 0.4718 83.0 159 19.65 0.3350 1.7048 0.810 0.4751 83.6 158 19.65 0.4584 2.3328 I. no 0.4758 83-7 158 19.65 0.6195 3.1527 1.494 0.4738 83.4 159 Metaldehyde recrystallized from chloroform and allowed to stand for two weeks : 19.00 0.0756 0.3979 0.182 0.4574 80.5 164 1900 0.1535 0.8079 0.365 0.4517 79-5 166 19.00 0.2443 1.2858 0.580 0.45 1 1 79-4 166 19.00 0.4560 2.4000 1. 104 0.4600 80.9 163 58 Orndorff and White. Metaldehyde recrystallized from chloroform and allowed to stand for two or three days : 22.40 0.1212 0.5410 0.233 0,4306 75.8 174 Solvent, thymol; molecular depression, 73.9; depression-coeffi- cient, 0.4199. Metaldehyde recrystallized from chloroform and allowed to stand for two or three days : 17.78 0.1047 0.5888 0.253 0.4296 75.6 172 It will be seen from these determinations that the solid material remaining when metaldehyde is allowed to stand has the mole- cular weight 176 and the formula (C2H40)4. It is made up of four aldehyde molecules and might be called tetraldehyde. The results in nearly every case are a little lower than the theoretical values. This is probably due to the fact that the solid material still contained some paraldehyde. That this was the case was proved by powdering some of the compressed tablets which had stood for half an hour, when the odor of paraldehyde was quite apparent. The change which metaldehyde undergoes on standing was also noticed by Troeger,' who found that when metaldehyde had been kept some time it was largely converted into paraldehyde and aldehyde. The same observation has been made by Friedel,^ also by Mr. Kortright, in this laboratory. Troeger states that the solid substance remaining was metaldehyde, and he claims to have shown this by molecular-weight determinations of the substance in solution in phenol. He gives no results, however, and he also states that the results were not quite satisfactory, probably on account of the dissociation of the metaldehyde in the warm phenol. We have shown that metaldehyde undergoes no dissociation either in phenol or thymol solution during the determination of its mole- cular weight in these solvents, and we think that Troeger's unsat- isfactory results were due to the fact that he had a mixture of the trimolecular compound, paraldehyde, and the tetramolecular com- pound (C2H40)4, formed by the spontaneous decomposition of the metaldehyde. It is quite probable that the same explanation applies in the case of the results obtained by Zecchini in determining the molecular weight of metaldehyde in solution in phenol. He states that the experiments with phenol solutions would lead to the formula 1 Ber. d. chem. Ges. 35, 3316. « Bull. Soc. chim. (Paris) 9, 384. Acetic Aldehyde, Paraldehyde and Metaldehyde. 59 (GH40)4, or, better, to a formula between (C2H40)3and (GH40)4. This is exactly what we should expect if metaldehyde which had stood for some time were used. For we have shown that such metaldehyde is a mixture of paraldehyde and the tetramolecular compound (C2H40)4. Zecchini makes no statement regarding the purity of his metaldehyde, nor did he recrystallize it. We are justified therefore in concluding that he used metaldehyde which had stood for some time, and was probably a mixture of paral- dehyde and the tetramolecular compound (C2H40)4. The results obtained from the depression of the freezing-point of solutions in phenol and thymol were then confirmed by vapor- density determinations, using the Hofmann method and apparatus. The same precautions were used as in the determination of the vapor-density of the freshly crystallized product. Metaldehyde freshly crystallized from chloroform was allowed to stand for about two weeks. It was then pulverized and kept for some time until the odor of paraldehyde was no longer perceptible. The vapor-density was then determined in the manner already described. The data and results follow : I. 11 Ill [. w 0.0400 gram 0.0434 gram 0.0368 gram p,< 737 mm. 739 mm. 738 mm. p\, 528 mm. 538 mm. 557 mm. PI 201 mm. 199 mm. 183 mm. vl 41-5 cc. 39-7 cc. 39-4 cc. t„ 20.8° 19.1° 20° t 180° 180.1° 179.7° tl 100.4° 99.6° 99.8° s II mm. II mm. II mm. p.„ 737 mm. 739.4 mm. 738 mm. PU 285.4 mm. 279 mm. 310 mm. pin 197-5 mm. 198 mm. 179-5 mm. v" 86 cc. 86.8 cc. 84.6 cc. Ur 20.8° 19.6° 20° in 180° 180.1° 179.7° i'J,' 100.4° 99.6° 99.8° P.,n 737 mm. 739-4 mm. 738 mm. P.,„ 556.5 mm. 554 mm. 557-5 mm. v^'i 72.6 cc. 72.8 cc. 72.2 cc. 'v'll 20.8° 19.6° 20° 'vlli 20° 19° 19° e 1.2 mm. 1.2 mm. 1.2 mm. X 0.00963 gram .01157 gram 0.00586 gram X 1-34 cc. 152 cc. 0.76 cc. 6o Orndorff and White. Whence we deduce the vapor-density : I. Found. II. III. Calculated Density for (CjH^O)^. 5-56 5-90 5-96 6.09 80.22 84.98 86.05 88.00 Referred to air " hydrogen These results show that metaldehyde undergoes decomposition when allowed to stand for some time, and that one of the products formed during this decomposition has the formula (C2H. 0)4. The other products are paraldehyde and a small amount of aldehyde, as Troeger and Friedel have already shown. That this change of the metaldehyde into tetraldehyde, (C2H40)4, paraldehyde, and aldehyde, takes place in the dark as well as in the light, was shown by allowing some of the freshly recrystallized metaldehyde to stand in a desiccator of brown glass in a dark room for three days. It was then taken out and pulverized. The odor of paraldehyde was noticed, as in previous cases, when the metaldehyde was not kept in the dark, and molecular-weight determinations showed that the solid product remaining was tetral- dehyde, (C2H40)4. Further investigation showed that this decom- position of the metaldehyde was increased by a rise in the temper- ature of the room. In this connection it should, perhaps, be stated that Friedel' succeeded in transforming more than half of the metal- dehyde into a liquid consisting of paraldehyde mixed with a little ordinary aldehyde, by heating some grams of metaldehyde in a sealed tube to a temperature of 6o°-65°. Friedel says nothing, however, about the solid product remaining after heating the metal- dehyde. This tetramolecular compound (C2H40)4, tetraldehyde, resem- bles metaldehyde very closely — so closely indeed that it was only discovered by making molecular-weight determinations with it in solution in phenol and in thymol. It has the same crystal form as metaldehyde, but differs from it in that the crystals are dull and opaque, while those of metaldehyde are clear and trans- parent. One may follow the change of metaldehyde into this tetramolecular compound, indeed, by noticing the change taking place in the crystals. This change takes place much more rapidly' in the product recrystallized from chloroform, than in the larger crystals formed by the action of calcium chloride on ordinary » Loc. cit. *The change is apparent in some cases (where the temperature of the room is high) within twenty-four hours, though it is by no means complete within this time. Acetic Aldehyde, Paraldehyde and Metaldehyde. 6i aldehyde. The freshly crystallized product is pulverized with difficulty ; the clear, transparent crystals, being tough and fibrous, merely pack together without becoming powdered. The tetral- dehyde, on the contrary, is easily reduced to powder. The crystals have become brittle as well as dull and opaque, on stand- ing. The solubility of the tetraldehyde in phenol and in thymol is much greater than that of metaldehyde. The tetraldehyde is soluble in warm chloroform to about the same extent as metalde- hyde. Several attempts were made to recrystallize the tetral- dehyde from chloroform, but without success, owing to the fact that the recrystallized product was found to have the same mole- cular weight and properties as the metaldehyde. It is quite prob- able that the tetraldehyde undergoes dissociation in solution in warm chloroform into the simple aldehyde, and that on cooling the hot chloroform solution in ice-water or in a freezing mixture — the method usually employed by us — metaldehyde crystallizes out, being the more stable form at this low temperature. That this is really the case seems likely, from the fact that the chloro- form solution of the tetraldehyde has a strong odor of aldehyde, and is found to contain considerable quantities of paraldehyde when the crystals have been filtered out of the solution. The chemical conduct of the tetraldehyde is practically the same as that of metaldehyde with the same reagents. It is decomposed parti- ally into the simple aldehyde when heated, and undergoes com- plete dissociation when heated in the presence of sulphuric or hydrochloric acid. Theoretical. Having thus shown that paraldehyde and metaldehyde are undoubtedly isomers as well as polymeric modifications of acetic aldehyde, and that metaldehyde is the unstable form, it yet remains to find some explanation of this peculiar kind of isomer- ism. Paraldehyde is certainly a ring-compound in which the oxygen atoms connect the various groups. The structural for- mula usually assigned to it, O O 62 Orndorff and White. is in strict accord with its formation from the simple aldehyde, its decomposition-products, and its chemical conduct in general, and may be regarded as correct. The recent work of Briihl' on the relation between the spectrometric constants and the chemical constitution of paraldehyde, points to this formula as the only one which will satisfy all the requirements. Metaldehyde has the same molecular weight and the same general formula as paralde- hyde, (C2H40)3 or CeHivOs. It is not possible to explain the isomerism existing between these two substances by any hypoth- esis which merely takes into consideration differences of structure in a plane. The plane structural formulas would indeed seem to be identical, as the chemical conduct of either isomer may be accurately represented by the same structural formula. After a complete and exhaustive study of the chemical conduct of these two substances, we have come to the conclusion that they are stereo-, or space-, isomers.^ Some of the reasons that led to this conclusion are as follows : 1. Paraldehyde and metaldehyde contain the same elements, have the same percentage composition and the same molecular weight. 2. Both paraldehyde and metaldehyde are formed from acetic aldehyde by the action of the same polymerizing agents, a low temperature (below o° C.) being favorable for the formation of metaldehyde; paraldehyde being formed under ordinary condi- tions. The unstable polymer, metaldehyde, is always formed in much smaller quantity than paraldehyde, and paraldehyde is always formed simultaneously with it. 3. The chemical conduct of both paraldehyde and metaldehyde is identical with the same reagents. Thus, with phosphorus pen- tachloride both give ethylidene chloride, while with hydrochloric acid both yield ethylidene oxychloride. Both remain unchanged when boiled with strong alkalies, and neither reacts with ammonia, hydrocyanic acid, sodium bisulphite, hydroxylamine or phenyl- hydrazine. Neither reduces Fehling's solution or an ammoniacal solution of silver nitrate. The formulas for these two substances must then show that they are identical so far as chemical conduct ' Ber. d. chem. Ges. 24, 653. - Meyer and Jacobson in their " Lehrbuch der organischen Chemie," suggest that paraldehyde and metaldehyde may be stereoisomers. Friedel makes the same suggestion in a paper (Bull. See. chim. (Paris) [3] 9,384) which appeared after the work here recorded had been completed. Acetic Aldehyde, Paraldehyde and Metaldehyde. 63 is concerned, and that they do not contain the characteristic O II aldehyde group — C — H. 4. Kekul6 and Zincke showed that metaldehyde in contact with a small amount of hydrochloric-acid gas is converted into paral- dehyde containing some aldehyde. Troeger and Friedel have both observed that in some cases this change takes place sponta- neously. We have noted above that metaldehyde can only be kept a short time owing to its liability to change into paraldehyde and the tetramolecular compound (C2H40)4, and have shown that this change is a function of the temperature. 5. Both paraldehyde and metaldehyde are decomposed when heated with the same reagents that bring about their formation from acetic aldehyde, and both yield the same product, viz., acetic aldehyde. All these facts point to the conclusion that we have here to deal with a simple case of stereo-, or geometrical, isomerism, of the same kind as that to which Baeyer referred the hexahydro- terephthalic acids. Referring now to the structural formula by which we must represent both the molecule of paraldehyde and that of metalde- hyde, it will be seen — more clearly by using models' — that it admits of two configurations in space, and only two: H»C,.. H HsC H HaC.., I J ,.QWz HsC. • 1 ^CH3 the one corresponding to the cis or maleinoid form, the other to the cistrans, or fumaroid form. A careful examination of the models will show that these two forms are the only ones possible. According to this view of the structure of metaldehyde and paraldehyde, the formation of these isomers from acetic aldehyde may be represented graphically as follows : ' The Kekule models are well adapted for this purpose. ^ The broken lines represent the groups below and the unbroken lines those above tl.e plane of the paper. 64 Orndorff and White. CHs H V CH3 'I - // ^ o Three molecules of aldehyde uniting to form either H3C H H3C H HsC A I CHs HsC L I .XH3 Cis form. Cistrans form. The next problem was to determine which of these stereo- or space-formulas corresponds to paraldehyde and which tometalde- hyde. Two facts may be made use of to aid in solving this question : {a) the relative stability of the paraldehyde and metaldehyde, and (J)) the relative amounts of paraldehyde and metaldehyde formed when acetic aldehyde undergoes polymeriza- tion. We should expect the cistrans form to be more stable than the cis form, if we reason from the conclusions derived from the study of other stereo-isomers, and if we take into consideration that in the cistrans form like groups or atoms which would have little or no attraction for each other, are represented as far away from each other as possible, while unlike groups, such as methyl and hydrogen — which would certainly exert some influence on each other — are represented as nearer to each other. The reverse is true in the cis formula. Here all the hydrogen atoms are repre- sented on one side of the molecule and all the methyl groups on the other. The hydrogen atoms would exert no attraction on each other nor would the methyl groups influence the other methyl groups. Like groups or atoms are in other words nearer to each other in the cis form than in the cistrans form, and unlike groups or atoms are farther apart, so that this form would naturally be the unstable one and would tend to pass into the other, pro- vided the conditions for the change are favorable. It has been demonstrated in the present paper that metaldehyde is the unstable polymeric modification of acetic aldehyde, and that it tends to pass Acetic Aldehyde, Paraldehyde and Metaldehyde. 65 into paraldehyde, particularly in the presence of a halogen acid or when the temperature is raised. We therefore assign the cis formula to the unstable modification, metaldehyde, and thecistrans formula to paraldeh)'de, the more stable modification. The second consideration is based on the theory indicating the relative amounts of the cis and the cistrans modifications that would be formed, provided the double bond between the carbon and oxygen atoms does not break preferably on the one side or other of the aldehyde molecule. It will be seen by refer- ring to the models and to the graphical representation of the polymerization of aldehyde that the three molecules of aldehyde can give but two forms of the trimolecular compound, the cis and the cistrans. Each molecule of the aldehyde may combine with the other two, and as each molecule in the nascent condition has two bonds or free affinities, the total number of ways in which combination may take placewill therefore be eight (2X 2 X 2=8). Of these eight combinations six will be of the cistrans form, but only two of the cis form. The proportion of the cistrans modifi- cation and the cis modification formed by the polymerization of aldehyde should then be, so far as we can see, three parts of the cistrans form to one of the cis. No actual measurements of the exact amount of paraldehyde and metaldehyde formed in the polymerization of aldehyde have ever been made, but every one who has made metaldehyde knows that paraldehyde is always formed simultaneously with the metaldehyde and in much larger quantity. Kekul6 and Zincke and many other chemists mention this fact which we have frequently confirmed. To paraldehyde therefore we give the cistrans formula, and to metaldehyde the cis formula. The heats of combustion of paraldehyde and of metaldehyde have been determined by Louguinine," and have been found to be nearly the same at constant pressure. The figures actually obtained were 813200 calories for paraldehyde and 805790 calories for metaldehyde. This difference in the heats of combustion of these two isomers is attributed by Louguinine to differences in the physical conditions of the two substances, but it seems more likely that it is due to the fact that paraldehyde is the more stable modifi- cation of the two, and this being the case it would probably, though not necessarily, give out more heat on combustion. The experi- > Compt. rend. 108, 620. Vol. XVI.-s. 66 Orndorff and White. ment of Friedel, moreover, shows that to pass from the metalde- hyde condition to that found in paraldehyde it is necessary to add energy in the form of heat. The study of the heats of combustion of paraldehyde and metaldehyde, then, show that paraldehyde is the stable form, while metaldehyde is the unstable form. Perhaps it may be well to mention here some evidence confirma- tory of this view of the relation existing between paraldehyde and metaldehyde. If these two substances are really stereo-isomers, then it will be seen by referring to the models that when the methyl groups of paraldehyde or metaldehyde are replaced by hydrogen there should be but one substance ; there is no longer any possibility of stereo-isomerism. This means that formic alde- hyde should give but one polymeric modification, having the for- mula (CH20)3. Of the polymeric modifications that have been described, paraformic aldehyde and a-trioxymethylene, only one, a-trioxymethylene, is a trialdehyde. In its properties both physical and chemical, it resembles paraldehyde very closely. The other polymer, paraformic aldehyde, will probably be found to have a greater molecular weight than the formula (CHsO)3 indicates, since it is insoluble and amorphous and does not have a constant melting-point. We may therefore state that the theory indicates the existence of only one triformic aldehyde, and that but one is known, a-trioxymethylene. Again, if we replace the methyl groups in paraldehyde and metaldehyde by ethyl groups, we ought to have two stereo-isomers. A compound in which the methyl groups of paraldehyde and metaldehyde are replaced by ethyl groups, we should expect to result from the polymerization of propionic aldehyde, CsHsCHO. Two polymeric modifications of propionic aldehyde are indeed known, and so closely do they resemble paraldehyde and metal- dehyde in their physical and chemical properties, that it is only by analysis or a molecular-weight determination that they can be distinguished. The facts and the theory are here again in accord. There are many other aldehydes which polymerize quite readily, and yield two varieties or forms of polymers. Polymerization indeed appears to be a characteristic property of the aldehydes. That these polymers, the trialdehydes, are related to each other in the same way that paraldehyde and metaldehyde are, there seems but little reason to doubt, but it will require further work to establish this hypothesis. Acetic Aldehyde, Paraldehyde and Metaldehyde. 67 The polymeric thioaldehydes have been very thoroughly studied by Baumann and Fromm. These chemists have shown that tri- thioformic aldehyde exists in only one form, while the polymeric modifications of thioacetic aldehyde, trithioacetic aldehyde, exist in two stereo forms corresponding to the cis and the cistrans forms. The cis modification agrees with metaldehyde in every respect, even to the transformation into the cistrans modification by heating, or in the presence of strong hydrochloric acid. Acetone contains the carbonyl group, and in most of its prop- erties shows a striking analogy to the aldehydes. We should expect it to polymerize and yield but one triacetone (CsHeOs. Unfortunately it has not been found possible to make a polymeric modification of acetone, though many attempts have been made, owing probably to the ease with which it undergoes condensation in the presence of polymerizing-agents. Baumann and Fromm have, however, prepared a polymeric modification of thioacetone having the formula (C3H6S)3, and have shown that it exists in but one form. By consulting the models it will be seen that stereo- isomerism is impossible here. The recent work of B6hal and Choay' on chloralimide and iso- chloralimide also confirms the view that paraldehyde and metal- dehyde are stereo-isomers. It will be seen from the formula of chloralimide, CCl3.CH=:NH, that, just as chloral itself and ordi- nary aldehyde contain the double union between the carbon and oxygen atoms, it contains the double union between the carbon and nitrogen atoms. From what we have 'learned of the poly- merization of aldehyde, we should expect three molecules of this chloralimide to combine to form a trichloralimide, and this tri- chloralimide ought to exist in two forms corresponding to the formulas : CI3C H CI3C H V V c c HN NH HN NH CUC^ I i^CClr. ClsC^L I .,CCU H"' ^ \/^*^-H H-'- ^ \j^Yi NH NH Cisform. Cistrans form. • Ann. chim. phys. [6] 26, 34. 68 Orndorff and White. B6hal and Choay have shown that two forms exist and that one corresponds to the cis compound, the other to the cistrans, and that one may be converted into the other. It is of interest to note that this polymerization of the aldehydes is in some way connected with the peculiar condition existing between the carbon and oxygen atoms, a condition which may be termed the unsaturated condition. In the case of the unsaturated hydrocarbons, with the exception of acetylene and ethylene, polymerization takes place very readily in the presence of sulphuric acid, thus allylene gives mesitylene, and dimethylacety- lene hexamethylbenzene. Propylene, isobutylene and amylene all polymerize in the presence of sulphuric acid. This tendency towards polymerization is particularly noticeable also in the case of cyanogen compounds. In the case of the formation of di-, tri- and tetramylene from amylene in the presence of sulphuric acid, it seems very likely that an addition-product is first formed and that this then breaks up into amylene and sulphuric acid. The amylene molecules being, as it were, in the nascent condition, unite to form the di-, tri-, and tetramylenes, as they are called. It seems quite likely that a similar process takes place here when polymeric modifications of the aldehydes are formed. Thus, in the case of acetic alde- hyde and hydrochloric-acid gas, we may assume that an addition- product is first formed according to the equation : CH.CHO + HCl = CH=CH<^j^. This hydrochloric-acid addition-product, being exceedingly unstable (as we know from the work of Hanriot and (Econo- medes), loses hydrochloric acid, and three molecules of the nascent aldehyde thus formed combine to form either paralde- hyde or metaldehyde, according to the conditions of the experi- ment. Adopting this view, it is easy to see why a few bubbles of hydrochloric-acid gas can convert a large quantity of aldehyde into paraldehyde and metaldehyde, for the hydrochloric acid is constantly regenerated. There would seem to be no good reason why dialdehydes and tetraldehydes should not be produced also by a similar method of formation. The dialdehyde would have the formulas : Acetic Aldehyde, Paraldehyde and Meialdehyde. 69 HaC H H CHs C C V V c c HsC H CHs H Cis form. Trans form. and would exist in the two forms, cis and trans. A dialdehyde, (C2H40)2, is indeed known, but whether it has one of the above formulas or not cannot now be stated. The tetraldehyde has been shown to exist also, and has in all probability a structural formula similar to that of paraldehyde, viz., CH3 H HsC V M>C-0-C— O ^ I I O— C— O-C— CHa A I H3C H H This formula will explain all the facts now known about the tetraldehyde. Summary. Some of the results obtained in this work may be summed up as follows : 1. Paraldehyde and metaldehyde have the same molecular weight, and from their chemical conduct 2. Both must be represented by the same structural formula in a plane, but 3. The difference between them may be readily understood if they are regarded as stereo-isomers : 4. Paraldehyde, the more stable modification, being the cistrans, while metaldehyde is the cis variety. Further work to determine the correctness of this theory of the relation of the polymeric aldehydes is now being carried on in this laboratory. Cornell University, Ithaca, N. Y., June, 1893. 70 Lengfeld and Stieglitz. Contributions from the Kent Chemical Laboratory of the University of Chicago. THE ACTION OF PHOSPHORUS PENTACHLORIDE ON URETHANES. By Felix Lengfeld and Julius Stieglitz. In the course of our work on nitrogen halogen compounds we found that urethanes are easily obtained by the action of alco- holates on amide bromides (R.CO.NHBr)/ and it seemed desira- ble to convert urethanes, prepared by this method, into the corresponding ureas. Attempts to prepare unsymmetrical ureas by substituting amidogen for alkoxyl in NHRCO-OR — making use of the direct action of ammonia on urethanes — have shown that at low temperatures there is no action, whereas at higher both the OR and NHR groups are replaced and urea results. The usual method of converting urethane into the amine and treating with potassium cyanate is not applicable to those urethanes whose amines cannot, on account of their instability, be obtained from them. Hofmann'' showed that urethanes may be converted into isocyanates by heating with phosphorus pentoxide; and isocyanates unite with ammonia to form unsymmetrical ureas. We found that a mixture of phenylurethane and • phosphorus pentoxide begins to react at 90°-95° and colors already at this temperature; it therefore seemed probable that the substitution of phosphorus pentachloride would enable us to work at lower temperatures and would avoid the formation of isocyanates which polymerize easily. The reaction promised to be interesting also as leading to a new method for preparing chlorformamides, R.NH.CO.CI,orchlorformimido-ethers, R.N:C.MI ; (C6H5.S03M)5.MI.I; (C6H5.SOsM)5.MI.L; (CeHo.SOsIVDs.MI.Ia; (C6H5.SO..M)5.MI.L; (C6H..S03M)5.MLl6; etc., etc.; and that the potassium and sodium compounds which have already been described really represent two different members of such a series. In this connection it might not be amiss perhaps to call attention to the fact that among the periodides of certain of the alkaloid bases — the only compounds, so far as we have been able to discover, which in any way bear a relation to the compounds under investigation — such a series has been found to exist. Jorgensen, who has made a careful study of this whole subject, distinguishes seven compounds of quinine, sulphuric acid, and iodine, four of which belong to the herapathite type. Letting Q stand for the molecule of quinine, OoH:4N202, the series stands as follows : 4Q.3H2SO4.2HI.I4 4Q.3H2SO4.2HI.I5 4Q.3H.SO4.2HI.l6 4Q.3H2SO4.2HI.I7 Barium Benzenesulphoyi-periodidc , [(C6H5S03)2Ba]5.Bal2.Iio. — This compound was obtained by a method exactly similar to those already given under the description of the potassium and sodium salts. Asaturated solution of barium benzenesulphonate was added to one of the required quantity of iodine and enough barium iodide to hold it in solution, when the barium salt was precipitated as a crystalline, bronze-green precipitate. This was then got into solution in its own mother-liquor by the addition of more water and the solution allowed to stand. After a time the substance made its appearance in the liquid in the form of fine needle- shaped crystals, having the characteristic greenish-bronze color and luster by reflected light. Up to the present the substance has been obtained only in small quantity, and hence has not been studied as closely as either of the substances described above ; a barium determination was made, however, with results which agree fairly well with the formula above given : Calculated for Found. [(CsH5S03).j.Ba]6.Bal5.1,o. Ba 20.55 20.99 122 Kastle and Hill. The attempt was also made to prepare the corresponding calcium salt. The constituents of the salt were brought together in water and the solution placed under a desiccator. Alter stand- ing for a time, a brownish-red substance made its appearance, altogether unlike any of the substances obtained in this investiga- tion. The substance was evidently impure and was not further examined. The dish was then allowed to remain under the desic- cator until all of the water had been absorbed, and there was left behind in the bottom of the dish a crystalline residue having the general appearance of the salts already described ; it was so admixed with impurities, however, that no analysis of the com- pound was attempted. As stated in the opening paragraphs of this paper, this study was first undertaken with the view to determine the action of the sulphonic acids upon potassium iodide, and while we have been diverted from our original- purpose by the study of this new and interesting class of derivatives obtained at the outset, it has not been lost sight of. From the facts at present at hand it would seem that the benzenesulphonic acid acts upon the potassium iodide with the production of potassium benzenesulphonate and hydriodic acid, and that then the hydriodic acid reduces a part of the benzenesulphonic acid to benzenesulphinic acid, with libera- tion of iodide, which in turn combines with some of the potas- sium benzenesulphonate and potassium iodide to form the com- pound which for lack of a better name we have called potassium benzenesulphon-periodide. It will be remembered that in the original experiment the benzenesulphonic acid and potassium iodide were brought together in aqueous solution, and it can easily be conceived that possibly other and highly interesting reactions may take place in more dilute or more concentrated solutions, or especially if the pure, dry acid be heated with the potassium iodide. It is our purpose to make a more complete study of this reac- tion than has been possible thus far. Other sulphonic acids are even now in the course of preparation, with the view to making iodine derivatives similar to those described above, in the hope that among a great number of such compounds some few at least will be found which will throw more light upon the nature of the whole class. State Collegk of Kentucky, Lexington, Janunry i, 1894. Amidophosphoric Acids. 123 Contribution from the Kent Chemical Laboratory of the University of Chicago. ON DIAMIDOORTHOPHOSPHORIC AND DIAMIDO- TRIHYDROXYLPHOSPHORIC ACIDS. By H. N. Stokes. The highest hydrate of phosphorus pentoxide, pentahydroxyl- phosphoric acid, P(OH>, as is well known, has not yet been obtained, and it is extremely questionable whether it is capable of existence under ordinary conditions. Although no salts or esters of this acid have been prepared, the probability of their existence is much greater, it being a general rule that salts and esters of polyhydroxyl acids are more stable than the acids themselves, if the hydroxyls are united to the same atom. A few basic phos- phates are known, it is true, but these are to be regarded as true basic salts, not as derivatives of pentabasic phosphoric acid. The object of this article is to describe two amidophosphoric acids, one of which is to be referred to the common orthophosphoric acid, and which may be named diamidoorihophosphoric acid, the other derivable from the hypothetical pentabasic acid, and which may be designated as diamidotrihydroxylphosphoric acid. The relation of these acids is shown in the following formulas : PO(OH)., PO^g^=>; P(OH)5, P[oH)^ Like orthophosphoric acid, diamidoorthophosphoricacid exists in the free state. The free diamidotrihydroxyl acid is apparently incapable of existence, passing at once on liberation into the ortho compound. A well characterized series of salts has been prepared, however, which indicate that it is a pentabasic acid, the amide hydrogen being in part replaceable by metal. In a recent article' I have described the preparation and prop- erties of monamidoorthophosphoric acid. By an analogous process the diamido acid is obtained. Phenyldichlorphosphate^ is converted by aqueous ammonia into phenyldiamidophosphate, POCNH^XOCeHs), a highly crystalline and stable body, which, on saponifying with alkali or baryta and acidifying with acetic 1 This Journal 15, 198. SI use the term phenyldichlorphosphate for PO.ClalGCeHj), in preference to the usual name phenylphosphoric dichloride, as being analogous to phenyldiamidophosphate . This, on the older system, should be named phenylphosphoric diamide. 1 24 Siokes. acid, yields the free acid, which has the composition represented by PN^OaHi. The direct product of saponification is unquestion- ably a salt of the trihydroxyl acid; from this, by loss of water, the free acid is formed. Its actual constitution may be represented by two formulas, according to the manner in which this occurs; Which of these corresponds to the acid obtained, or whether both forms can exist, cannot be decided with any certainty from present data. As the action of nitrous acid fails to indicate difference of function of the two nitrogen atoms, and in the absence of any experimental evidence of the contrary, I prefer to adopt (a) provisionally. One or two salts must, however, be referred to an acid having the constitution (b). The properties of the acid may be here stated briefly, leaving a fuller description to the second or experimental part of the article. It is crystalline, and is more stable than monamidophosphoric acid; it has a sour taste and decomposes carbonates in the cold, forming primary salts. The salts of the alkalies and earths are extremely soluble and devoid of crystallizing power ; the silver salts are the most characteristic; it forms salts with one, two, three, four and five atoms of silver, and double salts with silver and alkali metals. Some of these have characteristic colors, and with increasing amounts of silver show an increasing tendency to deflagrate or explode, that with five atoms detonating violently on gentle friction. Many of them are diamidotrihydroxylphosphates ; those which are not (including alkali salts), add bases directly, forming salts of this acid. Primary silver diamidophosphate, for example, does not give silver oxide when treated with an excess of potassium hydroxide ; the potash is added directly, forming a 1 In order to distinguish possible isomers of these bodies, I propose to designate those con- taining the phosphoryl group, PC, as phosphoric compounds, and those in which the group P(NH) occurs as phosphimic compounds. The three forms derivable from orthophosphoric acid by substituting amide for hydroxyl, and their three isomers would then be : PC I ^i^\ monamidophosphoric acid; P(NH)(0H)3, phosphimic acid. PC I [^^«^» diamidophosphoric acid; P(NH) | ^^^ amidophosphimic acid. P0(NHj)3, phosphoryl triamide; P(NH) | '^^2)11 diamidophosphimic acid. As this nomenclature cannot be rigidly adhered to in the absence of more definite evidence as to the nature of these bodies, where there is no presumption in favor of one formula rather than the other, the terms amidophosphoric.etc, will be used. Amidophosphoric Acids. 125 colorless gelatinous salt, which is undoubtedly POAg ; from this, (OK> by transformations to be described later, the following series is obtained : • (NH.> (NHO-^ (NHAg)2 (NHAg). P(OAg)., P(OAg)a, P(OAg>, P(OAg)3 , OK OH White. Yellow. Red-brown. Dark brown. as well as several other bodies of uncertain composition. A secondary silver salt has also been obtained, which, in the dry state, is apparently P(NH)(NH0(OAg)2, but which, in freshly (NH2). precipitated and moist condition, acts like P(OAg)2, a derivative OH of the trihydroxyl acid. These salts are readily converted into each other or back into the original substance. Diamidotrihydroxylphosphoric acid is therefore pentabasic, the amide hydrogen being readily replaceable by metal, the com- pounds thus formed showing no tendency whatever to decompose into oxide and salts of diamidoorthophosphoric acid. The forma- tion of the salt PN2H403KAg2 in thirty per cent, caustic potash solution seems to indicate, however, that the amide hydrogen is not replaceable by alkali metals. Diamidotrihydroxylphosphoric acid may be regarded as ortho- phosphoric acid in which the oxygen of the phosphoryl group is replaced by two amido groups : = P(OH>; (NH2)2 = P(OH)=. The analogy appears in the yellow tertiary silver salt, which was at first mistaken for silver phosphate : O = P(0 Ag).. ; (NH2)2 z= P(0 Ag)=, the yellow color being possibly due to the common chromophore = P(OAg)3. A further study of monamidophosphoric acid, now being made, shows that besides the white primary and secondary salts described," others of yellow, red and brown color exist, which may possibly be referred to a monamidotetrahydroxylphosphoric acid. ' This Journal 15, 211. 1 26 Stokes. Phosphoryltriamide, if it can be obtained, will probably act in a similar manner, existing not only as PO(NH2)3, but in salts as triamidodihydroxylphosphoric acid, ^rOHV ' Experimental Part. Phenyldiamidophosphate, PO Qp^tj • — ^ vaoA. wt. phenol is boiled with somewhat more than i mol. wt. phosphorus oxy- chloride, until the evolution of hydrochloric acid ceases. The phenyldichlorphosphate, PO.CI2.OC6H5,' is approximately puri- fied by two distillations, the portions passing over between 200°- 300° being reserved. The product is poured, drop by drop, with constant stirring, into ammonia water of about sp. gr. 0.9, which is kept cool in ice,'' 400 cc. ammonia being taken for 100 grams dichloride. The reaction is accompanied by a crackling noise, and separation of phenyldiamidophosphate as a pure white gran- ular solid. This is readily filtered off by suction, washed a few times with cold water, dried on clay plates and recrystallized twice from a large volume of boiling 95-per cent, alcohol. The ammo- niacal mother-liquor retains a small portion in solution, which cannot be recovered by evaporation. The yield of crude diamide is 45-55 per cent, of the theoretical, calculated from the amount of dichloride, or 30-35 per cent, of the amount theoretically obtain- able from the phosphorus oxychloride used, a portion of this having been lost in the form of unchanged substance or of phe- nylphosphates. The rest of the dichloride is probably converted mostly into phenylphosphoric acid, but a small quantity of phenyl- amidophosphoric acid, PO.NHa.OH.OCeHs' is formed, which may be separated by removing most of the ammonia by a strong current of air, neutralizing and precipitating by cupric sulphate. On decomposing the copper salt with sodium sulphide, saponify- ing the filtrate with caustic soda and adding acetic acid, the diffi- cultly soluble primary sodium amidophosphate* is obtained, recognizable by its crystalline form and by conversion into other characteristic salts. The yield is not great. Phenyldiamidophosphate is permanent in the air, and loses nothing at 100°. The analysis of two preparations gave : 1 G. Jacobsen : Ber. d. chem. Ges. 8, 1521. Boiling-point 24i°-243°. 2 Cooling with ice and salt does not increase the yield. sThis Journal 15, 201. *Ibid. 15, 204. * Amidophosphoric Acids. 127 Calculated for Found. PO.(NHj)2.0C,H5. (i). (2). P 18.04 18.31 18.13 N 16.32 16.08 15.92 16,18 (i). P: N= I : 1.94 and 1.92. (2). P:N= 1:1.97. Phosphorus was determined as magnesium pyrophosphate, after repeated evaporation with caustic soda and hydrochloric acid ; nitrogen as ammonia, after boiling a quarter of an hour with dilute hydrochloric acid. Phenyldiamidophosphate crystallizes from alcohol in brilliant colorless rhombic scales, or plates which seem to be monoclinic, and which in masses much resemble naphthalene. Cold water dis- solves it but slightly, although much more readily than diphenyl- monamidophosphate ; it may be recrystallized unchanged from boiling water. It dissolves in about 100 parts cold and 10 parts boiling 95-per cent, alcohol; other neutral solvents dissolve it but very slightly. It fuses, but not sharply, at i85°-i90°, undergoing partial decomposition, indicated by evolution of ammonia and falling of the melting-point ; at a higher tempera- ture it gives off ammonia and phenol, and leaves a white phos- phatic residue. Dilute mineral acids dissolve it very slowly in the cold, readily on boiling, decomposing it into ammonia and phenylphosphoric acid; dilute and concentrated acetic acid, hot or cold, act as solvents merely, the glacial acid dissolving it quite readily. Caustic potash, soda, ammonia and baryta saponify it readily on boiling, giving salts of diamidophosphoric, or more probably, diamidotrihydroxylphosphoric acid ; it is but slowly soluble in cold dilute alkali. The saponification is accompanied with scarcely any evolution of ammonia, unless in the case of extremely concentrated boiling solutions, when it may be regarded almost as a fusion. This is not to be reconciled with the opinion of Mente' that the non-appearance of ammonia is proof that the nitrogen in these compounds exists in the form of imido groups. The acid under consideration must contain at least one amido group, yet it is scarcely decomposed by alkali. Action of Dry Ammonia on Phenyldichlor phosphate. Hoping to obtain a better yield of phenyldiamidophosphate, phenyldichlorphosphate was subjected to the action of dry ammo- nia gas. The following reactions were expected : 'Ann. Chem. (Liebig) 348, 240, 247, Vol. XVI. -9. 1 28 Slokes. ^, NH. (i). PO>,'iTT +2NHs=P0Cl +NH.C1. ^"^'^^ OCeHa (^)- P08ceH.+4NH3=P0[N^>+2NH.Cl. Calculated increase of weight in (i) 16.1 per cent., in (2) 32.2 per cent. The temperature was at first kept down by immersing the flask in cold water. There resulted a semi-fluid mass with an increased weight of 14.7 per cent., which remained nearly unchanged after prolonged action at 100°. The product, covered with strong aqueous ammonia, became warm, but there was no strong reaction ; it became thick and pasty, but did not solidify even after pro- longed action. No diamide could be obtained from this by extraction with hot alcohol; had it been formed, it would certainly have been detected. It might be concluded from the increase of weight observed that the dry ammonia had replaced only one-half of the chlorine by amide; this conclusion, however, is unjustifiable. Had the reaction taken place as in equation (i), aqueous ammonia would have generated some diamide. Moreover, other reactions could occur, requiring the same increase as equation (i), for example : POQ^^j^^).NH.(PO.OCeH5,).NH.(POQ^^j^J-f4NH4Cl, where the ratio of chloride to ammonia is also i : 2, yet two-thirds of the chlorine have been acted on. It is highly probable that the action of dry ammonia occurs in a sense analogous to equation (3), a mixture of more or less complex imidophosphates being produced. The reaction was not studied further. Saponification of Phenyldiamidophosphate. Diamidoorthophosphoric Acid, PO qu • — On placing two or three molecular weights solid caustic potash on one molecular weight ether and adding a little cold water, the heat of solution of the alkali is sufficient to start the reaction, and the ether dissolves completely. The solution is boiled five minutes to make sure of complete decomposition. It is noteworthy that if Arddophosphoric Acids. 129 less than three molecular weights potash be used much phenol is liberated, but if the amount of potash equal three molecular weights all phenol is held in combination as phenylate ; an indi- cation that the salt in solution contains two atoms of potassium. The solution is cooled in ice and a little more than the calculated quantity of strong acetic acid added gradually, when the diamido- phosphoric acid crystallizes out at once as a coarse white powder, which, under the microscope, is seen to consist of six-sided plates, united to pairs by overlapping or fusign at the angles. More may be obtained by adding tive or six volumes of alcohol to the mother-liquor. It is then sucked out and washed with alcohol. It may be redissolved in water and reprecipitated by alcohol, but the analyses indicate that this does not result in a purer product. When precipitated by alcohol it usually forms X -shaped skeleton plates, with parallel striations in the direction of the longer axis, or frequently large thick X -shaped forms with dentate edges; other forms were occasionally observed. The yield is about 65 per cent, of the theoretical, the rest remaining in solution in the dilute alcohol, and separating but very slowly and imperfectly in impure form. Whether prepared as above or by decomposition of the silver, salt with sulphuretted hydrogen, it always evolves a small amount of ammonia on treatment with alkali in the cold, which does not increase on boiling, and hence is not due to direct decomposition, but probably to the presence of a little ammonium phosphate or monamidophosphate. This is indicated in the analyses, in which the phosphorus and nitrogen, but especially the latter, fall some- what too low. All analyses were made with substance dried in vacuo over sulphuric acid. Calculated for Found. PO.(NH2)j.OH. (I). (2). P 32.28 32.15 N 29.22 27.66, 27.66 28.10, 28.16, 28.10 Found. (3«)- yzl')- (3^i- (4). p 31.70 31.67 32.17 32.16 N 28.13 27.56 27.93 27.70 (i). P:N= 1:1.90. From acetate solution by alcohol, dis- solved in water and reprecipitated by alcohol. (2). Prepared like (i). 1 30 Stokes. (3a). P : N= I : 1.96. Direct crystallization from acetate solu- tion, without alcohol. (3(5). P : N = I : 1.97. From mother-liquor of (3a) by alcohol. (3<:). P:N::= 1:1.92. (3a:) dissolved and reprecipitated by alcohol. (4). P : N=: I : 1.90. From primary silver salt by sulphuretted hydrogen ; precipitated by alcohol. For analysis, the substance may be decomposed by boiling for ten or fifteen minutes with dilute hydrochloric or sulphuric acid; longer than this is useless, as the decomposition takes place easily. Concentrated sulphuric acid must not be used, as the action is violent. The decomposition of the primary silver salt by sulphuretted hydrogen and precipitation by alcohol gives a good yield, and the acid shows the characteristic form. Diamidoorthophosphcric acid is moderately soluble in cold water, and nearly insoluble in alcohol. It is permanent in the air, and loses nothing over sulphuric acid. At 100° it slowly fuses, undergoing a slight increase of weight, probably due to absorp- tion of moisture, and is converted into ammonium salts of unknown nature, as it gives a copious white precipitate with silver nitrate, which is without action on the acid itself. At a higher temperature it gives off ammonia, and finally volatilizes without residue, attacking platinum. It is much more stable than monamido- phosphoric acid, and apparently maybe kept indefinitely without change. It also differs from the latter, which has a sweetish taste, in being decidedly sour. Its reaction towards litmus and methyl orange is acid, but no sharp end-reaction can be obtained with either indicator. It readily decomposes insoluble carbonates in the cold, apparently forming primary salts only. Boiling with moder- ately concentrated alkali liberates ammonia but very slowly after the small quantity due to impurity has been expelled. A weighed portion was boiled an hour with an excess of baryta, the ammonia being collected and determined ; it corresponded to a decomposition into phosphoric acid of 4.5 per cent., but part of this was due to original contamination. This is in marked dis- tinction from monamidophosphoric acid, which is instantly decom- posed on boiling with excess of baryta,' although almost unacted on by boiling alkali." Even the primary salts are quite slowly IThis Journal 15, 206, 207. '^Ibid. 1.5. 203. Amidophosphoric Acids. 131 decomposed on boiling, and sometimes separate in nearly pure form from boiling solutions. The free acid, on the contrary, is rapidly destroyed by boiling water, with evolution of ammonia, but passes through at least two intermediate stages before being converted into phosphoric acid. Boiled for less than a minute and then cooled, the solution gives no precipitate with barium chloride, but a white curdy precipitate with silver nitrate, which melts under hot water to a transparent liquid, solidifying on cooling to a hard and often transparent glass, suggesting silver hexameta- phosphate ; it is not this, however, but the salt of a nitrogenous acid. If boiled for a minute or two, both barium chloride and silver nitrate give white precipitates, the latter turning yellow under boiling water, but not melting. These products have not been further examined. On longer boiling it gives ammonium phosphate. Sapo7iification by Ammonia. — The action of ammonia on the phenyl ether differs from that of the fixed alkalies. Diamidophos- phoric acid is formed at first, but owing to the instability of the ammonium salts, these act more like the free acid or primary salts on boiling. After short boiling, the acid is formed which gives the easily fusible silver salt just referred to; on longer boil- ing (about an hour) this acid changes to another, the silver salt of which is white, amorphous, and infusible under boiling water ; at the same time phosphoric acid is formed. On precipitating with silver nitrate, dissolving in ammonia and precipitating in fractions by nitric acid, a colorless salt was obtained, which gave on analysis P, 11. 51; N, 3.87 ; Ag, 66.25, or P: N : Ag = 4:2.94:6.53. The substance was obviously a mixture, but the figures suggest a possible imido acid, having the same relation to tetraphosphoric acid as pyrimidophosphoric^ does to pyrophos- phoric acid. Saponification by Baryta. — The action of baryta on diphenyl- monoamidophosphate is to saponify readily one phenoxyl group only -^ the second is attacked but slowly, and as fast as this occurs the barium amidophosphate decomposes. Markedly different is its behavior towards the diamido ether. One part ether is boiled for ten or fifteen minutes with 2 parts crystallized baryta and 10 parts water. There occurs a very slight evolution of ammonia and formation of a precipitate, but unless the boiling be long con- ' This Journal 15, 212. '^ Ibid. ■202. 132 stokes. tinued the decomposition is trifling. On cooling and treating with carbon dioxide, the primary sah remains in solution, and this affords the best starting material for preparing the silver salts below mentioned. Action of Nitrons Acid on Diamidophosphoric Acid. The action of nitrous acid on diamidophosphoric acid was studied, hoping to obtain some indication as to whether it has the constitution expressed by PO qtt or by PNH.qttn • Primary diamidophosphates are without action on sodium nitrite, but the free diamidophosphoric acid decomposes it readily, nitrogen being given off in the cold, rapidly from concentrated, slowly from dilute solutions, and primary sodium monamido- phosphate, which is difficultly soluble, crystallizes out. As the yield of the latter is considerably below the theoretical, it would seem that the liberated nitrous acid does not attack only one amido (or imido) group by preference, but that it acts indiscrimi- nately, a mixture of primary phosphate, mono- and diamido- phosphate resulting. These primary salts are not sufficiently acid to decompose any excess of sodium nitrite, and the reaction ceases here. If, however, a little acetic acid (which does not decompose the amidophosphoric acids) be added, more nitrous acid is liberated and decomposition to phosphate is complete. The latter is recognized, and found to be free from amidophos- phates, by diluting largely and precipitating with ammonia and silver nitrate. There being no apparent difference in the ease with which the two nitrogen atoms are removed, there is no evi- dence in this direction of their having different functions. The difficultly soluble sodium salt was purified by dissolving in caustic soda and precipitating by acetic acid, and then dis- solved in ammonia together with two molecular weights silver nitrate, and precipitated by nitric acid. A white salt was obtained which gave on analysis figures corresponding to secondary silver niona nidophosphate :' or P : Ag Calculated for PO(NH„)(OAg)2. Found. p 9.98 10.10 Ag 69.40 68.95 :i.96. ' This Journal 15, 21 Amidophosphoric Acids. 133 A peculiar difference was observed between the monamido- phosphates made in the usual way (a series) and those prepared as above from diamidophosphoric acid (/? series), and which was at first supposed to be due to their being derived from the isomeric forms PO.NH2(OH)2 and PNH(OH)=. Primary sodium a-amidophosphate' crystallizes in unusually well formed hex- agonal pyramids or prisms ; the /5-salt sometimes forms oblongs with rounded ends, sometimes circular disks, sometimes mulberry shapes, but always devoid of sharp angles even after several recrystallizations ; both salts have apparently the same solubility. The secondary a-silver salt forms rhombic plates, or prisms ; the /S-salt, spherical aggregates of innumerable extremely fine needles, distinguishable only under a high power, and this form persists after recrystallization. In the case of several other salts the a- and /3-compounds were of different crystalline habit, in others they were identical, though with a tendency on the part of the /3-series to form smaller and less distinct crystals. The action of boiling water was identical. Equal weights of «- and /5-sodium salts were then dissolved together in caustic soda and precipitated by an excess of acetic acid. If different, they would probably crystallize separately ; if identical, the difference being due to impurity, an intermediate product should be obtained. The result was a mixture of the extreme forms, with transitional shapes in every possible stage. Finally, the a-salts were found to be quite as readily attacked by nitrous acid as the /5-salts. There seems, therefore, to be no positive evidence in these experiments of any essential difference. Salts of Diamidophosphoric Acid. Owing to the extreme solubility and absence of crystallizing power of the alkaline and earthy salts, they cannot be obtained in a form suitable for analysis. The primary salts have doubtless the general formula POCNH2).OM, (or possibly P(NH)(NH2) (OH)(OM); those with more base are in some cases possibly derivatives of the same acid, in others they are unquestionably diamidotrihydroxylphosphates, reacting with silver nitrate directly to form salts of this acid. They are all much more stable than the corresponding monamidophosphates. Poiassiicm and Sodium Salts. — By adding to i molecular weight 1 This Journal 15, 204. 1 34 Stokes. free acid, i molecular weight potassium or sodium hydroxide, solutions are obtained, which, on evaporation in small quantities on the water-bath or in vacuo, leave perfectly transparent vitreous masses, readily soluble in water, but not deliquescent, insoluble in alcohol, and which are primary salts ; silver nitrate precipi- tates from the solution the characteristic primary silver salt. With 2 molecular weights alkali, the secondary salts, are formed, which have similar properties, but which give the secondary silver salt. With increasing quantities of alkali, the silver precipi- tate changes to yellow and finally brown, the yellow body being the tertiary salt of the pentabasic acid. As seen in the experi- ments on the saponification of the ether, tertiary potassium diamidotrihydroxylphosphate is converted into secondary salt by phenol, into primary salt by carbon dioxide, and into free acid by acetic acid. Avimo)iium Salts. — i molecular weight free acid is able to com- bine completely with i molecular weight ammonia; with 2 molec- ular weights the solution has an ammoniacal odor and gives with silver nitrate a mixture of secondary and primary silver salts. Barium Salts. — The primary salt, ( PO^q ^^-^' ] Ba, is formed by adding to the free acid the calculated amount of baryta solution, or by using an excess of the latter, and treating with carbon diox- ide, or by digesting the acid with excess of barium carbonate. It is most conveniently made directly (mixed with phenol) by sapon- ifying the ether as above described. It is extremely soluble in water, and on careful evaporation is left as a transparent vitreous mass. Alcohol precipitates it in viscous form. The solution decomposes very slowly in the cold, but quite rapidly on heating, forming a white granular precipitate ; but even after boiling an hour or more, much remains undecomposed, as. shown by its yielding the characteristic primary silver salt. By using i molec- ular weight acid and i molecular weight baryta the secondary salt is formed, which in physical properties resembles the primary salt, but which gives secondary silver salt, and is decomposed by carbon dioxide. The solutions, with excess of baryta, remain clear, and are but very slowly decomposed by boiling. Magnesium Salts. — The free acid is not precipitated by mag- nesia mixture ; it dissolves magnesia and decomposes the carbon- ate in the cold, forming primary salt. If the oxide be used, a Amidophosphoric Acids. 135 small quantity of secondary salt seems to form, as the solution gives, besides primary, a small amount of secondary silver diam- idophosphate. A determination of the ratio in the solution gave P : Mg=:2 : 1.17, or nearly that required for fPO^Q_^'^ j^Mg. The salt resembles the barium and alkali compounds, leaving a transparent glass on evaporation, which is permanent in the air. Alcohol precipitates it as a clear liquid. On boiling, it slowly deposits a flocculent precipitate. Other Salts. — Primary barium salt gives an abundant granular light-blue precipitate with cupric nitrate, and a white pulverulent precipitate with lead nitrate ; with mercuric chloride no precipitate is formed, and with nickel and cobalt nitrates a precipitate only on long standing. Silver Salts. These are by far the best characterized salts observed. They have the advantages of being for the most part insoluble in water, of having sometimes characteristic crystalline form, in other cases characteristic color, of being usually free from crystal water, and of containing a univalent metal, thus giving indications as to the basicity of the acid. Some are to be referred to diamidoortho- phosphoric- acid, others are diamidotrihydroxylphosphates. They are all unaffected by light. Primary Silver DiaviidopJwsphate , PO^q a "^". — This highly crystalline, very characteristic and stable salt is obtained in several ways, but most conveniently from the primary barium salt, pre- pared by direct saponification of the ether. To the dilute solution (which may contain a little barium dicarbonate) is added, drop by drop, the calculated quantity of silver nitrate solution, when the salt rapidly separates as a coarse white sand, which is washed, dissolved in ammonia, diluted to about one liter for each fifty grams, and precipitated by adding dilute nitric acid, drop by drop. Each drop produces a white amorphous precipitate, which at once redissolves, and presently the primary salt crystallizes out. The addition of nitric acid should be continued to neutral reaction. The yield, thus purified, is about equal to the weight of phenyl ether taken, or 85 per cent, of the theoretical. Carbon dioxide also precipitates it in apparently pure form from the ammonia solution. It loses nothing over sulphuric acid or at 100°. 1 36 Stokes. Calculated for Found. PO(NH2),.OAg. (>). (2)- (3>. p 15-29 15-32 14-57 15.28 N I3-83 12.97 I3.II 13.48, 13.44, 13.42 Ag 5314 52.81 5389 53-01 (i). P : N : Ag= I : 1.87 and 1.89 : 0.99. (2). P:Ag=i: 1.06. (i) and (2), from free acid and i molecular weight ammonia, (i) with calculated, (2) with large excess of silver nitrate; not recrystallized. (3). P: N:Ag= 1:1.94:1. Prepared like (i), and recrystallized. The analysis of this and other silver salts is best effected by dissolving in dilute nitric acid for silver and phosphorus, and in hot dilute sulphuric acid for nitrogen, the solution being heated a half-hour, the silver removed by hydrochloric acid, and the filtrate distilled with alkali. Concentrated sulphuric acid acts violently, usually causing evolution of free nitrogen, and in some cases explosion. Prepared as above, the salt forms a white sand, but by spon- taneous decomposition of the secondary salt under water, or by dissolving either salt in ammonia and evaporating over sulphuric acid, crystals of two or three millimeters in length are obtained. The crystals are thus described by Prof. J. P. Iddings; " The habit is sometimes short prismatic, with prism faces horizontally striated and terminated at both ends by a single, somewhat acute rhombohedron. In other cases the habit is rhombohedral, consisting of the same rhombohedron with or without small prismatic truncations of the angles. It often sepa- rates in acicular forms, which, on closer examination, are seen to consist of groups of rhombohedra united in the direction of the principal axis, and occasionally showing a tendency to form stel- late and cruciform groups." The salt is also obtained in characteristic form by dissolving any of the silver salts described below in ammonia and precipi- tating by nitric acid, but as the higher salts often give at first amorphous precipitates, the process has frequently to be repeated ; also by digesting the free acid with silver oxide. It is nearly insoluble in water, readily in ammonia, of which it requires exactly two molecular weights. The ammoniacal solu- tion gives with silver nitrate a precipitate of more or less impure Amidophosphoric Acids. 137 secondary salt. Sulphuretted hydrogen liberates the free acid. If. rapidly heated it swells up, evolves ammonia, and finally fuses to a grayish mass. If heated at i50°-i6o° it slowly loses weight, the loss approximating to that required by the equation .P0(NH=). = NH<^<| + NH., that is, to that required by the formation of the silver salt of diamidopyrimidophosphoric acid. Calculated Loss. Found. 2 mols. — NH3 4.20 After 2 hours at i50°-i6o° 4.65 2 mols. — 2NH3 8.40 " 6 " " " '6.11 The rapid loss of the first molecule of ammonia is accompanied by a very slow loss of the second ; at 200° the rate of loss dimin- ishes and is finally succeeded by an actual increase of weight of about 3.8 per cent., which is followed by a further loss at 230°, with fusion. Although the experiment was not carried further, observations on pyrimidophosphoric acid,' which will appear in a separate paper, show that its silver salt is oxidized at 200° by dry oxygen to silver pyrophosphate. The increase of weight in the above case is doubtless due to oxidation of the imido groups to free nitrogen and water, and combination with the latter to form acid silver pyrophosphate, which fuses with loss of water at 225°. The salt formed by the loss of the first molecule of ammo- nia will be studied further when opportunity admits, as its acid has an empirical formula identical with that of Gladstone's PO(NH-.0. pyrophosphoiriamic acid,"" 0<:^pp.OH , the existence of which is denied by Mente,' who regards Gladstone's compound as NHNH . Both observers prepared the acid in the same manner, by acting on phosphorus oxychloride with dry ammonia, and subsequent treatment with water, and each obtained analytical data agreeing well with his own formula. If Gladstone's acid has really the formula PsN^HiO^, it is probably the diamide of pyrimidophosphoric acid, not the triamide of pyrophosphoric acid, I This Journal 15, 212. - Jour. Chem. Soc. [2] 4, i ; [2] 6, 68. 3Ann. Chem. (Liebig) 348, 246. 1 38 Stokes. and hence identical with the acid of the compound obtained by heating primary silver diamidophosphate. Experiments on phenyldichlorphosphate, above alluded to, indicate that the link- age of the two phosphorus atoms probably occurs during treat- ment with ammonia gas, and not by the subsequent action of water. Primary silver diamidophosphate probably contains silver united with oxygen, this conclusion being based on the slowness with which it decomposes on heating, as distinguished from those salts known to contain silver joined to nitrogen, which deflagrate or explode, and also on the absence of color, the known nitrogen salts being colored. A silver salt crystallizing in large colorless scales was frequently obtained by dissolving various silver diamidotrihydroxylphos- phates in ammonia and precipitating by nitric acid, and also by neutralizing a solution of the free acid in caustic potash, and pre- cipitating by silver nitrate. It forms more or less elongated plates with parallel edges and ragged ends, and quickly passes into rhombohedra of the primary salt. The exact conditions under which it forms have not yet been made out ; it is perhaps (NHO2 the primary salt of the pentabasic acid, P OAg . (OH> Secondary Silver Diamidophosphate , or Amidophosphimatey P(NH),Q . \^. — A solution containing i mol. wt. free acid and I mol. wt. barium hydroxide was poured into a solution of 2 or more mol. wts. silver nitrate, when a perfectly amorphous straw- colored precipitate was obtained. This was washed, dried on a porous plate and over sulphuric acid. It lost practically nothing at 100° and gave on analysis: Calculated for PN,,H302Ag„. Found. P 10.01 9.98 Ag 69.61 7i'04 or P : Ag=i I : 2.05. If the primary salt be dissolved in ammonia, and silver nitrate added, an amorphous precipitate is obtained, which, when an excess of ammonia has been used, is often colorless, but becomes straw-colored on washing, and frequently has this color from the first. This is doubtless the secondary salt, colorless when pure, Amidopiwsphoric Acids. 139 but usually contaminated with small quantities of the yellow salt described below. There is a strong tendency to the formation of more or less primary salt at the same time, which is easily recog- nized under the microscope. It has been already mentioned that' diamidophosphoric acid is able to hold i mol. ammonia, but that with more than this the solution has an ammoniacal odor ; hence it appears that the secondary salt dissociates easily. The propor- tion of primary and secondary silver salt formed from this solution depends on the relative amounts of acid, ammonia and silver present. Many attempts were made to prepare pure secondary salt in this way, but although some preparations seemed to be free from primary salt, the analyses were always unsatisfactory and may be omitted. One experiment only may be mentioned : I mol. wt. acid, 3 mol. wt. ammonia and 2 mol. wt. silver nitrate gave a curdy precipitate which was taken for secondary salt ; on warming, however, it was at once converted into a crystalline powder which proved to be only primary salt : Calculated for PO(NH,)50Ag. Found. p 15.29 15-16 N 13.83 13.61 Ag 53-14 53-21 Ag=:i; : 1.99: i.oi. or P:N Whenever the amount of silver was relatively increased, there was an increase in the quantity of secondary salt. Dry secondary silver diamidophosphate, if heated, swells up, gives off ammonia and water, and leaves a residue grayish from metallic silver; it is totally devoid of explosive properties. Its solution in ammonia, if allowed to evaporate over sulphuric acid, deposits large crystals of primary salt, but no secondary salt. It shows a remarkable behavior on standing or heating under water. In the course of a few hours, dark-red specks appear in the nearly colorless amorphous mass, and in the course of a day or two the whole has become converted into a mixture of dark-red granules, and large colorless crystals of the primary salt, the change being accompanied by great diminution of bulk ; the same transforma- tion occurs in a few moments on boiling. It also occurs to a slight extent in the course of weeks or months if the salt be pre- served in an imperfectly dried condition, as indicated by its acquiring a pink or reddish tint; if absolutely freed from moisture 140 Siokes. by careful drying over sulphuric acid, it can be kept without change indefinitely. The red substance thus formed, which is referred to hereafter as "red salt," has a color almost identical with that of silver dichromate. It is formed in several other ways, which will be described below. Although very stable, it is difficult to obtain with constant composition, and it cannot be recrystallized ; hence some doubt still exists as to its nature; it is, however, certain that it contains phosphorus, nitrogen and silver in the ratio 1:2:4, and it may be provisionally regarded as tetra-silver amidophos- NHAg phimate: P(NAg)OAg . OAg The secondary salt, in the dry state, has the formula PN-iHs- OsAgs, which makes it a salt of diamidophosphoric acid. Its white color and the absence of explosive properties make it almost certain that the silver is wholly united with oxygen, as NH" represented by P(NH),j^ , \ , i. diamidotrihydroxylphosphate, (OAg)2, which loses i molecule OH water on drying. Mixtures of free diamidophosphoric acid and alkali behave in every respect as pentabasic compounds, as will be shown below, and this salt is made from such a mixture. The strong tendency which the fresh, moist salt shows 10 pass into others containing more silver, is shown by the trihydroxyl com- ^ pounds only ; the primary salt, which is a diamidoorthophosphate, \ does not show it in the least. The action of caustic potash and of ammonia on this salt will be stated below. Action of Caustic Potash o?t Primary Silver Diamidophosphate. Diamidotrihydroxylphosphates. The behavior of potassium hydroxide towards primary silver diamidophosphate leaves no doubt that diamidophosphoric acid readily forms addition-products with bases, which are salts of the diamide of pentabasic phosphoric acid. The action of alkaline carbonates, which is somewhat different, is stated in a separate Amidophosphoric Acids. 141 section; boiling baryta-solution acts only superficially on the silver salt. If 30-per cent, caustic potash solution be poured on the primary silver salt, a coherent white pasty mass results. It is better to add the salt in small portions to the alkali, with stirring, about 10 cc. being taken for each gram salt; as made in this way the paste is less coherent, forming small fragments. Not a trace of silver oxide is formed. On adding water cautiously, with constant gentle agitation, the paste swells up to transparent masses of jelly, and if about ten volumes of water have been added, this appears gradually to dissolve, and finally a solution is obtained which is as viscous as thick starch-paste or mucilage, and perfectly colorless and transparent. Portions of undissolved jelly may be brought into solution by breaking up with a rod. Further addition of water, up to a certain point, causes no visible effect further than to thin the liquid ; on adding more, a yellowish tint appears, which may be destroyed by adding more potash. On diluting further, the color increases until finally a deep wine- red, transparent solution is obtained. It requires much water to bring this about, but the color is intense, one gram of silver salt giving a deep color to one liter of water. The jelly is also obtained by adding potash solution, drop by drop, to the silver salt covered with water. Properties of the Potash Jelly. — The clear gelatinous solution remains at first unchang-ed, but in the course of half an hour, or sooner, if violendy shaken, it begins to grow turbid, and after several hours has deposited a voluminous snow-white mass,^ the liquid at the same time losing its viscosity. The new body is washed with 3-per cent, potash — pure water decomposes it instandy — and is then quickly pressed out on a clay plate, trans- ferred to another clay plate and dried in vacuo over sulphuric acid. Owing to the necessity of washing with potash solution, the potassium determination falls somewhat too high. Analysis shows it to be Mono-potassitmi-di-silver Diamidotrihydroxyiphosphate, P^p.^'^" -|- 2H2O. — The substance, dried over sulphuric acid, OK^' lost nothing at 100° and gave : 'It is not absolutely necessary to dilute the alkaline liquid to obtain this compound; it forms directly and more rapidly, but not in as well crystallized form, in the undiluted 30-per cent, potash solution. 142 Stokes. Calculated for Found. PN^H^OaKAg^-l-aHjO. (i). (2). P 7.72 8.23 7.72 7.59 K 9.73 ... 11.68 11.69 Ag 53-68 53.57 53.39 52.84 (i). P: Ag=: I : 1.87. Washed with 3-per cent, potash, alcohol and ether. (2). P : K : Ag= i : 1.22 : 1.99. Washed with 3-per cent, pot- ash, and dried on porous plate over sulphuric acid. The alkaline mother-liquor, which contains no silver, on neutral- izing and adding silver nitrate, gives an abundant precipitate of primary silver diamidophosphate, indicating that only a portion of the acid has gone to form the double salt. The reason for this appears from the following considerations : The action of the potash, in forming the jelly, is doubtless simply one of addition: Po(NHO^ (NH.). *^*^ OAg -h 2KOH =: P OAg -fH.O. (OK)2 Primary Salt. Je"y- The double salt with two atoms silver is then formed thus : (NH.> (NHO^ .NH.V 2POAg =P(OAg). + P^2{J^|. (OK> OK ^^^^' Jelly. Needles. The last remains in solution, and on neutralizing, gives primary potassium diamidophosphate, precipitated by silver nitrate as primary silver salt. Occasionally the scaly salt above referred to was obtained, which rapidly changed to rhombohedra of the primary salt. The same crystalline double salt is formed by direct action of caustic potash on the secondary silver salt : NH. (NH,). and a probably analogous sodium compound, also crystallizing in needles, is obtained from the primary salt and caustic soda ; the formation of a jelly was not observed. Mono-potassium-di-silver diamidotrihydroxylphosphate forms snow-white voluminous masses, which, under the microscope, are seen to consist of tufts of long, straight, radiating prisms. The Amidophosphoric Acids. 143 dried salt is usually not absolutely white. The water of crystal- lization is held tenaciously, and is scarcely given off below the point of decomposition. On heating, the salt decomposes rather suddenly, with a faint crackling sound, giving off much water and ammonia. It may be boiled any length of time with 30-per cent, potash, without undergoing any appreciable change except a trifling discoloration. In fact, potash of this strength converts all the other silver salts into this compound, slowly in the cold, rapidly on heating, any excess of silver above two atoms being liberated as oxide. If, however, sufficient free diamidophosphoric acid be added, the conversion is complete. Silver oxide is converted into the same salt by digesting or boiling with strong potash and diamidophosphoric acid. Even dilute potash, as 3-per cent., effects the same changes slowly in the cold ; if, however, the liquid be boiled, the five-atom explosive silver salt is formed, which is grad- ually reconverted into the needles on standing in the cold. These observations explain the following: if heated in a test-tube under the mother-liquor (which usually contains about 3 per cent, potash), the needles turn yellow, then red, and then brown, at the bottom of the tube, where the heat is greatest, but on removing from the flame they instantly become white again ; this cham- eleon effect may be repeated indefinitely. If the whole contents of the tube be allowed to become hot, the needles are wholly con- verted into the brown explosive salt; on cooling and allowing to stand several hours, this passes back into the colorless needles ; this, too, may be repeated indefinitely. On dissolving in ammonia and neutralizing with nitric acid, the primary salt is regenerated. Tri-silver Diamidotrihydroxylphosphate, P/q a \ -f- 2H2O. — The above acicular double salt, if washed with cold water, becomes yellow, especially if it has been dried ; if freshly prepared and still moist, it sometimes becomes yellow, but usually a yellow- ish-flesh color. When pure yellow, it suggests silver phosphate. This change of color is accompanied by a deep decomposition, a salt with three atoms of silver being formed, while the filtrate is found to contain potassium diamidophosphate in large quantities. The change is represented by the equation : OK ^ ^0^^> ^(OK)3 • Vol. XVI.-io. 144 Stokes. The new substance is in the form of needles, which are, how- ever, merely pseudomorphs after the potassium silver salt. Com- plete decomposition is difficult to effect, as no actual solution occurs; hence hot water extracts a little alkali, and the analytical results are not very close. The air-dried substance contains approximately two molecules water, which it loses but very slowly in the air, more rapidly over sulphuric acid, and at once at ioo°. This and the potassium silver salt are the only amidophosphates of silver which have been found to contain crystal water. Dried at ioo° its color is orange. Three preparations, dried at this temperature, gave : Calculated for Found. P(NH,),(OAg),. (i). (2). f3). P 7.14 7.40 7.38 7.26 Ag 74.46 74.23 73.75 74.33 2.88 P:Ag=^(2). 1:2.75 2.94 C(i). = i(2). U3). The water lost by the air-dried substance at 100°, expressed in percentage of the anhydrous substance was Calculated for Found. 2H,0. (1). (2). (3). 8.28 8.09 7.69 7.60 6.71 Loss ((i). 1 : 1.89 and 1.79 (2;. 1:1.77 (3)- 1 : 1-59 If free from water, as when dried at 100° or for a long time over sulphuric acid, it may be kept indefinitely without change, but if preserved after simply drying in the air, that is, with crystal water, it remains unchanged for a week or two, and then suddenly, in the course of a day or two, undergoes a change, turning red and becoming granular. It also passes gradually into a red salt when kept under water. Whether the red compound formed by this change is identical with that obtained by other methods is uncertain. It is much lighter in color, but this might be caused by admixture with a colorless substance. No free acid is formed, and microscopic examination fails to indicate the presence of primary salt. If dry, the yellow salt turns brown on heating and then suddenly deflagrates, giving off" ammonia; if not anhydrous, i\ Amidophosphoric Acids. 145 it turns red before deflagrating. Ammonia dissolves it with some difficulty, as is the case with the other silver diamidotrihydroxyl- phosphates, first turning it white; from this solution nitric acid frequently precipitates the white scaly salt, which gives rhombo- hedra of primary salt on boiling. I regard this yellow salt as containing silver united with oxygen only, because of its resemblance in color to silver phosphate, assuming the color to be due to the common chromophore = P(OAg)s, and because its formation would take place by a process analogous to that often occurring in the decomposition of double salts; also because it is formed by digesting the acicular mono-potassium di-silver salt with silver nitrate. The yellowish color frequently shown by the secondary salt, and which it always takes on momentarily on heating, before passing into the red compound, is possibly to be ascribed to the formation of this body. If boiling water be poured on the yellow salt, it is instantly transformed into a red-brown amorphous substance which, on heating a few minutes under water, changes to the red salt with 4 atoms of silver. The same change is shown by the crystalline potassium silver salt, but the product in this case is not pure, owing to the further action of the alkali. The first stage of the decomposition may be expressed by the equation : 4P(OAg)3 = 3P(OAg> +PO ^oj^ +H.O. ^-red-brown salt. The /J-red-brown salt then decomposes further, as described below, into red salt and water. Decomposition of the Potash Jelly by Cold Water. a- and /J- Teira-silver Diamidotrihydroxylphosphate, PNzHsOj- Ag4. — The effect of diluting the potash jelly has been already described. Any lumps of jelly remaining undissolved are grad- ually converted into an insoluble red-brown amorphous substance. The wine-red solution is very stable, and may be kept any length of time or boiled without visible change, but is decolorized by adding a large quantity of potash. On passing carbon dioxide into the cold solution it soon becomes turbid and finally colorless, depositing a flocculent substance, which in color and general 146 Stokes. appearance closely resembles precipitated ferric hydroxide. The action of carbon dioxide must be discontinued as soon as the pre- cipitate has completely formed, otherwise some silver carbonate is formed. The substance is readily washed by decantation with cold water and on the filter; the washing should be thorough, as it retains a trace of alkali somewhat persistently. It is then sucked out and dried in vacuo over sulphuric acid, when it forms a lump with vitreous fracture, easily pulverized to a red-brown powder. It usually loses about 2 per cent, water on drying at ioo°, but appears to be otherwise unchanged. The analysis of two prepar- ations, dried at ioo°, gave : P Ag Calculated for PN^HaOjAg,. .5-73 79-68 Found, (i). (2). 5.51 5.73 79.36 79.20 (I). (2). P:Ag=i P:Ag=i . 3-98- The substance dried over sulphuric acid for a few days shows some chemical difference from the freshly precipitated and still moist compound, the cause of which is not clear, but which can- not consist in a difference of empirical composition, except in as far as this may be due to loss of water. The fresh substance may be distinguished as a, the dried as /?. a-tetra-silver diamidotrihydroxylphosphate does not change by standing under cold water, but under boiling water it changes slowly, becoming more compact, and after long heating (an hour) being converted into a heavy, nearly black powder, which does not explode by friction, and which under the microscope is seen to be granular, with indications of being crystalline; no red salt is formed. A strongly alkaline solution of potassium diamidophos- phate instantly converts the a-salt into needles of the potassium silver compound ; potash alone, in the cold, slowly converts it into the same body, liberating silver oxide. /3-tetra-silver diamidotrihydroxylphosphate does not explode on friction; on heating it detonates rather weakly, with a reddish flash, giving a little moisture and ammonia. Dropped into strong sulphuric acid it also flashes, but gives no sound, a little metallic silver separating. When boiled with water, it changes in two or three minutes into red salt (see analysis), which contains a higher percentage of phosphorus and silver, but in the same ratio. The Amidophosphoric Acids. 147 y?-salt is probably identical with the red-brown substance obtained by treating the tertiary silver salt with hot water, as they have the same appearance and both yield the red salt very quickly on boiling. Towards potash and other reagents it shows the same general behavior as the other salts. For a body of the above composition but two formulas are possible : NH'2 (NHAg> (a) PNHAg and (b) P(OAg> . (0Ag)3 OH A substance of formula (b) would probably lose water quite readily, giving red salt, PNAg ,q , s^, which is true of the /3-salt. The isomeric body (a) would probably change much more slowly, and the tendency would be not to lose water but silver oxide, thereby becoming black, and this is the behavior shown by the a-salt. Such reasoning is unsatisfactory, and positive evidence of their constitution must be sought in further experi- ments. The undiluted potash jelly is also decomposed by carbon dioxide, but the liquid remains colorless ; there is formed a pre- cipitate which closely resembles, and is possibly identical with the «-tetra-silver salt. By long-continued action of carbon dioxide this is finally converted into silver carbonate. The nature of the dark-red solution there is, of course, no means of directly determining, as it contains a large excess of potash. From the solubility of the colored body we may infer that it is a potassium salt, and from its dark color that it contains silver united with nitrogen. There is no reason for supposing that it contains, like the a-precipitate, four atoms of silver, on the con- trary it must contain less, as this would require at least two. to be joined to oxygen, and such a body, like the white salt with two atoms of silver and one of potassium, would be insoluble. Decomposition of the Potash Jelly by Heat. Penta-silver Diamidotrihydroxylphosphate, P/q a \ • — If the flask containing the colorless, transparent potash jelly (diluted to a strength of about 2-3 per cent, potash) be placed in boiling water, the solution becomes red-brown, then turbid, and in a few Found. (2). (3). 4-95 4.71 83-33 82.30 148 stokes. minutes deposits a dark-brown amorphous precipitate, which might be mistaken for silver oxide. In order to obtain it pure it is neces- sary to heat the liquid for about fifteen minutes at 100°, or until the precipitate appears homogeneous, otherwise it is apt to con- tain some of the red-brown salt and its appearance is streaky. It is then washed once or twice with boiling water, by decantation, again digested at 100° with 2-3 per cent, potash, washed with boiling water, sucked out and dried z« vacuo over sulphuric acid. Owing to its dangerous properties when dry, it is advisable to break up the moist cake. It is not even necessary to gelatinize the primary silver salt ; simple boiling with 2-3 per cent, potash suffices to effect the transformation, but it is apt to be incomplete. The substance lost nothing at 100°, and gave the following figures on analysis : Calculated for PNjHjOsAge. (I). P 4.78 4.87 Ag 83.18 81.64 (i). P : Ag=: I 14.82. Prepared from jelly by heating only. (2). P : Ag=r 1 :4.84. From primary salt without previously gelatinizing. (3). P : Ag =: 1 : 5.03. From jelly after repeated treatment with potash. A second method consists in boiling the needles (potassium- silver salt) with dilute potash ; a third in treating the tertiary silver salt in the same way ; the salts with four atoms of silver undergo the same transformation. These changes may be repre- sented as follows : (3). 5P[si+^KOH = 3P(g^^,?>+.p(S{^0= + ,H=O. Reaction (2) is reversible, reading from left to right when the potash is hot and dilute ; on cooling, it -occurs in the opposite sense, as it also does when the five-atom salt is boiled with 30-per cent, potash and diamidophosphoric acid. Reactions (i) and (3) Amidophosphoric Acids. 149 are not strictly reversible, but on standing in the cold, the potas- sium salt (3 molecules) acts on the silver salt (2 molecules), con- verting it in (i) wholly, in (3) partly, into the crystalline double salt, the remainder being converted by any excess of potash into the same body and silver oxide. This salt, which appears to contain the greatest amount of silver with which diamidotrihydroxylphosphoric acid can combine, and which proves its pentabasic nature, may be safely handled or heated in moist condition, or dried at 100°; when dry it may often be rubbed quite strongly in a smooth agate mortar, but finally explodes unexpectedly on lightly touching, giving a reddish flash. It instantly detonates with a flash on percussion, on rubbing on a rough surface, on heating, or on dropping into strong sulphuric acid. One decigram gives a report like the discharge of a pistol, but the explosion of that amount would scarcely break a thick test-tube. Owing to the unexpectedness with which it explodes, however, it is advisable to protect the eyes when handling it, as the solid decomposition-products are thrown to a considerable distance. The ease with which this body explodes, as compared with the other compounds, is remarkable. None of the others explode on friction, although the colored salts deflagrate or explode weakly on heating. The presence of a trace of BerthoUet's compound, which, according to Raschig,' is silver nitride, NAgs, is out of the question, as this is formed only in strongly ammoniacal solution, while in the present case no ammonia was present, unless it were a minute trace formed by decomposition of the diamide. More probable is the presence of a body containing the group NAgs, which would probably be explosive, and which in small amounts would not be indicated by the analysis. It is conceivable that a small quantity of penta-silver amidophosphi- mate, P(NAg)/Q A^l^ might be formed by dehydration which would probably be explosive. The subject, which is rendered difficult by the impossibility of purifying the products, will receive further attention. ^' Red Salt,'' Tetra-silver Amidophosphimate, PNAg.Q« -f? — Notwithstanding the variety of conditions under which this sub- 1 Ann. Chem. ( Liebig ) 333, 93. 1 50 stokes. stance is formed, and its great stability, none of these compounds have given greater difficulty in obtaining satisfactory analytical data. It must be remembered that with the single exception of the primary salt, none of these bodies can be recrystallized ; with this exception, however, the analysis leaves no doubt as to their composition; the nature of the red salt must be inferred. The body has been obtained in the following ways : (i). By boiling secondary silver diamidophosphate, or by allowing it to stand forty-eight hours under its mother-liquor, or under water, whereby it is transformed into primary salt and red salt. (2). By boiling tri-silver diamidotrihydroxylphosphate, where- by free diamidophosphoric acid and a red-brown substance, probably /3-tetra-silver diamidotrihydroxylphosphate, result, the latter passing on further boiling into red salt. (3). By boiling /S-tetra-silver diamidotrihydroxylphosphate a few minutes with water. (4). By boiling primary silver diamidophosphate with potassium or sodium carbonate ; the same changes of color result as stated under (2). (5). By boiling secondary silver salt with an amount of ammonia insufficient for complete solution ; here, too, the red-brown com- pound is an intermediate product. The following analyses of this salt, prepared by these methods, and dried at 100°. may be given. By method (i): (')■ Found. M- (3)- (4)- (5). p 6.09 6.08 6.12 6. 1 1 6.04 N ... 547 Ag 80.21 80.69 80-09 80.83 81.14 (I). P : Ag = 1 : 3.78 (2). P:N: Ag= i: 1.99: 3.82 (3). P:Ag =1:3.76 (4). P:Ag =1:3.81 (5). P:Ag =1:3.87 By method (2) : Found. (6). (71. P 6.05 5.98 Ag 81.06 81.34 (6). P:Ag = :i:3.85 (7). P:Ag = i:3.82 Amidophosphoric Acids. 151 By method (3) : Found. (8). P 5.84 Ag 80.7 1 (8). P:Ag= 1:3.97 In all but the last, the amount of silver is markedly below, but approximating to four atoms. In (8) the ratio is 1:4, and this may be regarded as the purest; it was made by boiling a sample of /3-tetra-silver salt of known composition, and the figures are given for comparison. They show that the change is not a simple conversion into an isomer, but that the red salt contains more phosphorus and silver, but in the same ratio, which is only to be explained by loss of water. The same is true of all the prepara- tions, which contain too much phosphorus and silver to be diamidotrihydroxylphosphates. The loss of water would be as represented in one of the following equations : (a). (NHAg) aPCO Ag). — H2O : OH = P(NAg)^NHAg (b). (NHAg). 2P(OAg)2 — H2O OH _ /p(NHAg)A -^l^(OAg). )^ Calculated for red-brovrn salt, PN.HjOaAg,. Calculated for Calculated for PNjH05Ag4. P^N^H^OjAga. P 5-73 5-93 5-83 Ag 79.68 82.41 81.02 Found for red-brown salt. Found for derived red salt. P 5-73 5-84 N 79.20 80.71 P:N=i:3.98 P:Nz= 1:3.97 If formed according to (a), it would be tetra-silver amidophos- phimate; if according to (b), a derivative of pyrophosphoric acid. Although the figures agree better with the latter, the ease with which it is reconverted into primary diamidophosphate speaks in favor of (a) ; a pyrophosphate would probably be more stable. The nature of the compound must therefore be left undecided. The red salt, formed from hot solution, consists of irregular granules : if formed in the cold, it forms either spherical trans- 152 Stakes. parent grains without a sign of crystallization, or spherulites with indistinctly radial structure. It is very compact. It undergoes no change at 100°, except an occasional loss of 0.2-0.3 per cent. Heated in a test-tube it deflagrates with a red flash, but no deton- ation, leaving a mossy residue containing metallic silver. Dropped into concentrated sulphuric acid it flashes or becomes red-hot ; at the same time there is an evolution of nitrogen and separation of metallic silver. It is difficultly soluble in ammonia, requiring a large excess and considerable time; its separation from the primary salt, when formed by method (i), is best effected by weak ammonia. The ammoniacal solution on evaporation or careful neutralization with nitric acid gives an amorphous white precipi- tate, probably secondary salt, which on redissolving is precipitated as primary salt, or sometimes in the form of the large scales above referred to which soon change to primary salt. Potassium diamidophosphate, in presence of potash, soon converts it into needles of the potassium-silver salt. Action of Ammonia on Secondary Silver Amidophosphimate. — The secondary salt, if treated with an amount of ammonia insuffi- cient for complete solution, forms a transparent pasty or slimy substance, which soon hardens to nearly transparent lumps, while the solution becomes thick, froths like soap, and gradually deposits a white substance. These transformations are doubtless due to the formation of addition-products, or double salts. On boiling, the solid is converted into an amorphous brown substance, while a similar substance separates from the solution ; these change to red salt on further heating, Actioji of Alkaline Carbonates oti the Silver Salts. — The action of alkaline carbonates on the primary silver salt differs markedly from that of caustic alkalies, but its interpretation is similar. If the salt be covered with a strong solution of potassium or sodium carbonate, no action is apparent in the cold, except perhaps a slight discoloration. On heating, it instantly turns red-brown, and after a few minutes' boiling is completely converted into red salt, which is unchanged by prolonged boiling. No five-atom compound appears to be formed, as when caustic alkali is used. As the primary salt is unaffected by boiling water, we must assume the formation of a silver-potassium diamidotrihydroxyl- phosphate and potassium dicarbonate : 154 Stokes. P0(^"'>^ + K.CO3 + H.O = P OAcJ ' + KHCO». OH the double salt undergoing the usual transformations into the tetra-silver salt and potassium salt, the former then losing water and forming the red body. All the silver salts are unacted on by sodium dicarbonate in the cold, but on warming they are completely converted into sodium salt and silver carbonate. The accompanying table shows the relations of the silver salts and their transformations. January, 1894. NOTE ON MONAMIDOPHOSPHORIC ACID. By H. N. Stokes. The researches of Gladstone' and of Mente" have shown that the action of dry ammonia on phosphorus oxychloride results mainly in the formation of complex amido- or imidophosphoric acids. Whether amides of orthophosphoric acid are formed in small amounts at the same time is left undecided by these inves- tigators. These bodies were totally unknown at the time, and I can find nothing in their writings which would answer the question negatively. Supplementary to a work on monamidophosphoric acid,' I made a few experiments in order to ascertain if this acid could be produced by the action of dry ammonia on ethyl monochlor- phosphate, PO.C1(OC5Hb)2, and if so, whether it would be a better method than that actually used. Having satisfied myself that dry ammonia is without action on ethyl phosphate at i6o°, and hence does not easily replace the ethoxyl group by amide, the only reactions to be expected were : (a). PO.Cl(OGH5> -f 2NH3 = PO(NHs)(OOH6> + NH^Cl. (b). 2PO.Cl(OGH5), + 3NH3 =: NH<^g^g^^^g;); + 2NH4CI. ' Journ. Chem. Soc. [2] 3, 225; 4-, i, 290; 6, 64, 261; 7, 15. "Ann. Chem. (I.iebig) fJ48, 232. 3 This Journal 15, 198. Note on Monamidophosphoric Acid. 155 Analogy with the action of this reagent on phosphorus oxy chloride would lead us to expect the reaction to occur only in the sense of the second equation. The theoretical increase of weight would be for (a) 19.8 per cent., for (b) 14.8 per cent. Ethyl monochlorphosphate, which was prepared as directed by Wichelhaus,' was subjected to the action of dry ammonia at 0° until no further increase of weight was observed. The total increase was 17.2 per cent., or intermediate between (a) and (b). The semi-fluid mass, directly saponified with an excess of caustic soda, and acidified with acetic acid, gave an amount of primary sodium amidophosphate corresponding to about 20 per cent, of the theoretical. It seems, then, that the reaction occurs partially in the sense of equation (a), but that the yield is not so good as to make this method preferable to the other. At 100°, the action of ammonia was totally different; instead of a gain in weight of between 14.8 and 19.8 per cent., there was a loss of 15.8 per cent. A combustible gas came off in streams, which, after treatment with concentrated sulphuric acid, proved to be ethyl chloride. The solid residue, which dissolved, though slowly, in water, was not further examined. The reaction POCl(OGH5)2 + 2NH3 = PO.NHo. BrC— CO BrC— CO H Toennies found that his product could be converted into muco- bromic acid by oxidation with chromic acid, and therefore con- sidered it to be the double aldehyde of dibromfumaric acid. We found, on the other hand, that the body showed none of the ord- inary characters of an aldehyde. On warming with chromic acid it was slowly oxidized, but mucobromic acid was not formed in quantity sufficient to enable us to identify it with precision. On long boiling with bromine and water, mucobromic acid was formed. From concentrated nitric acid it could be recrystallized unchanged, ■ Ber. d. chem. Ges. 12, 1202. ^ Proc. Am. Acad. 21, 172. Crotonolactones aiid Mucobrornic Acid. 191 and only after continued boiling was it oxidized to mucobrornic and dibrommaleic acids. It did not combine with acid sodic sul- phite, and did not react with hydroxylamine. Aniline in alcoholic solution removed one atom of bromine and gave a phenylamido derivative of the form C4H2(C6H5NH)Br02, and phenylhydrazine also removed bromine. Aqueous alkalies dissolved it with the formation of a deep yellow solution, but decomposition soon ensued with the elimination of hydrobromic acid. The whole behavior of the body was in direct opposition to the assumption that it was a double aldehyde, and the unsymmetrical structure was further established by the existence of two isomeric bodies of the same general constitution which contained a single bromine atom. The conclusion was inevitable that the body was a dibrom- crotonolactone. The rigorous proof of its lactone-structure, through its conversion into a salt of the corresponding oxy-acid, was rendered difficult by the presence of the halogen, since the halogen itself was rapidly removed in alkaline solution. By adding a decinormal solution of potassic hydrate to a cold dilute solution of the substance, it is true that it was easy to prove that very nearly two molecules of potassic hydrate were neutralized in the reaction, while but one molecule of potassic bromide was formed, but it was impossible to isolate definite products of the decomposition. On reducing the body with zinc and dilute sulphuric acid the bromine was completely removed, and on distillation a feebly acid solution was obtained, which on titration with decinormal potassic hydrate proved to contain a lactone. By the usual methods an amorphous barium salt was obtained, which when thoroughly dried had the percentage composition required by a baric oxycrotonate. The isolation of the crotonolactone itself proved to be a matter of such difficulty that we attempted to find some derivative which would be more manageable. The bodies containing the aniline residue in place of one of the halogen atoms were found to be useless for our purpose, since they were decomposed by boiling in alkaline solu- tion with the formation of phenyl isocyanide, and the phenylamido- crotonolactone formed from them by reduction was also decom- posed with the formation of aniline under the same conditions. It seemed probable that the corresponding derivatives containing the phenoxy group would prove to be much more stable. While we have not yet succeeded in replacing the bromine of the body C4H2Br203 directly by the phenoxy group, we have had no diffi- 192 Hill and Cornelison. culty in preparing such a body by the reduction of the brom- anhydride of the inucophenoxybromic acid which was described by Hill and Stevens.' The body thus formed was easily shown to be a phenoxybromcrotonolactone, since it dissolved in hot alkaline solutions with the formation of the salts of the corresponding oxy- acid. On acidifying the well cooled alkaline solution the phenoxy- bromoxycrotonic acid was obtained, which was stable under ordinary conditions, but which was again converted into the lactone by heat. These results were fully confirmed by a study ot the phenoxychlorcrotonolactone, which was prepared by the reduc- tion of mucophenoxychloryl bromide. The position which the phenoxy group takes in entering mucobromic acid has already been shown with sufficient precision by Hill and Stevens, since a remarkably stable phenoxybromacrylic acid is readily formed from this product by the action of alkalies. The phenoxylactones containing halogen must therefore have the constitution BrC— CH2 CIC— CH2 II >0, II >0; CeHfiOC— CO CeHsOC— CO and the analogous bodies containing halogen alone, the structure BrC— CH2 CIC— CH. II >o, II >o. BrC— CO CIC— CO The formation of these bodies from tribrom- and trichlorpyrc- mucic acids seems to us to be most readily explained by assuming that a trisubstituted furfuran is first formed with the loss of car- bonic dioxide : XC = C— COOH XC = CH I >0 =1 >0-fCO.. XC = CX XC = CX This then adds a molecule of haloid acid with the shifting of the double bond in a manner identical with that observed by Von Baeyer and Rupe* in the reduction of dichlormuconic acid : XC = CH H XC — CH= i >o + I = II >o. xc=cx X XC — ex. The addition-product is then decomposed by water, giving the 1 Proc. Am. Acad. 19, 262. ^Ann. Chem. (Liebig) 256, 25. Croionolactones and Mucobromic Acid. 193 lactone. Possibly a molecule of water is directly added to the trisubstituted furfuran, giving XC— CH2 II >o xc— cx OH which would at once pass into the lactone by the loss of haloid acid. The conversion of the/5;'-dibrompyromucicacid into the dibrom- crotonolactone, and of the /5;'-dichlorpyromucic acid into the cor- responding dichlorcrotonolactone by the action of bromine in aque- ous solution likewise finds its explanation in the formation of a tri- substituted furfuran. In aqueous acid or alkaline solution the car- boxyl of the substituted pyromucic acids is readily eliminated and replaced by bromine, as was first shown by Hill and Hartshorn' in the case of the o-brompyromucic acid : XC = C— COOH XC = CBr I >0 4- Br. = I >0 + CO2 -4- HBr . XC = CH XC = CH By the addition of water the substituted lactone is formed as before. We have shown that the bromanhydrides of mucobromic and mucochloric acids are also formed in the same reaction. This formation is evidently due to the addition of bromine to the trisub- stituted furfuran and the decomposition of this product by water : XC = CBr XC — CBr2 I >0 -f Brs = II >0, and XC = CH XC — CHBr XC — CBr^ XC — CO II >0 -f H2O = II >0-|-2HBr. XC— CHBr XC — CHBr It has already been said that two isomeric monosubstituted crotonolactones may be made by appropriate means from the disubstituted pyromucic acids containing the halogens in the /? and '5 positions. By heating these substituted pyromucic acids with mineral acids the brom- and chlorcrotonolactones melting at 58° and 26° respectively, may be made. The reaction is evidently perfectly analogous to the decomposition of the trisubstituted pyromucic acids under like conditions which has just been 1 Ber. d. chem. Ges. 18,448. 194 ///// a7id Coryielison. discussed. The carboxyl group is replaced by hydrogen with the loss of carbonic dioxide, and the substituted furfuran thus formed then passes into the lactone as before : XC = C — COOH XC = CH >o = I >o + co, CX HC = CX A: XC = CH XC — CH. I >0 + H.O = II >0 + HBr. HC = CX HC — CO The /^-substituted lactones which are thus formed may be made much more conveniently by the partial reduction of the disub- stituted lactones. In order to prepare the monohalogenized cro- tonolactones with the halogen in the ^-position from the /9(J- dichlor- and dibrompyromucic acids, derivatives of maleic acid must first be formed. The first step in the reaction is the forma- tion of a trisubstituted furfuran through the replacement of the carboxyl by bromine : XC = C — COOH XC = CBr I >0-fBr2 = I >0-i-C02H-HBr. HC=CX HC = CX Through the addition of bromine with the shift of the double bond, as before, XC = CBr XC — CBr. I >0 + Br. = II >0; HC = CX HC — CXBr a body is formed which through the action of water yields the derivative of maleic acid : XC — CBr. XC — CO II >0 + H20 = II >0 + 2HBr. HC — CXBr HC — CXBr This in its turn gives the lactone by reduction : XC — CO II >o. HC — CH2 In the case of the /?(5-dibrompyromucic acid, the tribromfurfuran, which should appear as the intermediate product in this reaction, has been isolated in a pure condition by Mr. W. M. Booth in this laboratory, and has been found to yield, when treated with bro- Crotonolactones and Miicobromic Acid. 195 mine in aqueous solution, the brommaleyl bromide described by- Hill and Sanger. The a-bromcrotonolactone may also be formed by the action of bromine in aqueous solution upon /3-brompyromucic acid. The reaction is evidently identical with that through which the dibrom- crotonolactone is formed from /S^'-dibrompyromucic acid. A bibromfurfuran is first formed, which by the fixation of water and the elimination of hydrobromic acid passes into the lactone; BrC = C — COOH BrC = CBr I >0 + Brs = I >0+CO: + HBr; HC = CH HC = CH BrC = CBr BrC — CO I >0 + H20 = II >0 + HBr. HC = CH HC— CH2 The isomerism observed in the monohalogenized crotonolac- tones must be due to the position of the halogen atoms, since, according to the views at present held concerning geometric isomerism, lactones can be formed from the maleinoid forms only. A definite proof that the halogen atoms in the a- and /5-bromcrotonolactones are attached to different carbon atoms is easily given. The a/3-dibromcrotonolactone yields with hydriodic acid an iodine derivative which can be reduced to /J-bromcrotonolactone. This same iodine compound gives with aniline a phenylamidobromcrotonolactone which on reduction is converted into the same phenylamidocrotonolactone that may be made by the action of aniline upon the a-bromCrotonolactone : BrC- -CHs BrC- -CH2 it -CO -> Hli- >o -co ~^ BrC- -CH. HC- -CH. C6H5. NHi- -c^° -^ C^Hs .NH.c!L >o -CO HC- -CH2 ^ B.!!- -cS° It is evident that the acid described by Hill and L. L. Jackson under the provisional name of the ;f-dichlorpyromucic acid, since it readily gives the «-chlorcrotonolactone on decomposition with Vol. XVI.-14. 196 Hill and Cornelison. \ mineral acids, must in fact be the ^5-dichlorpyromucic acid, and its structure I HC = C — COOH. .] I >o \ C1C = CC1 \ It is perhaps worthy of note that, although the /5^- and the yo- \ dichlorpyromucic acids give the two isomeric chlorcrotonolactones when heated with acids, they give the same a-chlorcrotonolactone when they are treated with bromine in aqueous solution, and the I resulting product it reduced. Evidently two isomeric dichlor- ' bromfurfurans are first formed, which yield the corresponding addition-products with bromine, but these addition-products are j attacked by water in such a way that in each case the chlorine is | left in the a-position with regard to the oxidized carbon : ClC = CBr CIC — CBr. CIC — CO I >0 -> II >0 -> II >0; ; HC = CC1 HC — CClBr HC — CClBr HC = CBr HC — CBn HC-CBr, II >o - II >o - II >o. CIC = CCl CIC -CClBr CIC = CO The close relationship between mucobromic acid and the bodies which have thus been shown to be derivatives of crotonolactone, naturally recalled the suggestions which had already been made as to the constitution of mucobromic acid itself. As early as 1882 Roser' pointed out that the so-called fumaric aldehyde-acid of Limpricht might in. reality be an oxylactone, ^^-^■• HC— CO and afterward, in 1887, Anschiitz,' in his interesting and sugges- tive discussion of the constitution of maleic and fumaric acids, was led to the conclusion that mucobromic acid was an oxy- dibromcrotonolactone : II y BrC — CO > Ber. d. chem. Ges. 15, 1523- ^ Ann. Chem. (Liebig) 339, i6i. Croionoladones and Mtuobromic Acid. 197 It was evident that the ahnost quantitative reduction of muco- bromyl bromide by such a reducing-agent as stannous chloride in the cold, was perfectly intelligible if the formula of this bromide were ll >■' BrC — CO while the formation of a lactone under these conditions from the bromanhydride derived from the aldehyde-acid, BrC-Co BrC — CO might well prove more refractory. Several years ago Lieberman' tried the behavior of the ethyl ester of opianic acid with hydroxyl- amine, and obtained only the opianic oxime anhydride. We have been able to find described no other experiments with the esters of the acids in question. Our own experiments with muco- bromic acid, and its methyl ester, showed that the ester was attacked with much more difficulty. While mucobromic acid is rapidly converted into its oxime and oxime anhydride, we failed to discover any appreciable action when hydroxylamine is added to a solution of methyl mucobromate in methyl alcohol at ordinary temperatures, even after the lapse of several days. On boiling, the methyl ester of mucobromoxime was formed, identical with the body formed by heating mucobromoxime with methyl alcohol. Further experiments in this direction can alone show whether the difference in behavior between the acids of this series and their esters is sufficiently general to warrant definite conclusions as to their structure. The action of phenylhydrazine upon mucobromic and mucochloric acids and their esters, which we hoped to study in detail, proved to yield bodies which were not suitable for investigation. Many years ago Mr. C. W. Andrews, who was at that time an assistant in this laboratory, made some experiments as to the action of aniline upon mucobromic acid and its ethyl ester, which he was unable to complete at the time, and which have never been published. He found that the reaction with mucobromic acid was complex, in that bromine was partially replaced by the aniline residue, and also that oxygen was eliminated by conden- sation. In marked contrast with this reaction was the behavior of ethyl mucobromate with aniline. One atom of bromine was ' Ber. d. chein. Ges. 19, 2926. 200 Hill and Cornelison. here replaced, but the rest of the molecule remained unaltered, so that the body C4H2Br(NHC6H5)03C2Hs was formed in nearly theoretical quantity. It therefore seemed to us of interest to determine whether ethyl mucobromate would behave in an analogous way with ammonia. We found, however, in this case, that the reaction took a different course, and that the ethoxy group was first attacked. The body which is thus formed has the formula of the amide of mucobromic acid, but its behavior is in some respects anomalous. It dissolves readily in caustic alk- alies, and may be reprecipitated unchanged by immediate acidifi- cation. It is but slowly converted into mucobromic acid by boiling with mineral acids, and with oxidizing agents it yields dibrommaleinimide. From mucochloric acid a body of similar properties was obtained. a/3-DlBROMCROTONOLACTONE. Although tribrompyromucic acid is little affected by boiling hydrochloric acid, it is quite readily decomposed, with the escape of carbonic dioxide, when heated to 130° with diluted sulphuric acid (sp. gr. 1.43), or when boiled with concentrated hydro- bromic acid. The reaction seemed to be more neatly effected with the latter reagent, and we therefore heated tribrompyro- mucic acid with from four to five times its weight of concen- trated hydrobromic acid with a return-condenser. When no further escape of carbonic dioxide could be detected, the some- what dark-colored clear solution was cooled, and diluted with water. A heavy crystalline precipitate was thus thrown down, which could easily be purified by recrystallization from alcohol or ligroin. The crude product obtained in this way amounted to about half the weight of the tribrompyromucic acid taken. Analyses of the body after several recrystallizations from alcohol showed that it had the formula CiHsBrsO.;. For the analyses of this dibromcrotonolactone, and of several of its derivatives, which we publish, we are indebted to Mr. H. N, Herman, who made a preliminary study of this body some two years ago, but was unable to continue the investigation. I. 0.2472 gram substance gave 0.1788 gram CO2 and 0.0209 gram H2O. II. 0.2394 gram substance gave 0.37 11 gram AgBr. Calculated foi C,H,Br,0,. c 19.83 H 0.83 Br 66.12 Croioyiolaciones and Mucobromic Acid. 201 Found, r. II. 19.73 0.94 65.96 This same body may also be made directly from mucobromic acid by the action of phosphorous iodide. Phosphorus is dissolved in five times its weight of carbonic disulphide, equivalent weights of iodine and mucobromic acid well ground together, are then added, and the nearly solid mass heated on the water-bath with reverse-cooler. The reaction sets in slowly and, frequently, only after a part of the carbonic disulphide has been allowed to escape through the cooler. With no solvent present the reaction is violent and yields little or no product. When the mass is dbm- pletely liquefied, the rest of the carbonic disulphide is distilled off, the flasked well cooled, and cold water added in not too small quantity. On shaking, the crude lactone separates in a granular condition. In this way it is easy to obtain a product which amounts to two-thirds of the weight of the mucobromic acid taken, but it contains a decided percentage of iodine, and the preparation of pure dibromcrotonolactone from it is a matter of great difficulty. By fractional crystallization from various solvents we found it impossible to eliminate the iodine, but on distillation with steam the substance first carried over contained but little, and the percentage of iodine did not become large until about one-half of the material taken had passed over. The material which had thus been partially purified by fractional distillation with steam, could not be further purified by recrystallization, but after boiling for some time with bromine-water the iodine com- pound was oxidized, and a substance was then obtained which after several recrystallizations from alcohol possessed the proper- ties and the composition of pure dibromcrotonolactone. 0.2439 gram substance gave 0.3779 gram AgBr. Calculated for C4H,Br20„. Found. Br 66.12 65.93 The dibromcrotonolactone can much more readily be made from mucobromic acid by the reduction of its bromanhydride. Mucobromyl bromide was described many years ago by Hill and O. R. Jackson,^ who made it by the action of phosphoric penta- 1 Proc. Am. Acad. 16, 174 202 Hill and Corneliso7i. bromide upon mucobromic acid. We have found that phos- phorous tribromide is in many respects more advantageous for its preparation. The reaction runs somewhat slowly at ioo°, but so smoothly that one molecule of the tribromide is sufficient for three molecules of mucobromic acid, and the yield is about 90 per cent, of the theoretical amount. After the reaction is over the flask is well cooled, cold water is added, and the whole vigorously shaken until the oil which first separates solidifies in a granular form. The crude product melts at 54°-55°, and by repeated recrystalliza- tion from small quantities of hot alcohol this melting-point may be raised to 56°-57°, one degree higher than the point given by Hill and O. R. Jackson. We attempted to purify the crude product by distillation in vacuo, but we found that the melting- point was depressed rather than raised by this treatment. We have not further studied the change which is thus apparently brought about by dii^tillation, for we soon found that a perfectly pure dibromcrotonolactone could be made directly from the mucobromyl bromide, as it was precipitated by water, while it was difficult, if not impossible, to do this with the distilled sub- stance of low melting-point. Hydriodic acid, or zinc-dust with glacial acetic acid, reduces the mucobromyl bromide to the lactone, although the former reagent yields a product which contains iodine. A far more efficient and convenient reducing- agent we found to be stannous chloride with hydrochloric acid, since this gives at once an essentially pure product in satisfactory quantity. Mucobromyl bromide is added to an equal weight of stannous chloride dissolved in the same amount of concentrated hydrochloric acid. The reduction proceeds rapidly with the evolution of heat, although in working with small quantities the reaction is greatly facilitated by warming gently at first. When the melting-point of the mucobromyl bromide is reached, the flask must be well shaken until the oil has completely disappeared. As the solution cools the dibromcrotonolactone crystallizes out, and still more separates on dilution. The yield which may be obtained in this way amounts to from 75 to 80 per cent, of the weight calculated from the mucobromyl bromide taken, or about 70 per cent, of that theoretically required by the mucobromic acid employed. A single recrystallization from alcohol is sufficient for its complete purification. Dibromcrotonolactone is very sparingly soluble in cold water, Crotonoladones and Mucobromic Acid. 203 and dissolves in from 30 to 40 times its weight of boiling water. As the hot aqueous solution cools, the lactone crystallizes in small stx-sided plates, or in pointed prisms crossing at an angle of 60°. From alcohol, in which it is somewhat sparingly soluble at ordinary temperatures, although very readily soluble on heating, it crystallizes in bundles of long friable prisms. From concen- trated nitric acid it crystallizes in large, clear, brilliant oblique prisms. In boiling ligroin it dissolves somewhat sparingly on boiling, and as the solution cools the greater part of it separates in finely felted needles. It is readily soluble in chloroform or benzol, more sparingly soluble in ether or carbonic bisulphide. It melts at 90°-9i°, and boils, under a pressure of 18 mm., at 145°. With steam it volatilizes rapidly. Although it is remarkably stable in acid solutions, it is easily attacked by alkalies. The alkaline hydrates dissolve it, forming deep yellow solutions, and at the same time alkaline bromides are formed. In studying the action of decinormal potassic hydrate in the cold, we have found that approximately two molecules of potassic hydrate are neutral- ized by each molecule of the lactone taken, while but one mole- cule of potassic bromide is formed ; but have been able to isolate no definite products of the reaction. The lactone was not attacked by hydroxylamine ; with aniline it yielded, with the elimination of bromine, the well crystallized aniline derivative which will be described later. Phenylhydrazine also removed bromine, but gave no well characterized product. Dibromfutnaric Aldehyde of Toennies. The melting-point and other properties of the dibromcrotono- lactone recalled a body of like composition which was obtained in 1879 by Toennies,' through the action of bromine-water upon /9^-dibrompyromucic acid. He gives no analytical data to support his formula, and apparently had but small quantities of material at his disposal. From its behavior he considered the body to be the aldehyde of mucobromic acid, or the double aldehyde of dibromfumaric acid. In 1886 Hill and Sanger' again prepared this body from y3^-dibrompyromucic acid, but were unable to obtain it in satisfactory quantities, and hoped to return to it at some future time. A further study of the reaction has shown us that it is by no means simple in its nature. While we have been ' Ber. d. chem. Ges. 13, 1202. - Proc. Am. Acad. »!, 172. 204 Hill and Coriielison. unable to obtain a satisfactory yield of the desired product, we have succeeded in preparing an amount amply sufficient for its identification, and at the same time we have isolated a second product of the reaction. Preliminary experiments showed us that it was most advantage- ous to add about 20 per cent, more than one molecule of bromine to the finely divided dibrompyromucic acid suspended in 20 times its weight of cold water. We therefore boiled 5 grams of pure /J^-dibrompyromucic acid with 100 cc. of water, and quickly cooled the solution in order to obtain the acid in a finely divided condition. When the temperature reached 16°, 1.2 cc. of bro- mine were added, and the whole well shaken. The color of the bromine gradually faded, but when it had completely disappeared a considerable amount of a well crystallized body remained undis- solved. Instead of allowing the solution to stand over night, as Hill and Sanger had done, we filtered out the insoluble substance at the end of three hours, and extracted the colorless aqueous solution thoroughly with ether. The ethereal extract left on distillation a syrupy residue, which partially solidified on standing- over night in vacuo over sulphuric acid. The crystals when thoroughly drained upon the pump and pressed weighed 0.85 gram. Two crystallizations from small quantities of hot alcohol yielded 0.65 gram of substance, which melted at 90°-9i°, and showed the characteristic behavior of the dibromcrotonolactone. In order to identify it with precision, we converted it into the a-iodo-/5-bromcrotonolactone, which is fully described later, by heating it with hydriodic acid, and found the product to melt at the proper point, Ii8°-ii9°. On reduction with zinc and acetic acid it yielded the /S-bromcrotonolactone melting at 57°-58°, and with aniline the a-phenylamido-/5-bromcrotonolactone with its characteristic properties, both of which bodies, as we shall after- wards show, may be made in the same way from the dibromcro- tonolactone. As we shall show more fully later, we were unable to confirm the statement of Toennies that the dibromcrotonolac- tone could easily be converted into mucobromic acid by oxida- tion with chromic acid. By the action of bromine in aqueous solution a part of the /9j'-dibrompyromucic acid had been con- verted into dibromcrotonolactone, according to the equation given by Toennies, aH^Br^On -f Br. + H.O = OH.Br.O. + CO= -f 2HBr. Crotonolactones and Mucobroniic Acid. 205 The viscous oil which had been drained from the crystalline product deposited a few more crystals of the lactone on long standing, but we have not yet examined it further. The insoluble matter, which had been removed by filtration before the extrac- tion with ether, was washed with a dilute solution of sodic carbonate. But a small amount dissolved, and upon acidification 0.2 gram of a sparingly soluble acid was precipitated, which melted at i89°-i90°, and was evidently unaltered dibrompyro- mucic acid. The residue, insoluble in the alkaline solution, weighed i gram, and consisted chiefly of mucobromyl bromide. It gave in alcoholic solution a deep blue evanescent color on the addition of sodic carbonate, and one crystallization from alcohol was sufficient to raise the melting-point to 56°-57°. A part of the dibrompyromucic acid had therefore been decomposed according to the equation OH4Br203 + 2Br2 -f H2O = CHBr^O^ -f CO2 -fsHBr. Mucobromyl bromide is so slowly attacked by water that little mucobromic acid could have been formed under the conditions chosen ; but it is found in abundance when the product of the reaction stands for a long time in dilute solution. The alcoholic mother-liquors obtained from the recrystallization of the muco- bromyl bromide were evaporated, and the residue boiled with water under a reverse-cooler until the bromanhydride had been converted into mucobromic acid. A small amount of insoluble material was then left, which had the odor of tetrabromfurfuran, but we were unable to identify it with precision. A preliminary experiment under essentially the same conditions had already yielded us a crystalline body, which was proved by its melting- point (64°), and other characters, to be tetrabromfurfuran. Hill and Sanger' had previously found that this substance was formed in considerable quantities when bromine was gradually added to the dibrompyromucic acid suspended in water. Action of Oxidizing Agents. As we already have said, dibromcrotonolactone maybe crystal- lized without alteration from concentrated nitric acid, but on long boiling it is slowly oxidized, and mucobromic acid or dibrom- ' Proc. Am. Acad. 31, 172. 2o6 Hill and Corjielison. maleic acid formed. Even fuming nitric acid attacks it with diffi- culty, and after boiling for half an hour the greater part of the lactone taken may be recovered unchanged. We dissolved the lactone in 8 times its weight of concentrated nitric acid (sp. gr. 1.42), and boiled the solution for three hours under a reverse- condenser. The unaltered lactone was then driven off with steam, and the acid solution evaporated to small volume. On cooling, mucobromic acid separated in abundance, which melted at 119°- 120°, after recrystallization from hot water. Even after boiling for six hours with concentrated nitric acid the oxidation was far from complete, but it was then easy to establish the formation of dibrommaleic acid through the melting-point (114°-! 15°) of its anhydride. On boiling the lactone with bromine- water, it is slowly converted into mucobromic acid. The reaction may be hastened greatly by using concentrated hydrobromic acid as a solvent, and the oxidation is then completed in a comparatively short time. The mucobromic acid which we obtained, when recrystallized from water, melted at I20°-I2i°. With chromic acid the action is also very slow, and we have been unable to identify any products of the oxidation except carbonic dioxide. If an amount of chromic acid was used which corresponded to one atom of oxygen for each molecule of the lactone, several hours at 100° were needed for the complete reduction ofthe chromic acid. On distillation large quanti- ties of the unaltered lactone were obtained (melting-point 90°-9i°), and the retort-residue, when evaporated to small volume, deposited nothing on cooling. This residue was therefore extracted with ether, the ethereal extract shaken with a dilute solution of sodic carbonate, and this alkaline solution immediately acidified with hydrochloric acid, and again extracted with ether. The ether then left upon evaporation a small syrupy residue, which was strongly acid to test-paper. After standing for several days, several minute clusters of rhombic plates could be seen under the microscope. In appearance they closely resembled mucobromic acid, but they were insufficient in quantity even for a melting-point determination. Since the dibromcrotonolactone reduced an ammo- niacal solution of argentic nitrate on heating, we made one attempt to effect its oxidation with argentic oxide. On boiling an aqueous solution of the lactone with a large excess of well washed argentic oxide, metallic silver was formed, but at the same time argentic bromide in large quantity. An approximate quantitative deter- Crotonoladones and Mucobromic Acid. 207 mination showed that 87.5 per cent, of the total bromine contained in the lactone had been converted into argentic bromide. After a careful search we failed to find any products of the oxidation except an amount of dibrommaleic acid which was just sufficient for its complete identification through the microscopic appearance of its barium salt and the melting-point of its anhydride, 1 14°-! 15°. Action of Bromine. Dibromcrotonolactone is not attacked by bromine at ordinary temperatures, but at 100° substitution is quite rapidly effected. If equal molecules of bromine and the lactone are taken, a colorless oil is obtained after several hours' heating, which on standing gradually solidifies, and the crystalline solid may be purified by recrystallization from small quantities of alcohol. The substance which was thus obtained melted at 56°-57°, and gave with alkalies in alcoholic solution the characteristic blue color described by Hill and O. R. Jackson.' Mucobromyl bromide had therefore been formed from the dibromcrotonolactone according to the equation C4H.Br202 + Br2 = C4HBr302 + HBr. When heated with an excess of bromine further substitution is effected and a product is obtained in which both the hydrogen atoms of the dibromcrotonolactone are replaced. For the prep- aration of this body we heated equal molecules of mucobromyl bromide and bromine in a sealed tube at 125°-! 30°. At this temperature the reaction proceeds rapidly, but it is also completed at 100° on longer heating. The colorless oil which we obtained gave no blue color in alcoholic solution with sodic carbonate, and on standing gradually solidified. The carefully pressed solid proved to be extremely soluble in alcohol, ether, chloroform, benzol, or carbonic disulphide, but it was somewhat more spar- ingly soluble in ligroin and could be purified by recrystalliza- tion from this solvent. The percentage of bromine which the substance contained corresponded to the formula C«Br<0«. I. 0.1685 gram substance gave 0.3177 gram AgBr. II. 0.1884 gram substance gave 0.3552 gram AgBr. Calculated for Found. C^Br^O,. I. II. 80.00 80.23 80.: Br ' Proc. Am. Acad. 16, 175. 2o8 Hill and Cornelison. This substance crystallizes from ligroin in clustered leaflets which melt at sS^'-sg". It has a strong suffocating odor like that of the acid bromanhydrides. When heated with water it is slowly dissolved and the solution then contains dibrommaleic acid. The acid was as usual identified by the crystalline form of the barium salt, and by the melting-point (114°-! 15°) of its anhydride. On treating the body with stannous chloride and hydrochloric acid the dibromcrotonolactone is again formed by reduction. The mode of formation, the composition, and the behavior of this body justify the conclusion that it is the unsymmetrical form of dibrom- maleyl bromide. Its behavior will be further studied in this laboratory. We have made many unsuccessful attempts to pre- pare an identical or an isomeric body directly from dibrommaleic acid. We have tried the action of phosphoric pentabromide upon dibrommaleic anhydride at temperatures which varied from 100° to 225°, and either obtained the unaltered anhydride, or else carbonization ensued. We were equally unsuccessful in our attempts to prepare such a product through the salts of the acid. Action of A^iiline. Aniline reacts upon the a/S'-dibromcrotonolactone at ordinary temperatures and forms w-phenylamido-zJ-bromcrotonolactone. This substance is most readily prepared by dissolving the lactone in twenty parts of alcohol, diluting this solution with an equal weight of water, and then adding somewhat more than two mole- cules of aniline. On standing, long needles of the aniline deriva- tive separate in abundance, which may be purified by recrystal- lization from 60-per cent, acetic acid, and afterward from alcohol. I. 0.2570 gram substance gave 0.4426 gram CO2 and 0.0763 gram H2O. II. 0.2393 gram substance gave 0.1775 gram AgBr. III. 0.2727 gram substance gave 13.4 cc. moist nitrogen at 21° and 763 mm. Calculated for Found. C.oHgBrNOj. I. II. c • 47.24 46.97 ... H 3.15 3-30 ... Br 31-50 31-56 N 5.51 ... 5-63 a-Phenylamido-/5-bromcrotonolactone is almost insoluble in ether, carbonic disulphide, or ligroin, but dissolves somewhat Crotonolactones and Mucobromic Acid. 209 more readily in boiling benzol or chloroform. It is somewhat sparingly soluble even in boiling alcohol, still less freely in boiling water, and as the solutions cool is deposited in each case in long briUiant needles. In hot glacial acetic acid the substance dis- solves readily ; from chloroform it crystallizes in thin transparent plates. When slowly heated it melts with decomposition at about 165°, but if the capillary tube containing the substance is plunged into the heated bath it melts promptly at i86°-i87° and immediately decomposes. In alkaline solutions it dissolves readily on warming, and if the solution be quickly cooled it crys- tallizes out apparently unchanged, but on heating for a longer time decomposition ensues, as the strong odor of phenyl iso- cyanide shows. By the action of sodium amalgam bromine is removed and a-phenylamidocrotonolactone, which we shall describe later, is formed. We have been unable to replace the second atom of bromine by heating with an excess of aniline. Action of Hydriodic Acid. When a/9-dibromcrotonolactone is dissolved in ordinary dis- tillable hydriodic acid and the solution is boiled for a short time, or heated for a longer time at 100°, one atom of bromine is replaced by iodine, and as the solution cools the a-iodo-/3-brom- crotonolactone separates in long prisms. This same body is also formed by the action of phosphorous iodide upon mucobromic acid or when hydriodic acid acts upon mucobromyl bromide, and it may easily be made the chief product of the latter reaction. For analysis it was recrystallized from alcohol. I. 0.2393 gram substance gave 0.1453 grairi CO2 and 0.0229 gram H2O. II. 0.1874 gram substance gave 0.2747 gram AgBr ■\- Agl. III. 0,2031 gram substance gave 0.2978 gram AgBr -|- Agl. c Calculated for C^HjBrlOj. 16.61 I. 16.56 Found. II. H Br+I 0.69 71.63 1.05 71.76 71.78 The a-iodo-/5-bromcrotonolactone is quite readily soluble in benzol or chloroform, more sparingly soluble in ether, carbonic disulphide, or ligroin. It dissolves freely in boiling alcohol, and 2IO Hill and Cornelison. as the solution cools is deposited in quite large, oblique, colorless prisms, which gradually change color on exposure. With steam it distils with difficulty. The melting-point of most preparations of this substance we have found to be constant at US'*-! 19°. Still we were frequently unable to raise by recrystallization alone the melting-point of material which melted at too low a tempera- ture, and on one occasion a preparation made in the usual way melted one degree higher, at 119°-! 20°. Toward aqueous alk- alies this substance behaves like the corresponding bromine derivative, an alkaline iodide being formed in the decomposition. Concentrated nitric acid or bromine-water on boiling liberates iodine. When heated with one molecule of dry bromine in a sealed tube at 100°, iodine is also liberated, and but a small quantity of hydrobromic acid is formed. With aniline in dilute alcoholic solution it yields the a-phenylamido-z^-bromcrotonolactone which has already been described, although not quite as smoothly as the dibromlactone, since dark-colored viscous products, which we have not further examined, are formed at the same time. On reduction with zinc-dust and glacial acetic acid, the /5-brom- crotonolactone melting at 58°, which we shall presently describe, was formed. After dilution with water the acid solution was extracted with ether, the ethereal extract washed with a dilute solution of sodic carbonate, and the crystalline residue obtained by the evaporation of the ether recrystallized several times from small quantities of alcohol. The body thus obtained had the properties of the /S-bromcrotonolactone, melted at 57°-58°, con- tained but an unweighable trace of iodine, and gave on analysis the required percentage of bromine : 0.2078 gram substance gave 0.2407 gram AgBr. Calculated for C4H3Br03. Found. Br 49.08 49-27 When the iodobromcrotonolactone was boiled with hydriodic acid the separation of iodine soon ensued. After long boiling with the addition of red phosphorus, although traces of volatile products which had the odor of fat acids had been formed, the main product of the reduction was a viscous oily body, which could be extracted by ether from the diluted solution, but from which we were unable to prepare any material suitable for analysis. With distillable hydriodic acid and red phosphorus Croio7iolacto7ies and Mucobromic Acid. 211 in sealed tubes at temperatures below 180° we obtained substan- tially the same result, while prolonged heating at 200° brought about decomposition. We had no better success when we employed hydriodic acid saturated at 0°, although in this case decomposition set in at a lower temperature. /3-Bromcrotonolactone. When /?r5-dibrompyromucic acid is boiled with concentrated hydrobromic acid, it is slowly decomposed, carbonic dioxide is evolved, and /?-bromcrotonolactone is formed. The reaction does not run as smoothly as it does with tribrompyromucic acid, and more or less carbonization ensues. When the reaction appears to be completed, the dilution of the dark brown solution usually precipitates a small amount of dark-colored unaltered acid, but no lactone is thus thrown down. The filtered solution is then extracted with ether, the ethereal extract washed with a dilute solution of sodic carbonate, and dried with calcic chloride. The residue left after distilling off the ether gradually solidifies on standing, and the crude product, which amounts to about one quarter of the dibrompyromucic acid taken, may be recrystallized from small quantities of alcohol or from ether. The same body may be prepared much more conveniently by the reduction of the a/3'-dibromcr©tonolactone with zinc-dust and acetic acid. The dibromcrotonolactone is suspended in its own weight of 80-per cent, acetic acid, and somewhat more than the calculated weight of zinc-dust is then added with careful cooling. At first the I'educ- tion proceeds rapidly with the evolution of heat, but several hours at ordinary temperatures are necessary for its completion. When the zinc has nearly disappeared the viscous solution is warmed, filtered, and the clear filtrate cautiously diluted with water. The bronilactone is thus precipitated as an oil, which after cooling and shaking soon solidifies in the form of colorless feathery crystals. The precipitated lactone amounts to 45 per cent, of the weight of the dibromcrotonolactone taken, and some- what more may be obtained by extracting the mother-liquor with ether. I. 0.2417 gram substance gave 0.2609 gram CO2 and 0.0440 gram H2O. II. 0.1913 gram substance gave 0.2214 gram AgBr. Vol. XVI.-is. Hill and Cornelison. Calculated for C,H,BrO,. I. c 2945 29.44 H 1.86 2.02 Br 49.08 49.24 The y3-bromcrotonolactone is readily soluble in alcohol, chloro- form, benzol or carbonic disulphide. It is decidedly less soluble in ether, and very sparingly soluble in ligroin. From small quan- tities of alcohol it crystallizes in colorless clustered prisms, from ether by slow evaporation in large transparent six-sided plates. It is quite readily soluble in hot water, and as the hot aqueous solution cools it separates in clear obliquely truncated prisms. It melts at 58°, and boils, under a pressure of 18 mm., at 140°. It distils with steam, although with some difificulty. Aqueous alk- alies dissolve it with the formation of a deep yellow color, and the solution then contains an alkaline bromide. In dilute alcoholic solution aniline also removes bromine, but forms at the same time a dark-colored viscous oil, from which we have been able to isolate no crystalline product. Action of Bromine. /9-Bromcrotonolactone is but slowly attacked by bromine in the cold. If one molecule of bromine is added to the powdered lactone, a clear deep red solution is soon obtained, but the color fades so slowly that several days are required to complete the reaction at ordinary temperatures. On opening the tube a small quantity of hydrobromic acid escaped, but the weight of the crystalline product was substantially the same as that of the mat- erials employed. After several recrystallizations from alcohol an analysis also showed that the substance had been formed by the addition of bromine: 0.2965 gram substance gave 0.5175 gram AgBr. Calculated for C^HaBrjOa. Found. Br 74.30 74.27 This body, which from the mode of its formation must be the ai32-tribrombutyrolactone, is readily soluble in alcohol, ether, benzol, carbonic disulphide, or chloroform, but is more sparingly soluble in ligroin. From alcohol it is deposited in large, well-formed, brilliant prisms, which melt at 63°-64°. On boiling with water a Crotonoladones and Mucobromic Acid. 213 part distils unchanged, but decomposition soon sets in, carbonic dioxide is evolved, and hydrobromic acid is formed. If the /5-bromcrotonolactone is heated to ico° with one molecule of bromine, the color of the bromine soon disappears, but at the same time hydrobromic acid is formed in considerable quantities. With an excess of bromine at 100° mucobromyl bromide is formed in abundance. The product which we obtained melted at 56°-57°, gave in alcoholic solution a deep blue color with sodic carbonate, and when heated with water yielded mucobromic acid. Action of Oxidizing -Agents. We have been unable to obtain any characteristic products by the oxidation of the /5-bromcrotonolactone, On boiling with con- centrated nitric acid oxidation takes place, but carbonic dioxide is evolved in abundance, and no other definite products were isolated. When heated with bromine in aqueous solution the addition- product is apparently formed at first, but on boiling this is soon broken up with the evolution of carbonic dioxide. We also noticed in this case the formation of a small amount of a highly crystaUine body which on recrystallization from alcohol formed long lustrous prisms. The melting-point (54°), the camphor-like odor, and other physical properties, render it extremely probable that this substance was pentabromethane, but it was insufficient in quantity for analysis. The aqueous solution upon evaporation gave a small viscous residue, from which on long standing a few microscopic rhombic plates separated which appeared to be mucobromic acid. [TV be continued.'\ 214 Linebarger. AN ISOTHERMAL CURVE OF SOLUBILITY OF MERCURIC AND SODIUM CHLORIDES IN ACETIC ETHER. By C. E. Linebarger. At the conclusion of a paper " On the Existence of Double Salts in Solution,'" it was stated that the author's intention was to make a study of all the states of equilibrium of the system : corrosive sublimate, common salt, acetic ether. It is found to be impossible, however, to complete the investigation. The results that have already been obtained are here given. The first point to be determined was the solubility of each of the salts in the ether. Sodium chloride was found to be but slightly soluble : lOO molecules of the liquid dissolving 0.031 molecule of the salt at 17°, and 0.037 molecule at 40°. Its solu- bility being so slight may be neglected. The solubility of mer- curic chloride is given in Table I. Table I. Solubility of Mercuric Chloride i7i Acetic Ether. Temperatures 0° 13° 30° 40.5° 50.2° Molecules HgCl, to 100 I ^^ ^ .^^ ^^^ ,g , ,5^ mols. L4H80-2 ) As the table shows, the solubility varies but slightly with the temperature. All of the subsequent experiments were carried out at about 40° C; the temperature did not vary more than 0.05° during an experiment. At 40°, 100 molecules of acetic ether dissolve 16.0 molecules of mercuric chloride, or 0.037 molecule of common salt. If to the saturated solution of the first salt, sufficient quantities of the second be added, it enters into solution, and the resultant solution is able to dissolve still more of the first salt. The ether in con- tact with an excess of each of the salts forms a solution in which are 40 molecules of corrosive sublimate and 20 molecules of sodium chloride to 100 molecules of the solvent. In general, 1 This Journal 15, 337. Mercuric and Soditim Chlorides in Acetic Ether. 215 when a solution of mercuric chloride in acetic ether of any con- centration possible is brought into contact with an excess of sodium chloride, a solution results in which twice as many mole- cules of the former salt are present as molecules of the latter, that is, the double salt (HgClOs.NaCl is formed. But what will be the composition of the solution if there be a deficiency of sodium chloride? To get an answer to this question, the following series of experiments were performed: A solution saturated with both the salts was put in a flask set in a thermostat, and a little more acetic ether, together with a large excess of corrosive sublimate, was added. This was frequently shaken for about 12 hours, and then a sample of the clear solution was drawn off and analyzed. More ether was added to the contents of the flask, care being taken that a large excess of mercuric chloride was always present. In this way, solutions saturated with mercuric chloride, but con- taining less and less sodium chloride, were obtained. As a " con- trol," several solutions were prepared with weighed quantities of the constituents. The results obtained with them were entirely concordant with those obtained by the other method. The com- positions of the solutions analyzed are given in Table II. Table II. Solubility of Sodium Chloride in Solutions of Mercuric Chloride in Acetic Ether. 100 molecules acetic ether dissolve : Mols. HgClj. Mols. NaCl. Mols. HgClj. Mols. NaCl. 40.0 20.0 18.0 5-1 38.1 19.6 16.4 4.3 36.0 19.2 14. 1 3.8 34.9 18.5 13.2 2.9 34-8 18.3 12.4 2.3 32.1 13.8 12.0 1.6 28.0 9.1 12.2 1-3 22.8 7.0 12.9 0.8 22.9 7.0 The results are shown graphically in Fig. i, where the axis of abscissae represents the number of molecules of mercuric chloride, and that of ordinates the number of molecules of sodium chloride to 100 molecules of acetic ether. 2i6 Linebarger. Fig. I. An inspection of the curve shows that the addition of a little sodium chloride diminishes the solubility of mercuric chloride. The maximum of diminution is reached when two molecules of common salt have gone into solution. The solubility of both the salts then increases at first slowly, then more rapidly. The remarkable thing about the curve is that for the values of the abscissae between 12 and 16, there are two values of the ordinates, or, in other words, two saturated solutions of common salt in corrosive sublimate solutions are possible for a certain range of concentrations. THE BENZOYL HALOGEN AMIDES. By C. E. Linebarger. The halogen amides of acid radicles have been studied princi- pally by Hofmann,' Bender,^ and Hoogewerff and van Dorp.' In these compounds the proximity of the nitrogen and halogen 1 Ber. d. chem. Ges. 14, 2725; l.'i, 407, 752, 977. -Ibid. 19, 2272. ' Recueil trav. chim. 6, 373 ; 8, 173 ; 9, 33 ; 10, 4. The Benzoyl Halogen Amides. 217 atoms, the latter of which are generally so capable of reaction, seems to indicate that certain important syntheses may be effected by their aid. Bender^ has sought, indeed, to substitute certain groups of atoms in the place of the chlorine atom in benzoyl chloramide, but his efforts were not successful. Thinking that perhaps the bromine atom of benzoyl bromamide would prove more tractable, the author has performed certain experiments upon that compound. The experiments did not, however, give the desired result, benzamide being the only crystallizable com- pound obtained. Although the results in general were negative, still in the course of the work facts were obtained in regard to the preparation and properties of the halogen amides which it may be well to make known. I. Benzoyl bromamide. — After a number of experiments upon the preparation of this substance, that which gave the best results was found to be the following : One molecule (80 grams) bromine is added to one molecule (60 grams) finely powdered benzamide contained in a large flask set in cold water. The amounts given in parentheses are convenient working-quantities, but it seems to make no difference in the yield if other quantities be taken. But little heat is evolved on the addition of the bromine. The reddish- brown mass that results must be well stirred together with a glass rod, in order that no benzamide may remain unattacked by the halogen. A solution of a little more than one molecule (30 grams) of caustic potash in 700-800 grams water is now added in small portions, the flask being well shaken after each addition and care being taken that no rise of temperature ensues. The resultant yellowish solution is filtered by the aid of a suction-pump from the solid product, and the latter is washed free from alkali with small quantities of water. The compound is slightly soluble in alkaline solutions, for on dilution of the first filtrate by the wash- water, a small quantity of a light flaky precipitate separates out. This is filtered off, washed and added to the main portion of product. The united masses are dried thoroughly by pression between folds of filter-paper, and dissolved in boiling benzene, from which on cooling it crystallizes in small plates. About 90 per cent, of the theoretical yield is thus obtained. A determina- tion of the nitrogen gave 7.14 per cent., instead of the theoretical 7.00 per cent. Benzoyl bromamide melts at 171°, with slight decomposition. It is soluble in alcohol, and in ether and alcohol, crystallizing best from the latter solvent. 2i8 De Chalmot. II. Benzoyl chlor amide. — As stated above, this compound has already been prepared by Bender. His method consisted in the action of chloride of lime in concentrated aqueous solution on benzamide dissolved in water acidified with acetic acid. It may also be easily obtained by the action of hydrochloric acid on benzoyl bromamide. When the strong acid is added to the bromamide contained in an evaporating-dish, the mass turns at first yellow and then red, from separation of bromine, which is expelled by heating on a water-bath. The residue is filtered and washed free from acid with water. It is then dissolved in boiling water and the solution concentrated until, on cooling, a crop of fine soft crystals separates out. By concentration of the mother- liquor, new crops may be obtained. The bromamide can in this way be almost quantitatively converted into the chloramide. Benzoyl chloramide melts at 113°, is but shghtly soluble in cold water, more so in boiling water; it is almost insoluble in alcohol or benzene even at their boiling-points. III. Benzoyl iodoaviide cannot be obtained by the application of the methods employed in the preparation of the brom- and chloramides. Only benzamide is obtained when iodine and benzamide are melted in a closed vessel and the resultant product treated with' caustic-potash solution. It is doubtful whether the substance is capable of existence. PENTOSANS IN PLANTS.* By G. DE Chalmot. I. — Notes concerning the Quantitative Estimation of Pentosans. I have used throughout with a slight alteration the method originally advised by Tollens and myself and somewhat modified by Flint.'^ Since this method was never fully published in an American journal, I shall describe it in brief here. 5 grams of the material to be analyzed are distilled with hydro- chloric acid (of i2-per cent. HCl), with constant renewal of that > This Journal 15, 21, 276 ; J. Am. Chem. Soc. 15, 618. 2 Landw. Vers. Stat. 48, 39S5. Peyitosans in Plants. 219 lost by distillation, until the distillate no longer shows the furfurol- reaction with aniline acetate. The distillation should be so regu- lated that 30 cc. of distillate are obtained in 10 to 15 minutes. The total volume of the distillate is measured, and after neutralizing with bicarbonate of sodium and slightly acidifying with acetic acid, enough sodium chloride is added for the liquid to contain about 16.3 per cent, of sodium chloride after its volume has been brought to 500 cc. by adding distilled water. A table in the original paper of Flint mentions the quantities of sodium chloride to be added to a certain volume of distillate. This addition of sodium chloride is made in order to bring the solution to a standard strength, and is necessary, since furfurol phenylhydrazone is less soluble when the sodium chloride solution is stronger. Phenylhydracine acetate is now added in order to precipitate the furfurol in the form of hydrazone.' The liquid is stirred for half an hour and filtered, and the hydrazone washed with about 100 cc. of distilled water. For filtering the same tubes are used as in Allihn's gravimetric method for determining glucose, the filter being of glass-wool. For drying the hydrazone the tubes are heated in an air-bath to 50°-6o° C. and are connected with an air- pump. The air is exhausted, and a slow current of dried air is passed through. The hydrazone will dry in li to 3 hours, accord- ing to the power of the pump. After weighing the tubes, the hydrazone is dissolved in warm alcohol and the tubes are weighed again. The difference between the weighings indicates the amount of hydrazone. The amount of furfurol phenylhydrazone (CuHioNaO^: 170) multiplied by y^\ or 0.516, indicates the amount of furfurol (C5H403=:96) found. Since, however, a small amount of furfurol phenylhydrazone is soluble in the sodium-chloride solution, ToUens and Flint multiply by 0.538.^ Empiric formulae are used for calculating the percentages of pentoses (or of pentosans) in the original substances. My slight modification of this method is, that instead of measur- ing the distillate and adding to it a calculated amount of sodium chloride, I at once bring the volume of the neutralized liquid to 500 cc. by adding a 19.3-per cent, sodium-chloride solution 1 I usually add 20 cc. of a solution containing 5 per cent, of phenylhydracine hydrochloride and 5 per cent, of sodium acetate. 2 This figure is in most instances too high. 220 De Chalmot. (19.3-per cent. NaCl=i2-per cent. HCl). I neglect the small amount of furfurol phenylhydrazone that is soluble in this solution, I further assume that pentosans yield 50 per cent, of furfurol. These pentosans may then be xylan, araban, riban ' (adonan ?) or others. I am warranted in doing so, for I do not care to obtain absolute figures, but only the relation between the amounts of pentosan in similar vegetable matters. For example, I compare the one wood with the other, but seldom wood with leaves or seeds. I no longer use the method of slow distillation" even for seeds, for the results obtained by it are perceptibly lower than those obtained by a quick distillation, A thoroughly mixed and very finely pulverized sample of oak leaves yielded by slow distillation 3.67 per cent, of furfurol =7.34 per cent, of pentosan, as the average of 8 analyses. The same sample yielded by quick dis- tillation 4.57 per cent, of furfurol ==9.14 per cent, of pentosan as the average of 4 analyses. Where great accuracy is necessary, I make the analyses of the substances to be compared, side by side. In this manner I was able to prove that pentosans are not formed by the assimilation- process.^ In the cases where the analyses have been made side by side, I have marked them both with x or with xx. As by distillation of vegetable matter with hydrochloric acid not only furfurol, but also methyl furfurol and other aldehydes are produced, the insoluble hydrazone obtained by adding phenyl- hydracine to the distillates is not pure furfurol phenylhydrazone. In the instance of some hard woods the hydrazone contains less impurities than where oak leaves or seeds have been used. If the hydrazone contains more impurities, its melting-point is lower, and it decomposes more readily, for which reason the results obtained by the analyses become less constant, unless great care is observed. I experienced this especially when estimating pen- tosan in corn seed. The hydrazone that I obtained could stand at the utmost a temperature of 55° C. It was therefore espe- cially necessary completely to exhaust the air in the tubes in order to dry the hydrazone quickly. Since the composition of the hydrazone varies according to the 1 E. Fischer : Ber. d. chem. Ges. 36, 633. "^ This Journal 15, 276. 5 J. Am. Chem. Soc. /. c. Pentosans in Plants. 221 material from which it is derived, its estimation does not give an absolute measure of the quantities of pentosan in the material. But since the larger part of those hydrazones that can stand heat- ing at 55° C. consists of furfurol hydrazone, we can safely use the gravimetric method for comparing the amounts of pentosan in similar material. II. — The Changes in the Percejitage of Pentosan duriyig the Growth of the Organs of the Plant. In a former investigation' I have found that the percentage of pentosan in clover hay increases from 9.2 per cent, at the begin- ning of the flowering to 10.5 per cent, at the end of this period. Stone obtained similar results.^ He found that the pentosans in Phleum pratense increased from 15.65 to 16.17 per cent., and in another case from 12.59 to 14.26. Since, however, these results were not obtained with a view to the subject now under discussion, I have investigated it anew and now communicate the results obtained. Experiment i. — Young corn plants, of which the tassels were still in the earliest stages of development, were investigated. Samples were made of: I. The tops of the stems, including the youngest undeveloped leaves and the undeveloped tassels. II. Full-grown parts of leaves picked at i p. m.^ III. The lowest two internodes of the stem without sheaths of leaves. I. II. III. Percentage of pentosan* 5.23 16.64 12.65 Experiment 2. — Red-oak leaves picked from the same shrubs at I p. m. of the same day contained : Percentage of Pentosan. Young leaves 5.29 Fully developed leaves 9.77 In both instances the full-grown organs of the plants contained much more pentosan than the organs still in development. I now investigated whether the percentage of pentosan increases in full-grown organs. 1 Author's Inaug. Diss., p. 33. * Agric, Sci. 7, 9. 'Leaves contain more pentosan in the morning than later on in the day. Vide J. Am. Chem. Soc. 15, 618. < The percentages given are relative to dry matter. 222 De Chalniot. Experiment 3. — Full-grown white-oak leaves contained on the morning of July 7, 7.16 per cent, of pentosan. Leaves of the same shrubs and branches contained in the middle of October, when they were still perfectly sound and healthy, 9.36 per cent, of pentosan. Experiment 4. — Full-grown white-oak leaves contained on the morning of July 13, 8.30 per cent, of pentosan. Perfectly healthy leaves of the same shrubs contained in the middle of October, 10.30X per cent, of pentosan. Leaves of which the color had turned into red, picked at the same time as the former, and from the same shrubs, contained 11.30X per cent, of pentosan. The pentosan in the living oak leaves therefore still increases after they are full-grown. In Experiment 4 we see that also after the death of the leaves the percentage of pentosan increases. This can be due to one or more of the following reasons : 1. A decomposition or washing-out of other substances (not pentosan) of the leaves. 2. A removal of other substances out of the leaves into the stems. 3. A formation of pentosan out of other substances after death. If the second reason was the only one, a rapid removal of sub- stances out of the leaves would probably take place during the last days before death, for the green and living leaves, picked at the same time as the red leaves, died soon afterwards. Such a quick removal seems to be contradicted by the following result, not obtained, however, with oak leaves. Experirnent 5. — Leaves of Madura aurayitiaca picked on October 22, perfectly green and sound, contained 8.04X per cent, of pentosan. Leaves off the same limb alternating with the former were picked on October 29, when the green color was fading in several spots, 8.09X per cent, of pentosan. The excess of pentosan in dead leaves is therefore most likely not alone due to a removal of other substances. That pentosans are formed after death seems most improbable, on taking into account the results obtained with woods, reported farther below. I therefore take for granted that the pentosans in the leaves are less subject to putrefaction and leaching out than other leaf-substances, and besides are not removed to any extent Pentosans in Pla^its. 223 out of the leaves before death. In leaves that decompose more quickly than oak leaves — for example, corn leaves — the difference is more apparent. Experiment 6. — Green and healthy corn leaves picked at 6 p. m., 14.5 per cent of pentosan. Dead leaves of the same plants picked at the same time, 24.6 per cent, of pentosan. In other organs than leaves I also found the percentage of pen- tosan to increase with age, as, for example, in corn-cobs and oak wood. Experiment 7. — Axil part of corn-cobs of which the seeds were in the earliest stages of development, 29.3 per cent, of pentosan. Axil part of corn-cobs of which the seeds were ripe, picked from the same lot of plants as the former, 33.3 per cent, of pentosan. Experiment 8. — Wood of Qiiercus nigra of i to 2 years old, including unfinished young wood, collected in the beginning of August from a standing tree, 17.3X per cent, of pentosan. Wood of 10 and more years old, collected at the same time off the same tree, 20.3X per cent, of pentosan. According to Experiment 8, the youngest unfinished wood con- tains less pentosan than when its formation is accomplished. Wieler' found in young wood of Pinus sylvestris 15.24 and 14.05 per cent, of wood-gum. Since pine wood does not contain more than 10 per cent, of pentosan, Wieler's result seems to indicate that the young unfinished wood of this tree contains more pento- san than the wood of which the formation is accomplished. This would contradict my result above. It is, however, probable that wood-gum from pine wood contains much mannan, since Tollens and Lindsay^ found mannose in an extract of pine wood which ' had been obtained with sodium bisulphite in the paper factories. I shall investigate this subject next summer. III. — Examination of Woods. Since wood is one of the vegetable matters most rich in pen- tosan, the study of different woods promised to reveal some facts as to the physiological significance of the pentosans. In the following list I have given the percentages of pentosan which I found in wood of different trees : > Landw. Vers. Stat. 33, 307. "- Ann. Chem. (Liebig) 267, 341. 224 De Chalmoi. Family. Magnoliacea Per cent, of Origin. Pentosan. Virginia collection' 19.1 Magitoliacece Rosacea See Exp. 11 Virginia collection 17-7 19.7 Legutninosce Sapindacea See Exp. 18 Virginia collection 21. 1 22.1 Hamainelidece Virginia collection 21. 1 Ilicinetz Virginia collection 24.6 Vitacece Virginia collection 20.9 CorfuicecE Virginia collection 21.6 Coniactce Virginia collection 20.8 OleacecE Virginia collection 17-5 Oleacece See Exp. 16 17.4 Juglajtdacea jfuglajidacea Salicinea: Virginia collection Virginia collection Virginia collection 19.2 21.0 Cupulifera Cupulifera Cupulifera Virginia collection Virginia collection Virginia collection 23-4 21.0 21.6 Cupulifera Cupulifera Cupulifera Urticacea Virginia collection Virginia collection Virginia collection Virginia collection 20.4 21.7 21.3 17.4 Plat ana Virginia collection 21.6 Conifera Conifera Virginia collection White pine timber 10.4 7-5 Conifera Yellow pine timber 8.8 Conifera 20-year old Christ- mas-tree 6.0 Species. Liriodendron tulipifera Magnolia acuminata Prwius Pennsylvanica Cercis Canadense Acer dasycarpum Liquidamhar Styraciflua Ilex opaca Ampelopsis quinquefolia Cornus florida Nyssa sylvatica Fraxinus Americana Fraxinus platicarpa Juglans cinerea Carya alba Salix spec. Betula spec. Fag us ferruginea Quercus Phellos Quercus alba Quercus rubra Quercus nigra Ulmus Americana Plaianus occidentalis yuniperus Virginiana Pinus strobus Pinus mitts Tsuga Canadensis Out of these figures I have not been able to trace any relation between the percentage of pentosan in wood and its hardness, durability, toughness or other qualities, — which of course does not imply that such relations cannot exist. The only regularity observed is that wood of Conifercs contains much less pentosan than that of Dicotyledonece, The former contains only 10 per cent., or less, of pentosan ; the latter from 17 per cent, to 25 per cent. Analyses which I made some years ago of German woods* confirm this conclusion. I found : Furfurol. Pentosan. Fagus sylvatica 12.6 25.2 per cent. Picea excelsa 5-0 lO.O " 1 " Virginia collection " means a collection of Virginia woods in blocks exhibited in New Orleans in 1885, samples of which were kindly put at my disposition by Colonel Whitehead, Commissioner of Agriculture in Virginia. ^Author's Inaug. Diss. p. 39. Pentosans in Plants. 225 These figures cannot be directly compared with the former, for they have been obtained by a somewhat different method. It was of consequence to know whether the amount of pentosan increases in wood after this is ready made. In order to investi- gate this I made use of discs of wood also belonging to the collection already mentioned. I have collected wood of different ages from the same discs by drilling holes in it with a brace and bit. The holes had a diameter of i cm. Experiment 9. — With Quercus nigra. I. Wood of 2-12 years, 21. i per cent, of pentosan. II. Wood of 69 years, 21.0 per cent, of pentosan. III. Wood of 109-110 years, 20.6 per. cent, of pentosan. Experiment 10. — With Liriodendroji tulipifera. I. Wood of 19-21 years, 19.6 per cent, of pentosan. II. Wood of 59-60 years, 18.8 per cent, of pentosan. Experiment 11. — With Magnolia acuminata. The year- rings of this tree were not distinctly discernible. I. Wood younger than 10 years, 17. 7x per cent, of pentosan. II. Old heart-wood, i6.8x per cent, of pentosan. Experiment 12. — With Platanus occidentalis. I. Wood of 4-10 years, 19.6X per cent, of pentosan. II. Heart-wood of 71-79 years, 19.4X per cent, of pentosan. Experiment 13. — With Primus Pennsylvanica. I. Wood of 4-12 years, 21. 5X per cent, of pentosan. II. Wood of 4-12 years, 20.8xx per cent, of pentosan. III. Heart-wood of 60-70 years, 19.5X per cent of pentosan. IV. Heart-wood of 60-70 years, i9.2xx per cent, of pentosan. I and II and also III and IV represent wood specimens origi- nating from different borings on the same cross-section of one disc. In these experiments the older wood always contained less pentosan than the younger. The difference is the greatest in Prunus Pennsylvanica, where it amounted to about 2 per cent. 226 De Chalmot. The analyses of the wood of this tree moreover show that the percentage of pentosan is not exactly the same in different. por- tions of wood of the same age and situated on the same cross- section. In other woods the older portions contain more pen- to-s in than the younger, as the following results show : Experiment 14. — 'W\\.\\ /uglans cinerea. I. Wood of 4-18 years, 18. ix per cent, of pentosan. II. Wood of 4-18 years, 17.9 per cent, of pentosan. III. Heart-wood of 60-65 years, 19. 2x percent, of pentosan. IV. Heart-wood of 60-65 years, 18.9 per cent, of pentosan. I and II and also III and IV are of different borings on the same cross-section of the same tree. Experiment 15. — With Fraximis Americanus. I. Wood of 2-7 years, 17.9X per cent, of pentosan. II. Heart-wood of 51-52 years, i8.8x per cent, of pentosan. In connection with these experiments I also ascertained whether the amount of pentosan in woods changes materially when a stronger decomposition than that into heart- wood takes place. I had 3 discs at my disposition which contained spots where the wood was dry-rotten. Experiment 16. — With Fraxinus platicarpa. I. Heart-wood, 17.4X per cent, of pentosan. II. Dry-rotten wood about 5 years older, i6.2x per cent, pen- tosan. Experiment ij. — With Prunus Pennsylvania a. I. Heart-wood, 19.8X per cent of pentosan. II. Dry-rotten wood of about the same age, 19.9X per cent, of pentosan. The specimens were of the same disc of wood as those in Experiment 13, but of the other cross-section. Experiment 18. — With Cercis Canade7ise. I. Heart-wood, 21. ix per cent, of pentosan. II. Dry-rotten, 21. gx per cent, of pentosan. Pentosans in Plants. 227 Experiment 19. — With a fence board of pine wood. I. Sound wood, 10. 2x per cent, of pentosan. II. Rotten wood, 7.9X per cent, of pentosan. We perceive no regularity in these results, and I therefore con- clude that the increase or decrease in the percentage of pentosan in ready-formed wood depends upon the individuality of the tree (or its species) to which it belongs. In one tree the pentosan therefore decomposes more easily than the other substances in the wood ; in another the reverse takes place. Perhaps also an infiltration of wood-gum into heart-wood, as Hartig claims,' takes place in some woods (Juglans). I believe it further to be safe to conclude that at least in the majority of cases, and most probably in all instances, a formation of pentosan does not take place after the wood ts ready made. Theoretical Conclusions. I have proved that pentosans are not formed by the assimi- lation process. They are therefore formed, direcdy or indirectly, out of hexoses. How and where does this take place ? Crosse, Bevan and Beadle^ have made a step towards solving the first part of this question. They found that if starch, cane- sugar, sugar of milk, or cotton-cellulose — in one word, if hexo- sans are gradually oxidized with chromic acid and afterwards dis- tilled with acids, large amounts of furfurol are obtained. By this oxidation of hexosans, substances are therefore formed that yield furfurol by hydrolysis. Whether these substances are pentosans is not as yet certain. It could as well be glucuronic acid or its isomeres, which also yield much furfurol by hydrolysis. Further communications on this point have been promised. However this may be, the formation of pentosan in plants is most probably bound up with the action of the living cell. After death pentosan is no longer formed ; the analyses of woods seem to indicate this clearly. In wood that is especially rich in pentosan the amount of these substances, as we just concluded, does not increase after its development is complete. The place where pentosans are formed may be the cell wall. Wood, dead corn leaves, straw and other materials where the contents of the cells have much diminished, and which therefore ' Botan. Centralbl. 56, 357. * Ber. d. chcm. Ges. 86, 2523. Vol. XVI.-16. 228 De Chalmoi. are rich in cell walls, contain also much pentosan. In many in- stances the pentosan seems, moreover, to be united with cellulose, since it cannot easily be separated from it.' Since, however, the presence of pentosan cannot be ascertained microscopically, the places where it is found in the living plant are only in a few instances known with absolute certainty. Schulze's paragalactan, which yields by hydrolysis arabinose, besides galactose,'' was found by Prof. Kramer^ to be contained in the walls of the cotyledonar cells of Ltipinus luteus. Amyloid, which occurs in the cell walls of many plants, contains, according to Winterstein,* more than 30 per cent, of pentosan. It is this matter that acts as a reserve substance in the seeds of Tropcsohim majus. Concerning the physiological significance of the pentosans in plants I might at present suggest the following: Pentosans accumulate during the whole life of the organs ; the plant seems to be profuse with pentosans. We have, however, no reason to believe that the pentosans are waste-products. They are transported to a limited extent, since we find small amounts of soluble pentosans in all plants. Pentosans are also resolved in the seeds and transported into the young plants, though not as readily as the hexosans (as starch). In seeds. of Propczolum the significance of the pentosans has been proven, since they conduct themselves therein entirely as reserve substances. It is probable that the pentosans have some importance for the formation of wood, since we find them especially in organs that have turned into wood. Large amounts of pentosan are, however, not essential for the wood-formation, since we find in some pine woods (hemlock) only 6 per cent. These investigations are to be continued. 'See E. Schuize: Ber. d, chem. Ges. 23, 2579, 3330; 34,2277. Ztschr. physiol. Cheni. IG, 387. Author's Inaiig. Dis., p. 40. Winterstein : Ztschr. physiol. Chem. 17, 381. ToUens and Schulze: Ann. Chem. (Liebig) 271, 55. W. Hoffmeister : Landw. Vers. Stat. 39, 466. Crosse and Bevan : J. Chem. Sec, 1889, p. 199. - Landw. Vers. Stat. 41, 223. ' Ibid. 36, 454. < Ber. d. chem. Ges. 35, 1237, where also the literature about this substance is referred to. On Phospho- Hydrocyanic Acid. 229 NOTE ON PENTOSANS IN SOILS. By G. dk Chalmot. There are pentosans which resist putrefaction, and for this reason we find pentosans in soils. This was already stated some years ago by a French investigator.' In the case of three soils I have estimated the percentage contents of pentosan and of humus, and therefore the ratio which the former bears to the latter. The soils were sifted through a i-mm. sieve. 25 or 50 grams were taken for analysis. Description of Soil. Humus. Pentosan. TOO parts of Humus contain . . . parts of Pentosan. Wood soil 23.42 per cent. 0.75 per cent. 3-2 Garden soil 9.85 0.39 4.0 Poor sandy soil 2.68 0.04 1-5 Richmond, Va., Ftlri mry, 18Q4. Contribution from the Chemical Laboratory of Lehigh University. ON PHOSPHO-HYDROCYANIC ACID. By W. B. Shober and F. W, Spanutius. Preliminary Paper. One of us, several years ago, working under the direction of Dr. Launcelot Andrews, attempted to prepare the phosphorus analogue of hydrocyanic acid — an acid the composition of which would be represented by the formula HCP — by heating chloro- form and zinc phosphide together in a sealed tube. The reaction was expected to take place in this way : 2CHCh + ZnsPa = 2HCP + 3ZnCk The results obtained were not satisfactory. Recently the work 1 1 have been unable to trace the paper alluded to, and I therefore cannot communicate where it was published. 230 SJiober and SpanuHus. has been taken up again, in this laboratory Other methods have been tried and we have succeeded in preparing the sodium salt of phospho-hydrocyanic acid. When dry ammonia is passed over heated sodium the reaction expressed by the following equation takes place : 2Na + 2NH3 = 2NH2Na H- H^.' When sodium amide is treated with carbon monoxide, under the proper conditions, sodium cyanide is formed : NH^Na + CO = NaCN + H2O. We repeated these experiments, substituting phosphine for ammo- nia, when the following reactions took place : 2Na -f 2PH3 = 2PH=Na= + H2 ; or, possibly, 2Na + PH3 rzNa^HP + Hs; or 3Na2 + 2PH3 = 2Na3P + 3H2. And then, PH=Na + CO = NaCP + H2O. The method of procedure was as follows : A clean, thoroughly dry piece of hard-glass tubing about 75 cm. long and 1.5 cm. in diameter was placed in a combustion-furnace. To the front end was attached a calcium-chloride tube, then a Woulff bottle con- taining concentrated sulphuric acid, the inlet and outlet tubes reaching almost to the surface of the acid ; finally, two Woulff bottles containing a saturated solution of copper sulphate. The object of these last was to decompose any phosphine which had escaped the action of the sodium. The last bottle was connected with a hood. To the rear end of the glass tube was attached a long (35 cm.) drying tube containing soda-lime; to this, ar.other tube of the same length filled with small pieces of potassium hydroxide ; to this two wash-bottles containing sulphuric acid, and finally a wash-bottle containing potassium-permanganate solution. Hydrogen was passed through the entire apparatus until all air was expelled. While a rapid current of hydrogen was still passing through, a small quantity (1.5 grams) of sodium, free from oxide and 1 Beilstein and Geuther ; Ann. Chem. (Liebig) 108, 88. 2 It is not known certainly that NaNHj is the correct forniula for the compound formed by the action of sodium upon ammonia. Assuming that NaNHj is the correct expression, the composition of the phosphorus compound formed under similar conditions is probably to be expressed by NaPHj. although it may be Na^PH or Na3P. In any case it is probably similar in its composition to the sodium compound, which, whatever it may be, yields sodium cyanide when treated with carbon monoxide. So this compound of phosphorus, whatever its composi- tion, yields, when treated with carbon monoxide, sodium fhosphocyanide. On Phospho- Hydrocyanic Acid. 231 hydroxide, was transferred to the glass tube. In order to prevent the action of the air and moisture and to keep the surfaces as bright as possible, the outside coating of the metal was removed under benzene. Without removing the benzene which adhered to it, the metal was quickly placed in the tube. The benzene was expelled by the aid of heat and the rapid current of hydrogen. When the benzene was completely removed, the hydrogen generator was replaced by a phosphine generator, from which all air had previously been expelled, and the permanganate and sulphuric-acid wash-bottles removed.' The phosphine was obtained by treating zinc phosphide with sulphuric acid. Heat was applied to the sodium until it melted. Phosphine was then generated slowly and passed through the apparatus. The bright silvery surface of the sodium began to blacken and show evidence of some action. Small black particles floated around on the surface of the metal, until finally the whole mass became covered with a thick black crust. The phosphine was passed until all of the sodium was converted into this black sub- stance, then, without allowing the apparatus to cool, dry carbon monoxide, free from air and carbon dioxide, was passed over this black substance for an hour. After the furnace was cool the glass tube was removed. On exposure to the air small particles of the substance ignited spontaneously. The tube was at once closed, when these small explosions ceased. It was then filled with absolute alcohol which had been standing over anhydrous copper sulphate for three months, corked up and allowed to stand for 12 hours. The alcohol became darker in color until it was a deep red. A sediment was deposited. Since the compound of sodium and phosphine is extremely unstable — decomposing upon exposure to the air — and since the difficulties in the way of estab- lishing its composition are very great — in fact, even greater than those that are encountered when an attempt is made to determine the composition of the compound which is usually represented by the formula NaNHz; — further, since sodium phosphocyanide is also extremely unstable ; we have found it impossible up to the present to analyze these compounds. Nevertheless, satisfactory evidence of the composition of sodium phosphocyanide is obtained by a study of its decomposition-products. ' Phosphine decomposes on passing through concentrated sulphuric acid. It deposits a reddish-yellow substance, presumably phosphorus, which on standing in contact with the acid explodes violently. 232 Shober and Spamdius. The alcohol and residue were transferred to a beaker and a small quantity of water added. Phosphine was immediately evolved in large quantites. This may have been due to the decomposition of sodium phosphocyanide, monosodium phosphine, disodium phosphine, trisodium phosphine, or to several of these compounds. The black substance above referred to probably consists of a mixture of at least two of these compounds. Since sodium cyanide is decomposed by alkalies into formic acid and ammonia, sodium phosphocyanide should be decomposed into formic acid and phosphine. The possible reactions giving rise to phosphine are these : NaH=P4-H=0 =NaOH4-PH3. Na^HP + 2H2O = 2NaOH + PH.% NasP + 3H2O — 3NaOH + PHs. NaCP + H.O = H.COONa + PH3. Each of these compounds would yield phosphine, but there is only one that could yield formic acid. Water was added until phosphine ceased to be evolved. The solution was then evaporated, a fluffy black substance was formed and filtered off. Upon drying it, the quantity obtained was too small to investigate. The filtrate, which was strongly alkaline, was neutralized with phosphoric acid and subjected to distillation. A portion of the distillate was heated to 6o°-70° and a solution of mercuric chloride was added. A heavy precipitate of mercu- rous chloride gave unmistakable evidence of the presence of formic acid. The presence of formic acid shows conclusively that the reaction referred to above had taken place, thus showing that the sodium salt of phosphohydrocyanic acid is capable of existence. With the advent of this compound a new field is opened. If a class of phosphorus compounds analogous to the cyanides, both organic and inorganic, can be prepared, they will form a remarkable series of compounds, each one of which will be an additional link in the chain of analogy between the related elements, nitrogen and phosphorus. This also foreshadows the existence of a compound the composi- tion of which will probably be represented by CP or C^Pi, analo- gous to cyanogen. We have been obliged to discontinue this investigation for the present, but in the near future we shall take it up again, SoUTii HETHi.niiKM, Pa., Fihriiaiy, 1804. REVIEW. Tabellarische Uebersicht der Naphtalinoerivate. Auf Grundlagc des Werkes : Sur la Constitution de la Naphtaline et de ses Derives, par F. Reverdin et E. Noelting, unter Beriicksichtigung der neucren Literatur, bearbeitet von F. Reverdin und H. Fulda. Basel, Genf, Lyon. Verlag von Georg & Co. 1894. (Erster Theil : Tabellen. Zweiter Theil : Literatur.) The authors of this work have attempted to give references to all articles treating of naphthalene derivatives, and to arrange these derivatives in tables in such a manner as to tell the story of their constitution in as few words as possible. The number of compounds included in the tables is 911, and these are all deriva- tives of naphthalene. Many of these compounds are of scientific interest, and many are now used in the dye-stuff industry, and in view of the extent of the literature it is, of course, desirable to have such an excellent work as this at hand as a guide to the sources. The earlier works upon which the present one is based are " Ueber die Constitution des Naphthalins und seiner Derivate," which appeared in 1880; and " Sur la constitution de la naph- taline et de ses deriv6s," which appeared in 1888. These have brilliantly withstood the test of use, so that it is fair to assume that the new edition will maintain the high standard of its prede- cessors. The manuscript was finished August i, 1893, ^"d articles which have appeared since this date have not received consideration. The authors do not claim thoroughness except for their refer- ences to German and French literature. They say: " Was die Literatur anderer Lander betrifft, so standen uns zwar nur die wichtigeren Zeitschriften zu Gebote, wir haben jedoch derselben noch ausserdem Rechnung getragen, soweit uns die Abhand- lungen im Original oder durch Referate in deutschen oder fran- zosischen Zeitschriften zuganglich waren." While probably no great harm has resulted from this off-hand treatment of the " Literatur anderer Lander," it seems to one who has the misfor- tune to belong to " another country " that, in a work of this kind, we have a right to expect as thorough a consideration of the "Literatur anderer Lander" as that which is accorded to the French and German. i. r NOTE On a neiv Class of Organic Bases, contammg Iodine but no Nitroge7i. Victor Meyer and C. Hartmann' have just described some remarkable compounds containing iodine. The principal one • Ber. d. chem. Ges. 87, 426. 234 Note. thus far obtained is made by treating iodosobenzene, CtHt.lO, with concentrated sulphuric acid at a low temperature. On diluting the solution thus obtained, and adding potassium iodide, a precipitate is formed which is the hydriodide of a base. The salts of the base resemble those of lead and of silver, but still more those of thallium. The sulphate is easily soluble in water, the nitrate more difificultly soluble. The iodide is a yellowish, in- soluble precipitate ; the bromide a pale yellowish, and the chloride a white precipitate. The aqueous solution of the free base has a strong alkaline reaction. It is obtained by shaking the solution of the hydriodide with silver oxide. The hydriodide, or iodide, was analyzed and shown to have the composition represented by the formula C12H9I3. When subjected to dry distillation, this breaks down into monoiodo- and diiodobenzene thus : C.iH9l3=C6H6l + C6H4l2. The salts studied, in addition to the iodide, are the bromide, CisHJzBr; the chloride, C12H9I2CI; the nitrate, C12H9I2.NO3; and the sulphate, wnich was not analyzed. The formation of the base is probably represented thus : 1. CeHsIO + H.O = GH6l^-^^ + ^'^ + ^• According to this, the base has the constitution expressed by /CeHs the formula, I — C6H4I, and it appears to be a derivative of the \OH compound, I — H , which, as far as the composition is concerned, \OH /-H is analogous to hydroxylamine, N — H . But the two must differ ^OH fundamentally in their conduct towards acids. While hydroxyl- amine forms salts in the same way that ammonia does, by direct addition to acids, the iodo-base reacts with acids like the ammo- nium bases with elimination of water : /CgHb /CeHs I— CGH4I -f HI == I— CeH^I -f H2O. ^^« ^^ /CeH. The authors have further obtained the compound, I — CeHs, and \0H are now engaged in studying these interesting products. Vol. XVI. [April, 1894.] No. 4. AMERICAN CHEMICAL JOURNAL. ON THE DECOMPOSITION OF DIAZO-COMPOUNDS. VIII.— A STUDY OF THE ACTION OF THE SALTS OF DIAZO- BENZENE ON METHYL AND ETHYL ALCOHOLS UNDER DIFFERENT CONDITIONS. > By J. L. Beeson. The work, the results of which are recorded in the following pages, was undertaken at the suggestion of Professor Remsen, and carried on under his guidance, and had for its object the careful study of the decomposition of the salts of diazo-benzene by methyl alcohol alone and by methyl and ethyl alcohols in the presence of certain other substances. Action of Diazo-benzene Nitrate on Methyl Alcohol. I. — Portions of the diazo-compound, varying from 30 to 80 grams, were placed in a one-litre flask, and anhydrous methyl alcohol added in the proportion of 5 cc. of the former to i gram of the diazo-salt, connected with an inverted condenser and slowly heated. A slow decomposition began at the room temperature. This grew more rapid as the temperature was raised, until at 30°-35° C. there was a rapid decomposition accompanied by a copious evolution of nitrogen gas and a gradual rise in tem- perature to the boiling-point of the alcohol without the further application of heat. 222 grams of the diazo salt and methyl ' From the author's thesis for the degree of Doctor of Philo.sophy. Submitted to the Board of University Studies of the Johns Hopkins University, June, 1893. 236 Bceson. alcohol were thus decomposed and the residues united. During all of these decompositions (six in number) no odor of aldehyde was perceptible, nor had the residues the odor of aldehyde. In the case of two or three the escaping gas was passed through either water or methyl alcohol, and the liquid examined for aldehyde by treating with an ammoniacal solution of silver nitrate, but there was no evidence of the presence of aldehyde. The united residues, which were strongly acid and had the odor of anisol, were poured into a one-litre flask and the alcohol and volatile products distilled off. The methyl alcohol thus obtained, which was slightly yellow, due to the presence of anisol, was poured into two-litre flasks, and six volumes of a saturated solu- tion of sodium chloride added. There rapidly rose into the neck of the flask an amber-colored oil, which was removed by means of a pipette. The methyl alcohol was then distilled off the salt- solution, and the first three portions of 20 cc. each were separately treated with six volumes of salt-solution and allowed to stand. There separated from the first two portions a small quantity of an oil having the same odor and appearance as that obtained in the first instance, to which it was added. The tarry contents of the flask from which the methyl alcohol had been distilled were made alkaline with caustic soda and distilled in steam. There was thus obtained a considerable quantity of oil resembling that obtained in the preceding cases. These portions were united, dried over fused calcium chloride, and subjected to fractional distillation. Almost the entire product passed over at i48''-i53° C, and proved to be anisol or methoxy-benzene. No trace of benzene was found. The contents of the flask which had been distilled in steam to remove the anisol were acidified with sulphuric acid and further distilled in steam. A yellow solid slowly collected in the con- denser and receiver, while the water had a deep yellow color. This operation was continued until the solid ceased to collect, and further, until the liquid which came over was almost colorless. The liquid was filtered from the solid, made alkaline with caustic soda, and evaporated to a small bulk. Upon acidifying with hydrochloric acid, a whitish-yellow substance was precipitated which, upon further examination, proved to be the same as the solid obtained in the first instance. These were united and subjected to careful crystallization, first from hot water and then from alcohol. In each case the substance was obtained in the On the Decomposition of Diazo- Compounds. 237 form of whitish-yellow plates which melted at 114° C. (Zincke thermometer) and otherwise had the properties of 2:4-dinitro- phenol. The liquid remaining in the flask through which steam had been passed was separated from the tarry substance which had been formed in the decomposition, was boiled with animal charcoal and evaporated down in stages. No definite substance was obtained from the tarry residue. From the 222 grams of diazo-benzene nitrate 58.2 grams of anisol (40 per cent.) and 20 grams of 2 : 4-dinitrophenol were obtained. Plainly the principal reaction that has taken place in this case is that represented by the following equation : aH5.N2.N03+CH30H = C6H50CH3 4-N3+HN03. Remsen and Orndorff have shown' that with ethyl alcohol this same diazo-compound gives principally the ethoxy- derivative together with a small quantity of benzene. No benzene is obtained when methyl alcohol is used. This is quite in accordance with other facts established in the course of this series of investigations, and points to the conclusion that the simpler the alcohol the greater the tendency to the alkoxy-reaction. As regards the dinitrophenol, it was thought improbable that the comparatively large quantity of this substance obtained was due to the presence of water in the alcohol, for this had been distilled from lime over which it had stood for some days. The presence of water would probably account for a small portion, but not for the entire 20 grams of the substance. It was thought more prob- able thatortho-nitrophenol may have been first formed by a mole- cular rearrangement of the diazo-compound, as was found to be true by Remsen and Orndorff in the case of the action of diazo- benzene nitrate on ethyl alcohol," and the mono-nitrophenol thus formed further nitrated to the dinitrophenol (a) either by the action of the nascent nitric acid during the decomposition, {b) or by standing in contact with the free nitric acid, {c) or while heating to distill off the alcohol and anisol. If ortho-nitrophenol is first formed and not further nitrated by the nascent nitric acid, then by- making alkaline as soon as the decomposition had ceased only ortho-nitrophenol should be found. In order to throw light upon this point the following experiment was performed. ' This Journal 9, 387. a /^,v. 9, 389. 238 Beeson. II. — Forty-two grams of diazo-benzene nitrate and methyl alcohol (off lime) were decomposed under conditions similar to those in the preceding case. The contents of the flask were made alkaline with caustic soda solution as soon as the decomposition had ceased. The alcohol was distilled off and the anisol obtained by pouring into a saturated solution of sodium chloride. The tarry residue was distilled in steam to remove the anisol, then made acid with sulphuric acid and further distilled in steam. A yellow solid rapidly collected in the condenser, which crystallized from hot alcohol in long yellow needles, melting at 45° C, and possessing the general properties of ortho-nitrophenol. The liquid remaining in the flask was separated from the tarry sub- stance, boiled with animal charcoal, evaporated to a small bulk, filtered, and about an equal volume of strong hydrochloric acid added. Upon standing, long, almost colorless needles crystal- lized out, which upon further crystallization from hot water were colorless, melted at 115° C, and otherwise had the properties of para-nitrophenol. Products : Anisol 12 grams, or 42.5 per cent. Ortho-nitrophenol 2 grams, or 5.7 per cent. Para-nitrophenol 0.4-0.5 gram, or 1.2 per cent. This indicates that mono-nitrophenols were first formed, and then nitrated to thedinitrophenol, either while standing in contact with the free nitric acid or when the alcohol is distilled off. The ortho-compound was doubtless formed by the molecular rearrange- ment of the diazo-benzene nitrate with the loss of nitrogen, as before mentioned, but the presence of the small amount of the para-compound was probably due to the presence of a small quantity of water in the alcohol. Remsen and Orndorff found that on heating diazo-benzene nitrate in a neutral liquid (toluene) 24 per cent, of it was converted into ortho-nitrophenol, as expressed by the following equation : CeHsN^NOs = Q^Wa^C^^ -\- N-2, while "not a trace of the para-compound was found.'" This in itself does not prove that no para-nitrophenol will be formed when the diazo-salt is decomposed with alcohol, for in the one case the decomposition is effected in the presence of a liquid supposed to be neutral, in the other case with a liquid which is known to be capable of reacting with it. 1 This Journafl 9, 391. On the Deco77iposUion of Diazo- Compounds. 239 III. — The preceding experiment was now repeated, using abso- lute methyl alcohol. This was prepared by treating the alcohol which had been distilled from lime, with anhydrous copper sul- phate, and letting them stand together for several days. The alcohol was then poured off and redistilled. The only products obtained were anisol and ortho-nitrophenol. A careful examina- tion was made for the para-compound by separating from the tarry mass the liquid remaining in the flask through which steam had been passed to remove the ortho-compound, and then evap- orating it down in stages, each time adding strong hydrochloric acid and letting cool. No organic matter crystallized out. The liquid was then evaporated to dryness and the residue of inor- ganic salts treated with hot alcohol. The alcohol was colored slightly yellow, but no para-nitrophenol was obtained from it. This experiment was repeated, but in this case also no para-nitro- phenol was found. Hence it was concluded that when diazo-ben- zene nitrate is decomposed with absolute methyl alcohol the pro- ducts of the reaction are anisol and ortho-nitrophenol. IV. — In order to determine whether the nitration of the ortho- nitrophenol to the dinitro-compound took place on standing in contact with the free nitric acid, or when heating to distill off the alcohol, 42 grams of the diazo-salt were decomposed with methyl alcohol, allowed to stand over night, and then made alkaline with caustic soda solution. Products: Anisol 13 grams, or 48 per cent. Ortho-nitrophenol 1.7 grams, or 5 per cent. Dinitrophenol i gram, or 2.2 per cent. Probably more of the mono-nitrophenol would be converted into the dinitro-compound by allowing the liquid to stand longer before neutralizing the nitric acid. The fact that only ortho-nitrophenol is formed when diazo- benzene nitrate is decomposed with absolute methyl alcohol (III) is evidence that no para-compound is formed by the molecular rearrangement, but it is only negative proof that the para-com- pound obtained in case II was due to the presence of a small quantity of water in the alcohol. If, however, this be the case, then dilute methyl alcohol should give an increased percentage of both the ortho- and para-nitrophenols, for both would be formed by the action of the free nitric acid on phenol. 240 Bee son. V. — Fifty grams of the diazo-salt were decomposed with 250 cc. of methyl alcohol diluted with 30 per cent, of water, at the room temperature, and the liquid made alkaline with caustic soda solution as soon as the action had ceased. The diazo-compound dissolved to a wine-colored liquid. The decomposition began without the application of heat and was allowed to decompose spontaneously. Products: Anisol 4.5 grams, or 14 per cent. Ortho-nitrophenol 5grams, or I2percent. "I „ er ce t Para-nitrophenol 2.4 grams, 6 per cent. J For comparison I repeat the results obtained in case II with alcohol off lime; Anisol 42 per cent. Ortho-nitrophenol 5.7 per cent. ) g^ ^^^ ^^^^^ Para-nitrophenol 1.2 per cent. ) With absolute alcohol the products obtained were anisol and ortho-nitrophenol. It will be noticed that there is a large increase in the per- centage of both the ortho- and para-nitrophenols in case V over that in case II, due to the presence of a larger amount of water in the former. But the decrease in the percentage of the anisol (28 per cent.) is too large to be accounted for by the assumption that the water present in case V uses up the diazo-salt to form phenol, of which action the increased percentage of the nitro- phenols (11 per cent.) is a measure. This discrepancy, it was thought, might be due to the difference in temperature at which the two decompositions were effected, for it has been shown by Remsen and Dashiell that increased temperature and pressure increase the alkoxy reaction.' VI. — In order to make the results in the two cases strictly com- parable, 71 grams of the diazo-salt were decomposed with methyl alcohol at the room temperature without dilution, and the con- tents of the flask made alkaline as soon as the decomposition had ceased. The products, separated as in the preceding cases, were : Anisol 27 per cent. Ortho-nitrophenol 5.6 per cent. Upon comparing these results with those obtained in the pre- ceding experiment it will be seen that the increase in the total \_ ' This Journal 15, 125-6. On the Decomposition of Diazo- Compounds. 241 percentage of the nitrophenols (12.4 per cent.) due to the presence of water, and the decrease in the percentage of anisol (13 per cent.) due to the using up of the diazo-salt by the water to form phenol and then nitrophenol, mutually account for each other. The presence of the water in case V also accounts for the formation of the para-nitrophenol, and for the increased per- centage of the ortho-compound, for both are produced by the nitration of phenol. Upon comparing the results obtained in cases II and VI there will be noticed a considerable decrease in the percentage of anisol formed at the lower temperature. This is of special interest, because here a high temperature apart from pressure appears to favor the alkoxy reaction. Remsen and Dashiell effected the decomposition of a diazo-compound with alcohol in a suitable flask, under increased pressure and temperature, but were unable to determine which was the primary cause in influ- encing the alkoxy reaction, whether the high temperature conditioned by the increased pressure, or the increased pressure due to an elevated temperature. In order to throw light upon this point they brought together a diazo-compound and alcohol under varying pressures without a corresponding change of temperature. But the decomposition would not take place except at the temperature of the boiling-point of the alcohol at that pressure.' While the results obtained in cases 11 and VI indicate that the elevated temperature favors the alkoxy reaction, they do not indicate that a lower temperature favors the hydrogen reaction. To test this point a diazo-compound and alcohol must be selected which give both the hydrogen and alkoxy reactions. Diazo- benzene nitrate and ethyl alcohol do this, and hence the following experiment was performed. VII. — 150 grams of diazo-benzene nitrate and ethyl alcohol distilled from lime were decomposed in two lots, A and B, at the room temperature. The action was slow, accompanied by a slight elevation of temperature and a gradual reddening and darkening of the alcohol. Froni two to three days were required for the completion of the decomposition. A. 119 grams of the diazo-salt and about 500 cc. of ethyl alcohol were allowed to stand together at the room temperature 1 This Journal 15, 135. 242 Bee son. until decomposition had ceased. The volatile products were distilled off and the residue then made alkaline with caustic soda solution, and distilled in steam to remove the phenetol. This distillate was added to the alcoholic distillate and set aside. Upon acidifying and further distilling in steam, 2 : 4-dinitrophenol was obtained, the quantity being less than was obtained in the corres- ponding case, when methyl alcohol distilled from lime was used. B^ 35 grams of the diazo-salt were decomposed with absolute ethyl alcohol at the room temperature, and made alkaline as soon as the decomposition had ceased. There were obtained 3 grams of ortho-nitrophenol, but no para-compound. These results confirm the conclusions drawn in the corresponding cases with diazo-benzene nitrate and methyl alcohol. The alcoholic distillates from the two decompositions were united and poured into six volumes of a saturated solution of sodium chloride. A rather dark-colored liquid, having the odor of aldehyde, rapidly rose. This was collected, dried over fused calcium chloride and subjected to careful fractional distillation. From the 30 grams of the dried liquid there were obtained 6.4 grams of phenetol. But the other constituents could not be separated by this means. Strong sulphuric acid was added, the liquid darkened, and upon gentle heating there began a copious evolution of ordinary aldehyde. When this had ceased strong nitric acid was added, and nitrobenzene was obtained. Phenetol, nitrophenols, diphenyl, paraldehyde, and benzene were thus obtained. The quantity of diphenyl formed was very small, probably not over 0.5 gram. It was doubtless formed by the union of the two phenyl residues in the nascent state, as it is found in many cases where benzene is being formed — e. g. from benzoic acid and lime. The ratio of the phenetol to the benzene was not determined in this case, as a portion of the latter was doubtless acted upon by the sulphuric acid, and a part of it carried off mechanically by the rapid escape of the aldehyde. So about the same quantity of the diazo-salt as was used in the preceding case was decom- posed with alcohol, and the decomposition-flask kept in a bath of cold water to prevent a rise in temperature during the decom- position. The products were obtained as in the preceding case, and the phenetol separated as before by fractional distillation. The mixture of paraldehyde and benzene was poured into a On the Decomposition of Diazo- Compounds. 243 small flask connected with an inverted condenser, strong nitric acid was added and the flask heated, when there began a copious evolution of aldehyde which continued until all of the paralde- hyde was broken down. The remaining liquid was separated from the acid and distilled. 12 grams of the liquid distilled over at 78°-85° C, the few drops remaining were nitrobenzene. The distillate, which had a slight odor of aldehyde, was treated with a mixture of concentrated nitric and sulphuric acids. There were thus obtained 18 grams of nitrobenzene, equal to 10 grams of benzene. From the 27.5 grams of the mixture there were obtained phenetol 2 grams, benzene 12 grams, and paraldehyde 13.5 grams. Here, it will be seen, the principal product of the reaction was benzene, while in the case of the decomposition of the same diazo-salt and alcohol at a higher temperature (above 50°-6o° C.) Remsen and Orndorfif got mainly the alkoxy pro- duct. For comparisons I give their results :' Phenetol, 22 per cent. Benzene, 4 per cent. This shows that, in this case at least, the temperature apart from pressure influences the character of the reaction, the higher temperature favoring the alkoxy reaction, the lower temperature the hydrogen reaction. Effect of the Presejice of Water tipoji the Diazo- Alcohol Reaction. In the case of the methyl alcohol and diazo-benzene nitrate, it will be seen, upon comparing experiments V and VI, that the presence of 30 per cent, of water had no effect upon the character of the reaction except to decrease the percentage of the anisol, and proportionally to increase that of the nitrophenols. It was thought desirable to study the action of dilute ethyl alcohol on diazo-benzene nitrate. A. 50 grams of the diazo-salt and 200 cc. of ethyl alcohol diluted with 30 per cent, of water were decomposed at the room temperature, and the liquid made alkaline as soon as the action had ceased. The products, obtained as in the preceding case, were: Ortho-nitrophenol 5 grams, or 12 per cent. Dinitrophenol 2.2 grams, or 5 per cent. Phenetol 4 grams, or 11 per cent. 1 This Journal 9, 389. 244 Beeson. and 2 grams of a liquid which distilled over between 75° and 85° C. and had the odor of aldehyde. This contained paralde- hyde, for upon adding strong sulphuric acid and heating an evolution of aldehyde took place. Nitric acid was then added to the remaining liquid and this heated for some time. Upon pour- ing the whole into a small beaker of water, a few small globules, apparently of nitrobenzene, collected. At any rate the principal product was the alkoxy compound. This experiment was repeated, using less dilute alcohol. ^- 35 grams of the diazo-salt were decomposed with ethyl alcohol diluted with 10 per cent, of water. The flask was kept as near the room temperature as possible, by placing the decom- position-flask in a bath of water. The products were collected as in the preceding case. From the 8 grams of the mixture there were obtained 5 grams of phenetol and 2 grams of a low-boiling liquid, which upon the addition of strong nitric acid and heating, was almost completely converted into aldehyde. In this case 0.5 gram of diphenyl was formed. No examination was made for nitrophenols. By comparing these results with those obtained in the preced- ing case it will be seen that the presence of a small percentage of water favors the formation of the ethoxy product. Actio7i of Diazo-benzene Nitrate on Sodium Methylate. In the preceding decompositions the yields of the alkoxy pro- ducts were rather small, in each case a large amount of tarry sub- stance and in some instances not an inconsiderable quantity of nitrophenols being formed. It was thought probable that if something were present during the reaction to combine with the acid, the formation of the nitrophenols might be prevented. Hence the action of sodium methylate on diazo-benzene nitrate was tried, in the expectation that the following reaction would take place: OH5N2NO3 + NaOCHs = CeH.OCHs + NaNOs + N2. A saturated solution of sodium methylate in methyl alcohol was prepared by dropping small pieces of pure metallic sodium into absolute methyl alcohol in a flask kept cool in an ice-water bath until the sodium ceased to react further with the alcohol. To this 87 grams of diazo-benzene nitrate were added little by little, stir- ring meanwhile. The flask was kept in the ice-water so that the On ihe Decomposition of Diazo Compounds. 245 temperature did not rise above 20°-25° C. The action took place energetically with the evolution of nitrogen gas and a rise in tem- perature. The diazo-body was partly changed into a yellowish gummy mass, while the liquid gradually darkened. A peculiar penetrating odor was observed during the decomposition. This was thought to be due to the formation of formic aldehyde. Some of the escaping gas was passed through water and some through methyl alcohol, in the hope of absorbing the aldehyde if present. These liquids were then tested for aldehyde by treating with an ammoniacal solution of silver nitrate, but there was no reduction. This test was made in the case of several of the decompositions, and with the same results. The methyl alcohol and volatile products were then distilled off and poured into about five volumes of water. There rose a considerable quantity of clear liquid having the odor of benzene. This was dried over fused calcium chloride and distilled. The whole passed over at 78°-83° C. The liquid solidified at i°-2° C, and, with strong nitric and sulphuric acids, gave nitrobenzene. Hence there can be no doubt that the substance was benzene. The liquid from which the benzene had been obtained was then satu- rated with sodium chloride, but no anisol separated out. Water was then added to the contents of the flask from which the alcohol and benzene had been distilled, and this distilled in steam. The first 100 cc. of the distillate were treated with six volumes of a saturated solution of sodium chloride, but nothing was thus obtained. A red solid rapidly collected in the condenser and receiver. The distillation in steam was continued until the solid ceased to collect, then the contents of the flask were acidified with sulphuric acid and further distilled in steam, but nothing was obtained. The liquid remaining in the flask was separated from the tarry substance and evaporated down in stages. The only compounds obtained were sodium nitrate arid sodium sulphate. The red solid was crystallized several times from hot alcohol, from which it was obtained in the form of small white pearly plates, which melted at 67.9° C. (uncorrected), and otherwise had the properties of diphenyl. The products of the decomposition were: Benzene 20 grams, or 50 per cent. Diphenyl, about 7 grams, or 15 per cent. 246 Bee son. In order to account for the formation of the diphenyl the fol- lowing reaction may be supposed to take place: C6H5N.NO3 NaOCHa Col-h + =1 +2N2+2NaN03 + 20CH,. C6H5N.NO3 NaOCHs C6H5 The two alkoxy residues would probably unite to form methyl alcohol and formic aldehyde. In regard to the formation of the benzene there are two ways in which it may be accounted for: {a) by supposing the action between the diazo-salt and the sodium methylate to give an inter- mediate product, and this to break down into formic aldehyde, nitrogen and benzene, according to the following equation: C6H5N2NO.. + NaOCHs = NaNOs -f C6H5N20CH3=C.H6 + N, -\- CH2O ; or, {U) by supposing a reaction to take place between the diazo-compound and the alcohol in which the sodium methylate was dissolved, giving the well-known hydrogen reaction. The latter supposition seems hardly probable, since there was a large excess of the sodium methylate present, and since methyl alcohol, at the same temperature at which this decomposition was effected, gave only the methoxy product. It may be, however, that the sodium methylate simply acts as an alkali, neutralizing the acid of the nitrate and setting free the diazo-compound. If this view be correct, then the presence of any alkali, or any other substance that would combine with the acid, ought to lead to the hydrogen reaction. To test this the next experiments were tried. Action of Diazo-benzene Nitrate on Methyl Alcohol in the Pres- ence of Alkalis and of Calcite. A. 70 grams of the diazo-salt were decomposed with absolute methyl alcohol in the presence of an excess of pure sodium hydroxide. The diazo-compound was slowly added under con- stant stirring, the temperature being kept below 30° C. by immersing the vessel in a bath of ice-water. The action took place with considerable energy. The general phenomena were the same as in the preceding case. The products obtained, as in the preceding case, were : Benzene 7 grams, or 22 per cent. Diphenyl (crude), 8-10 per cent. B. 60 grams of the diazo-compound and absolute methyl alcohol were decomposed at a temperature below 30° C. in the presence of On the Decomposition of Diazo- Compounds. 247 an excess of anhydrous sodium carbonate. The action was less energetic than in the two preceding cases. Products: Benzene 3.1 grams, or 10 per cent. Diphenyl (crude) 6 grams, or 21 per cent. C. 61 grams of the diazo-salt and methyl alcohol were allowed to decompose at the room temperature in the presence of an excess of pure calcite. The general phenomena were the same as when the diazo-compound was allowed to decompose at the room temperature with methyl alcohol alone. Products: Anisol 11 grams, or 25 percent. Ortho-nitrophenol, 0.5 gram, or i per cent. It appears from these experiments that the presence of alkalis leads to the hydrogen reaction, while the presence of calcite has no effect upon the character of the reaction. Action of Diazo-benzene Nitrate on Sodium Ethylaie. The experiments with sodium methylate and methyl alcohol in the presence of alkalis naturally suggested similar ones with sodium ethylate and ethyl alcohol in the presence of alkalis. A saturated solution of sodium ethylate in ethyl alcohol wasprepared in a similar manner to that of the sodium methylate, the experiment carried out under the same conditions and the products separated in the same way. The phenomena were in general the same as those in the experiment with sodium methylate, but the action was far more violent in this case, requiring cautious addition of the diazo- body to prevent an ignition of the alcohol. The alcoholic distil- late did not have the odor of aldehyde. From the 47 grams of the diazo-compound there were obtained the following products: Benzene 4.1 grams, or 19 per cent. Diphenyl (crude) 5 grams, or 17 per cent. It will be seen that the products are the same as those obtained with sodium methylate, though the yield of benzene was smaller. Relatively a much larger amount of tarry products was obtained in this case than in that of the sodium methylate. Action of Diazo-bcnzene Nitrate on Ethyl Alcohol in the Presence of Alkalis. The conditions under which these decompositions were made were the same as those of the preceding cases ; the phenomena in 1 15 grams. 248 Bee son. general were the same, and the products were collected in the same way. A. 47 grams of the diazo-salt and ethyl alcohol were decom- posed in the presence of an excess of anhydrous sodium hydroxide dissolved in the alcohol. Products : Benzene 5 grams, or 27 per cent., and diphenyl. The amount of the diphenyl was not determined, but the yield was smaller than that obtained in the case of the sodium ethylate. B. 50 grams of the diazo-compound were decomposed with ethyl alcohol in the presence of an excess of anhydrous sodium carbonate. The products obtained in this case were: Diphenyl 1.7 grams, or 6 per cent. Benzene Paraldehyde The ratio of the benzene to paraldehyde was not determined, since they could not be separated by fractional distillation. But the lower-boiling portions were heated with a mixture of strong nitric and sulphuric acids and converted into nitrobenzene. From the results obtained in these experiments it appears that the presence of alkalis in general leads to the hydrogen reaction. But since it has been shown that a low temperature favors the hydrogen reaction, it was thought possible that the low temperature at which these decompositions were effected was the determining factor and not the presence of the alkalis. In order to throw light upon this point the next experiments were tried. Action of Diazo-benzenc Nitrate on Methyl and Ethyl Alcohols in the Presejice of an Excess of Sodiu77i Hydroxide at an Elevated Temper atiire. Since all of these decompositions in the presence of alkalis gave the same products, it did not matter which was repeated at the elevated temperature. Merely for the sake of convenience the one in the presence of caustic soda was selected. A. 71 grams of the diazo-compound were decomposed with 400 cc. of absolute ethyl alcohol in the presence of an excess of sodium hydroxide at a temperature of 55°-70° C. This is the tempera- ture at which this same diazo-salt and alcohol alone gave mainly the alkoxy product.' The decomposition was effected in a flask » This Journal 9, 389. On the Decomposiiion of Diazo- Compounds. 249 with a very long neck, which was wrapped with a wet towel in order to condense any volatile products which tended to escape. The products obtained were: Benzene 5.5 grams, or 17.5 per cent. Diphenyl (crude) i.i grams, or 4 per cent. B. 37 grams of thediazo-compound were decomposed with 200 cc. of methyl alcohol in the presence of an excess of caustic soda at a temperature of 50°-55° C. Products : Benzene 3.5 grams, or 19 per cent. Diphenyl (crude) 0.9 gram, or 5.5 per cent. Relatively more tarry products were formed at the high tem- perature and correspondingly less benzene and diphenyl, than at the low temperature. Otherwise, it will be seen that the character of the reaction was the same in both cases. This series of experiments shows that in general the presence of alkalis leads to the hydrogen reaction, but just how, it is impos- sible to say until more facts bearing upon the question are accu- mulated. It may be that intermediate products are first formed, as was suggested in the case of sodium methylate. In the presence of alkaline hydroxides, then, diazo-benzene would probably first be formed, and this reacting with the alcohol give the hydrogen reaction, as expressed by the following equations: CeHs.N.NOs + NaOH = NaNO. + CeHnN.OH, and CeHs.Ns.OH -f CH.OH = CeHe -f Ns + H=0 + CH2O. But this cannot be claimed for the cases in which the decomposition was effected in the presence of sodium carbonate. Then it appears that, since the general phenomena and the products obtained in all these cases were the same, no intermediate products were formed in any of them, and that the hydrogen reaction is due to some other causes, possibly the prevention of the acidity of the medium by the neutralization of the acid at the moment of liberation. If this view be correct, then the presence of any substance which would combine at once with the acid as soon as set free ought to determine the hydrogen reaction. The study of the decomposi- tions in the presence of zinc-dust, given later on, confirms this view. And, on the other hand, the presence of a considerable quantity of free acid ought to increase the amount of the alkoxy products in those cases where both the hydrogen and alkoxy reactions take place. I am not aware that any experiments of this kind have been tried. 250 Bee son. Action of Diazo-benzeyie Sulphate on Methyl Alcohol. 144 grams of this diazo-compound were decomposed with 350 cc. of methyl alcohol distilled from lime. The application of heat was necessary to start the decomposition, which began at 4o"'-45° C. and continued spontaneously with a gradual rise in temperature to the boiling-point of the alcohol. The general phenomena of the decomposition were the same as in the case of the action between methyl alcohol and diazo-benzene nitrate. The action was a clean one, only a small quantity of tarry mate- rials being formed. The alcohol and volatile products were distilled off, and the remaining contents of the flask distilled in steam to remove the anisol. The two distillates were treated with about six volumes of a saturated solution of sodium chloride. The oil which rose was collected, dried over calcium chloride and subjected to fractional distillation. About 0.5-1 cc. of a colorless liquid came over below 100° C. when the thermometer rapidly rose to the boiling-point of the anisol, where almost the entire product distilled over. The low-boiling liquid was very volatile and had the odor of methyl alcohol, which it probably was. I think there can be no doubt that no benzene was present in the anisol. The contents of the flask which had been distilled in steam were neu- tralized with barium carbonate, filtered and evaporated down in stages, being allowed to cool slowly at each stage. The only thing found was aniline sulphate which had escaped diazotiza- tion. From the 144 grams of the diazo-benzene sulphate the only product obtained was 40 grams, or 53 per cent., of anisol. Action of Diazo-benzene Sulphate on Sodium Methylaie and Eihylaie. The reaction in these cases took place only to a slight extent, even upon heating to the boiling-points of the alcohols for an hour or more. Almost the whole of the diazo-body in both cases was changed into a hard, black, tarry mass. The alcohols were distilled off and treated with a saturated solution of sodium chloride. From each a small quantity of a dark-colored liquid rose, but the quantities were too small to determine what the pro- ducts were. The tarry residues were then distilled in steam. In each case a very small quantity of diphenyl was obtained. These On the Decomposition of Diazo- Compounds. 251 results were surprising, for the diazo-benzene nitrate and the sodium alcoholates reacted energetically at a low temperature, and gave good yields of benzene and diphenyl. Action of Diazo-benzene Sulphate on Methyl and Ethyl Alcohol in the Presence of an Excess of Ziyic-dust. The previously described decompositions made in the presence of alkalis suggested similar ones in the presence of finely divided metals. Experiments with finely divided copper and iron were not made with this compound, but iron by hydrogen had no effect upon the reaction between diazo-benzene nitrate and methyl alcohol. When diazo-benzene nitrate and sulphate were sepa- rately brought together with methyl or ethyl alcohol and zinc- dust added, an energetic reaction at once took place. A. ^i grams of diazo-benzene sulphate and methyl alcohol were decomposed in the presence of zinc-dust at a temperature below 30° C. The dry zinc-dust and alcohol were placed together in a flask immersed in a bath of ice-water and the diazo-salt slowly added while the liquid was stirred. The action was a clean one, almost no tarry products being formed. The alcohol and volatile products were distilled off, and treated with a saturated sodium chloride solution. A clear liquid rose, was collected, dried over calcium chloride and distilled. The liquid boiled at about 80° C, solidified at i°-2° C, and otherwise had the properties of ben- zene. The residue from which the alcohol and benzene had been distilled was now distilled in steam. The only substance thus obtained was diphenyl. Products : Benzene 3 grams, or 16 per cent. Diphenyl 1.4 grams, or 7 per cent. The products obtained in all of the following decompositions in the presence of zinc-dust were obtained by the method described in this case. B. jj grams of diazo-benzene sulphate and ethyl alcohol were decomposed in the presence of an excess of zinc-dust. The action in this case was somewhat less energetic than that in the preceding one. Products : Diphenyl 0.5 gram. Paraldehyde 26.5 grams. Benzene 4.5 grams. The benzene in the mixture of paraldehyde and benzene was estimated by getting rid of the paraldehyde by boiling the mixture Vol. XVI.-18. 252 Bee son. with strong nitric acid in a flask connected with an inverted con- , denser, and then converting the liquid remaining in the flask into nitrobenzene. The aldehyde was estimated by difference. Action of Diazo-benzene Nitrate on Methyl and Ethyl Alcohol in the Presence of an Excess of Zinc-dtist. A. 50 grams of diazo-benzene nitrate and methyl alcohol were ! decomposed in the presence of zinc-dust at a temperature below so'' C. Products : Benzene 7 grams, or 30 per cent. \ Diphenyl 3 grams, or 13 per cent. ; The contents of the flask which had been distilled in steam to \ remove the diphenyl were filtered from the remaining zinc-dust ; and the tarry products, evaporated to dryness and treated with ' hot alcohol in the hope of obtaining trioxymethylene or para- - formaldehyde, bui none was thus obtained. j B. 56 grams of diazo-benzene nitrate and ethyl alcohol were 1 decomposed in the presence of zinc-dust. The products obtained were diphenyl, 0.5 gram, and a mixture of paraldehyde and benzene. The ratio of these substances was ■ not determined, but the paraldehyde was broken down into \ aldehyde, and the remaining liquid converted into nitrobenzene. ; It will be seen that the results obtained in this series of decom- '; positions are in every way similar to those obtained by the action \ of diazo-benzene nitrate on the sodium alcoholates, and on methyl and ethyl alcohol in the presence of alkalis. They tend also to con- ' firm the view previously put forward, that the presence of any | substance which will combine with the acid at the moment of \ liberation leads to the hydrogen reaction. There are, however, 1 strong indications that the action is not between the free acid and | the neutralizing agent, but largely between the neutralizing agent « and the acid portion of the diazo-salt. For if either of the two :;.. diazo-compounds be brought together with methyl or ethyl % alcohol in the cold, only a slight reaction takes place, but when an ; alkali or zinc-dust is added an energetic action at once begins. \ This view would easily account for the formation of the diphenyl, \ but not for that of the benzene. The following reaction probably A takes place between zinc and diazo-benzene nitrate : j C6H5N.NO3 CeHs ' -f Zn = Zn(N03)s -f 2N« 4- I C6H5N2NO. CeHs On the Decomposition of Diazo- Compounds. 253 It therefore seems not improbable that there are two distinct reactions which take place : {a) one between the neutralizing agent and the acid portion of thediazo-compound leading to the forma- tion of diphenyl, and {b~) the other between the diazo-compound and the alcohol leading to the formation of benzene, the latter probably conditioned by the prevention of the acidity of the medium. If this view be correct, the acid set free in reaction {b) would combine with zinc and liberate hydrogen, which ought to reduce one-half of the aldehyde to alcohol. But this, of course, would not be true in the cases where alkalis were present, so there ought to be only one-half the percentage of aldehyde formed in the cases when zinc-dust was present as where the decompositions were made in the presence of alkalis. But the determinations of the paraldehyde were not sufficiently accurate in these cases to justify conclusions in regard to the point. A careful study of the action of these diazo-compounds upon water in the presence of zinc-dust would probably throw light upon this question. For if there be the two reactions, one should give diphenyl and the other phenol. Until further facts bearing upon this question are accumulated the question must remain open. It will be seen that more diphenyl is formed in those cases where methyl alcohol and the diazo-compound are decomposed in the presence of alkalis and zinc-dust, than in the corresponding cases where ethyl alcohol was used. For the preparation of diphenyl in large quantities I would suggest the decomposition of diazo- benzene nitrate and methyl alcohol in the presence of sodium carbonate. By comparing the results obtained in the two cases where diazo- benzene nitrate was decomposed with dilute ethyl alcohol it will be observed that the one which was diluted with 10 per cent, of water formed diphenyl, while the one diluted with 30 per cent, of water did not form the diphenyl. It would appear, therefore, that the presence of large quantities of water is unfavorable to the formation of the diphenyl. The formation of diphenyl by the decomposition of diazo- benzene nitrate in the presence of the sodium alcoholates, alkalis, and zinc-dust appears to be entirely similar to the iormation of this compound by the action of zinc-dust on diazo-benzene sulphate, as described by Gatterman.' ' Ber! d. chem. Ges. 23. 1226. 254 Beeson. Cojiclusions. The results which have been obtained in the course of this work justify, I think, the following conclusions : 1. When diazo-benzene nitrate is decomposed with methyl alcohol either at the room temperature or at the boiling-point of the alcohol, only the alkoxy product is formed. 2. Ortho-nitrophenol was also formed, probably by a molecular rearrangement of the diazo-benzene nitrate, with the loss of nitro- gen, and then was converted into 2:4-dinitrophenol by the action of the nitric acid set free in the course of the reaction, partially while standing in the cold, and completely while heating to distill off the alcohol and anisol. 3. Para-nitrophenol, when formed, was shown to be due to the presence of a small amount of water in the alcohol, first forming phenol, and then, by the action of the nitric acid present, forming both ortho- and para-nitrophenols. 4. In the case of the decomposition of diazo-benzene nitrate with ethyl alcohol, it was shown that an elevated temperature apart from pressure increased the alkoxy reaction, while a lower temperature increased the hydrogen reaction. 5. The presence of a small amount of water increased the ethoxy product. 6. In the decompositions of diazo-benzene nitrate with methyl and ethyl alcohols it was shown that in general the presence of an excess of alkalis led to the hydrogen reaction. 7. In the decompositions of diazo-benzene nitrate and sulphate with methyl and ethyl alcohols, the presence of an excess of zinc- dust led to the hydrogen reaction. 8. The decomposition of diazo-benzene nitrate with a solution of sodium methylate in methyl alcohol, and of sodium ethylate in ethyl alcohol, gave diphenyl and benzene. 9. In the three preceding cases there appear to be two reactions, one between the diazo-salt and the neutralizing agent, giving the diphenyl reaction, and the other between the diazo-salt and the alcohol, giving the hydrogen reaction. The latter reaction was possibly due to the prevention of the acidity of the medium. 10. The decomposition of diazo-benzene sulphate with methyl alcohol gave entirely the alkoxy product. Oxidation and Chemical Properties of Gases. 255 RESEARCHES UPON THE PHENOMENA OF OXIDA- TION AND CHEMICAL PROPERTIES OF GASES. By Francis C. Phillips. \^Continued from page 187.] II. — Qualitative Reactions of Gases. The recognition of any gas in a complex mixture is still a matter of difficulty in many cases, although in a few instances methods of identification are coming to be well known. Serious difficulties oppose all attempts at a system of qualitative analysis of gas-mixtures. There are but few groups of gases (if the name "group" be understood to include all gases chemically alike); moreover, the members of a group exhibit much closer relation- ships than are to be found among the metals of any one of the groups of Fresenius. The following classification of gases has been found convenient for purposes of study : Group I. Hydrogen. Group 2. Carbon monoxide. Group 3. Methane, ethane, propane, the butanes, etc. Group 4. Ethylene, propylene, trimethylene, the butylenes, etc. Group 5. Acetylene, allylene, etc. Group 6. Sulphur compounds : Hydrogen sulphide ; methyl hydrosulphide, (CH3)SH ; methyl sulphide, (CHsjsS ; carbon oxysulphide, COS ; carbon bisulphide. Group 7. Carbon dioxide. Unclassified : Nitrogen ; oxygen. In a study of the kind proposed, it "is of importance to take into account not only gases that are permanent under ordinary .condi- tions, but also vapors of liquids which are liable to occur in small quantities, such as carbon bisulphide, benzene, several of the lower paraffins and olefines, etc. Methods Eynployed. In the case of reactions between gases and solid substances, the solid to be tried was placed in a glass tube of i-in. diameter, which could then be heated to any given temperature in the iron oven previously described, while the gas was caused to stream through the tube. In the case of reactions in solution, two methods were used : 256 Phillips. 1. The gas was caused to flow through a capillary tube into the solution contained in a test-glass. The escaping gas could then be led into a second and, if necessary, a third test-glass in order to ascertain the action of the solution used in the first test-glass. 2. The gas was collected in glass-stoppered bottles over water, a small quantity of a solution introduced by means of a tap-funnel with the lower end of its stem bent upward, then the bottle closed and kept inverted at any given temperature for sufficient length of time (usually from a few days to several months) to ascertain if a reaction had occurred. The former method answers well for gases which are easily controlled in a slow, continuous stream and obtainable in large quantity. The latter method is more econom- ical as regards the gas to be used, and is better suited to gases where some slight but difficultly removable impurity is suspected to occur of a character liable to affect the regeanl — such as the traces of hydrocarbons present in hydrogen made from zinc and sulphuric acid. In such cases the smaller the volume of gas to be used in a trial the better. A reaction may usually be obtained with from 20 to 50 cc. of gas. Small bottles having well-ground, flat-topped glass stoppers answer well, as they may be kept standing inverted and may, if heat is to be applied, be placed inverted in boiling water. It is hardly necessary to add that, when inverted, such bottles may be used to hold gas in contact with a reagent for long periods without danger of loss. Hydrogen. Hydrogen for the following experiments was prepared and purified as already described.' Reactions were tried in bottles, and by causing the gas to bubble through the solutions, as just detailed. I. — Reactions in Solution. Reagent. Reactions. Palladium chloride The solution is slowly but completely reduced, cold or at 100°. The pre- cipitated palladium usually collects as a black powder. Sometimes it is deposited as a film on the glass. Platinum chloride** Very slow but complete reduction, cold or at 100°. The reduced metal appears as a black powder. ' Page 165. ' Mendeleeff: Principles of Chemistry, vol. II, p. 333- Oxidation and Chemical Properties of Gases. 257 Gold chloride Unchanged. Silver nitrate Unchanged if the fluid contains a trace of free nitric acid. Ammoniacal silver nitrate Slowly reduced, the silver appearing as a black powder. Iridium chloride Unchanged. Rhodium chloride Unchanged. Potassium rutheniate Slowly reduced. The orange color of the fluid disappears and metallic ruthenium is precipitated as a black powder. Cerium dioxide dissolved in dilute Unchanged, sulphuric acid. Potassium permanganate,' neutral. Extremely slow reduction, the purple color changing to brown. Permanganate, acidulated with sul- Bleached slowly, phuric acid. Permanganate, alkaline Slowly changes to brown. Potassium bichromate, acidulated Unchanged, cold or at 100°. with sulphuric acid. Mercuric chloride Unchanged. Osmic acid Unchanged. Prolonged contact in bright light yields traces of reduc- tion after two or three weeks. Ferric chloride Unchanged cold. Traces of reduc- tion to ferrous chloride after heat- ing for several hours at 100°. Potassium ferricyanide Unchanged. Ruthenium chloride Unchanged. Nitric acid, fuming^ Unchanged. Comments. — RusselP states that hydrogen reduces silver-nitrate solution, nitric acid being at the same time reduced to nitrous acid. Pellet* finds that this reduction is due to the silver salt containing silver oxide in excess, but that perfectly neutral silver nitrate is not altered. In a series of experiments I have obtained results corroborating those of Pellet. Silver nitrate containing a minute trace of free nitric acid is not altered by hydrogen. If some freshly precipitated and washed silver oxide is digested with solution of silver nitrate, and the liquid then filtered, it will have an alkaline reaction towards litmus and is slowly reduced by ' Meyer and Askenasy [Ann. Chem. (Liebie) 269, $6 (1892); Ber. d. chem. Ges. 25, 410 (Ref.)] find that electrolytic hydrogen reduces potassium permanganate. ^Winkler: Ztschr. anal. Chem. 28, 269 (1889!. •J. Chem. See. 27, 3 (1874). ,, these salts are unal- Iridium chloride ...V^ j u 1 j Stered by prolonged Rhodium chloride / ^ ^ .^, ^, . contact with methane, Cerium dioxide dissolved in sulphuric acid u ^ c ^ \ cold or at 100'^ Potassium bichromate acidulated with sulphuric acid ' Mercuric chloride Ferric chloride Ruthenium chloride ' Darvydowa: J. russ. phys.-chem. Ges. 1888, 362; Ber. d. chem. Ges. 81, 442 (Ref."* (1888). 'See p. 172. Oxidation and Chemical Properties of Gases. 263 Potassium permanganate (2-per cent, solution) is unchanged, whether neutral or acidulated by sulphuric acid. Osmic acid is not reduced by methane in the cold. Potassium ferricyanide is unchanged. Peroxide of hydrogen mixed with lime-water remains clear, proving that no oxidation to carbon dioxide occurs. Calcium-hypobromite solution remains free from any deposit of calcium carbonate. Bromine-water is not decolorized after prolonged contact. Chlorine attacks methane only at a temperature considerably above 100°. A mixture of methane and chlorine was exposed over water to bright sunlight on a July day without undergoing any noticeable contraction in volume or change of color.' Potassium rutheniate is slowly reduced with separation of metallic ruthenium. If methane is conducted into strong sulphuric acid to which crystals of permanganate have been added, immediate oxidation to carbon dioxide occurs, as proved by the action upon lime- water (in this experiment stoppers made of plaster of Paris were used). This reaction towards permanganic anhydride is a very delicate one. All hydrocarbon gases yield a similar result. 2. — Reactions at High Temper attires. Ferric oxide (prepared by ignition of ferric nitrate) heated in a glass tube over a strong Bunsen-burner flame underwent very slow and incomplete reduction, some carbon monoxide being formed in addition to carbon dioxide. Iodic acid (crystals) is not reduced by methane on heating nearly to its temperature of dissociation (250°), Neither iodine vapors nor carbon dioxide is formed. The action of methane upon the chloride, bromide, and iodide of silver was tried by the same method followed in the case of hydrogen — these substances contained in a glass tube being heated in methane and the gas then conducted into silver-nitrate solution. The temperatures of decomposition were as follows : Temperature of Decomposition. Silver chloride At melting-point of barium chlorate (414°). Silver bromide Above melting-point of thallium iodide (439°). Silver iodide Slightly volatile without reduction. 1 See experiments in " Chlorination of Methane." 264 Phillips. The order of reducibility by hydrogen and by methane is the same, therefore, as in the case of the action of light, the chloride being the most easily reduced, the iodide the most stable. Nickel chloride heated in natural gas (from Murrysville) under- went conversion into its nearly colorless, beautifully crystalline modification. An analysis made in the laboratory by Mr. H. T. Weed showed the composition of the salt to have been unchanged. At a dull-red heat reduction occurred, with liberation of carbon. Ethayie. This hydrocarbon was prepared from ethyl iodide by the process of Gladstone and Tribe, as described on p. 173. I. — Reactions in Solution. All the reagents used in the case of methane were tried. The reactions were so closely similar that a detailed statement is omitted as unnecessary. Kthane exhibits the same stability as the other paraffins towards reagents in solution. Oil of vitriol containing crystals of potas- sium permanganate causes prompt oxidation to carbon dioxide. Potassium rutheniate is quickly reduced with separation of metallic ruthenium. 2. — Readiojis at High Temperattcres. Towards iodic acid, silver bichromate, and the various metallic oxides, ethane closely resembles methane in its reactions. Propane. — Reactions. Experiments with propane made by the method of Gladstone and Tribe from propyl iodide, led to results closely similar to those obtained with ethane and methane. As already stated, propane is somewhat more easily oxidized than the paraffins lower in carbon. In a series of trials, using the same reagents in solution as mentioned in the preceding experi- ments, no reactions were obtained which would serve to distinguish between propane and methane or ethane. The results are there- fore omitted. IsobiUane. — Reactions. This hydrocarbon, prepared by the method of Gladstone and Tribe from isobutyl iodide, was found closely to resemble in its Oxidation and Chemical Properties of Gases. 265 reactions the paraffins already described. It is characterized, however, by a lower oxidation-temperature, as already stated in regard to the experiments with palladium asbestos. Heptane. Heptane obtained from "theoline"' was found to have the same general chemical properties as methane. Even cerium dioxide, osmic acid, gold chloride and potassium permanganate are unal- tered by prolonged contact with the liquid hydrocarbon. Hep- tane seems in fact to be almost, if not quite, as stable as methane towards reagents in solution. The paraffins as a group are, in the main, so proof against reac- tions that we can do little more than remove all other hydrocar- bons by suitable reagents and then test for paraffins by combus- tion over oxide of copper to carbon dioxide and water. This, of course, leaves the nature of the individual paraffins undetermined. Spongy palladium, heated in air so as to became partially con- verted to palladium oxide and transferred to an atmosphere of methane or other paraffin, undergoes a reduction. The reduced metal then combines with the carbon of the hydrocarbon." This carbide heated in air or oxygen yields carbon dioxide. The pro- duction of carbon dioxide in this case may be utilized as a test for hydrocarbons in a gas-mixture, provided no free oxygen is pres- ent. The carbide of palladium, formed by the above method, dissolves in aqua regia (containing but little nitric acid) with a camphor-like odor. Olefines: Ethylene. For preparation, see p. 176. The method of Erlenmeyer and Bunte was used. It is necessary for the purpose of studying its reactions to purify the gas from traces of the vapors of alcohol and ether by prolonged digestion with sulphuric acid. I. — Reactions in Solution. Reagents. Reactions. Palladium chloride Quickly reduced, the metal appearing as a black powder. No carbon dioxide is formed. Platinum chloride Unchanged. J See p. 175. * Graham-Otto : Vol. Ill, p. 995; Wilm : Ber. d. chem. Ges. 35, 220 (1892). / 266 Phillips. Gold chloride Extremely slow reduction, the gold ap- pearing as a brown powder. No car- bon dioxide. Gold chloride in excess of potas- Extremely slow reduction, sium hydroxide. Iridium chloride Unchanged. Ruthenium chloride After prolonged contact (several days) the solution is bleached. No deposi- tion of metal occurs. Rhodium chloride Unchanged. Silver nitrate Unchanged. Silver nitrate in ammoniacal so- Unchanged. lution. Potassium permanganate, neu- Quickly turns brown. tral solution. Potassium permanganate acidu- Quickly bleached. lated with sulphuric acid. Potassium permanganate crys- Prompt oxidation to carbon dioxide, tals in concentrated sulphuric acid. Potassium bichromate acidu- No change of color, cold or at ioo°, lated with sulphuric acid. Osmic acid Quickly reduced, with separation of metal as a black powder. Potassium rutheniate Quickly reduced, with separation of metal. Ferric chloride No change, cold or at ioo°. Calcium hypobromite contain- No precipitation of calcium carbonate, ing excess of lime-water. and hence no oxidation to carbon diox- ide. Potassium ferricyanide Unchanged. Bromine-water Rapid but incomplete absorption. Peroxide of hydrogen No oxidation to carbon dioxide. 2. — Reactions at High Temperatures. Silver oxide is reduced by ethylene with simultaneous forma- tion of silver carbonate at 140°.' Palladium chloride (dry) is reduced at about 140°. Iodic acid is reduced with liberation of iodine at about 270°. Comments. — As is well known, bromine-vapor and ethylene combine to form an oily liquid by the reaction so characteristic of the olefine group. Winkler' has shown that the absorption of ethylene by bromine is incomplete, and that the contraction in ' Darvydowa: J, russ. phys.-chem. Ges. 1888, 362 ; Ber. d. chem. Ges. 21, 442 (Ref.) (1888). - Ztschr. anal. Chem. 28, 269 (1889). Oxidation and Chemical Properties of Gases. 267 volume is by no means proportional to the volume of the ethylene present. I have tried experiments upon ethylene prepared from alcohol and from ethylene dibromide (by the action of zinc powder), and the results show that a considerable residue of hydrocarbon remains unabsorbed after prolonged contact with bromine-water in sunlight. The residual gas, on being mixed with air and passed over ignited oxide of copper, gave carbon dioxide and water at the outset. The reaction between ethylene and palladium chloride in solu- tion is of the second class and complete, the gas being rapidly absorbed. Palladium is deposited as a black powder, but no trace of oxidation to carbon dioxide occurs. The reaction is almost the same in the cold and at 100°. The gas escaping from the palladium-chloride solution (after complete reduction to metallic palladium) produces no precipitate in lime-water. The reaction between palladium chloride and ethylene leads to the production of aldehyde.' Of especial interest in this connection is a statement by Berthelot,'' that ethylene is oxidized to aldehyde by the action of chromic-acid solution at 120°. Gold chloride produces a similar result, the metal being slowly reduced. As in the case of palladium chloride, no carbon dioxide is formed. Rhodium chloride is remarkably stable towards ethylene (and other olefines) : after three months' contact with the gas, no trace of reduction was observed. Potassium permanganate (in weak solution) has been shown by Wagner' to convert olefines on digestion (cold) into glycols. The reaction may serve as a mode of preparing glycols, but could only in exceptional cases be utilized as a gas-reaction. Chromic-acid mixture is said to be reduced by ethylene.* In repeated experiments I have failed to show that chromic acid undergoes reduction by ethylene. No carbon dioxide is formed, the color of the solution remains unchanged, and no absorption occurs on prolonged contact of ethylene with a lo-per cent, solu- tion of chromic acid in a eudiometer. Similar results were 'A study of the changes here involved is yet in hand. 'Compt. Rend. 68, 334. »Ber. d. chem. Ges. 21, 1230 (i883). ^ sr \ ^ \ \ , 'V N, \, \ ^ \ \ N N \ \ \\ 775 ^ ■^ — N y ■ 75^ -^ ^ ^ V 7? ~^ "^^ --0 ^ / ~^^ ^ / Fig. 3. — Effect of Change of Temperature on Acetic Acid (above) and Sodic Hydrate (below). Curves started from same point. To eliminate the effects of the temperature on the diastase, and to represent graphically the effects of the temperature on the alkali, the curves were started from the same point. The regular approach of the curves to the axis of abscissae was thus rendered more apparent, crossing of the lines being obviated. Acetic acid acted in general in the same way, but the effects of changes in the temperature appeared to be much greater than in the case of the alkali, as will be seen in the diagram. These results show the necessity of working under exactly the same conditions, for a change of 0.5° in the temperature may be equivalent to a change often per cent, or more in the amount of the acid. Affinity- Constants of Acids. The action of the fatty acids on diastase was tried, with a view of determining the affinity-constants, and comparing them with the figures obtained by other methods. Hydrochloric acid was first tried, to obtain a standard for comparison. The solution was made equivalent to the sodic hydrate (0.0008 gram to the litre) by titration with phenolphthalein. For convenience I shall call o.i cc. of this solution one part. Quantities up to 0.6 cc, or 6 parts, were found to produce an increased yield of sugar. This effect has been noticed by Duggan' and others, and is probably due to a slight alkalinity of the starch. Above this point the acid >"0n the Determination of Absolute Neutrality" : This Journal 8, 211. The Ajffiftiiy- Cojisianis of Weak Acids. 319 \ \ 78 s ^^ .^ fc. 72 \ N A ^r 1 66 \ s ^ \ ^ \ ^ \ . 60 \ \^ r\ [^ :\ \ \ 1 •6 54 \ c N \ \ \ i < 48 V s N N \ i2 42 N N K ^ N 1 <2 36 N Ann. der Phys. Pogg. 136, I ; 135,177. 2 Am. J. Sci. 45, 478. 334 J<(^sil^- salts dissolve in water, for example, the iron ion Fe'" being one of the weakest would take up the hydroxy] from three parts of water to form hydroxide of iron, which would remain in the solu- tion in the colloidal state, while the resulting hydrogen would unite with the anion (N03)3 to form free nitric acid.' It will be observed that none of these facts and explanations seem capable of accounting for the differences in color observable in concentrated and dilute solutions of ferric salts. Of the colloidal modifications of iron hydroxide, two have already been isolated, one brown in color, the other yellow. Upon the assumption of the existence of still another variety of colloidal iron oxide, red in color, it would seem that every fact concerning the changes in the color of ferric salt-solutions might readily be explained somewhat as follows : When a colorless ferric salt dissolves in a small quantity of water it partially dissociates into colorless ions of iron and an acid residue. The ions of iron at once react with a certain portion of the water present to form iron oxide or hydroxide, red in color, and free nitric acid. As the dilution proceeds it may not be unreasonable to suppose that not only more dissociation of the iron salt takes place, but that also a portion of the iron oxide, or hydroxide already produced becomes so greatly hydrated as to produce a quantity of the yellow hydroxide, which mixing with the red variety gives the reddish-brown modification commonly existing in solutions of iron salts ; and, finally, as the dilution becomes very great, not only would the dissociation reach its maximum, but also the hydration of the hydroxide produced, until at last only the yellow variety of the hydroxide remained in solution. It would thus appear that all facts concerning the colors of ferric salts in solution, and the changes in color which these same salts undergo by alterations in the quantity of solvent and physical conditions, could be correctly interpreted upon the assumption of the existence and formation in solution, of a red, hydrated oxide of iron, soluble in water. Concerning this assumptiom it may be said that the fact that such a body has never yet been isolated is no proof of its non- existence; and, further, it has doubtless struck every one who has ever given the matter any attention, that were it possible to dissolve the purplish-red varieties of hematite in a neutral solvent, ' Oswald's Lehrbuch, 2, 794. The Color of Salts in Solution. 335 such as water, its solution would doubtless be of a deep blood-red color. There are several dyestufifs, dissolving in water with the production of a deep-red solution, which in their solid state are not unlike the purer varieties of hematite in appearance. The change which very dilute and yellow-colored solutions of ferric salts undergo on heating is in harmony with the views enunciated above. Such a change is attended with the production of a deeply colored red-brown solution. According to this hypothesis, such a change would signify that the greatly hydrated yellow variety of the hydroxide had lost a large portion of its water of hydration in consequence of the heat, and had therefore given rise to the less hydrated, red-brown modification. This view is strengthened by the fact that copper hydroxide loses its water on being gently heated, even in the presence and in contact with a thousand times the quantity of water necessary to form it ; and by the fact that all hydrates, crystalline and amorphous, part with their water of hydration on being heated. The decoloration of certain ferric salts, for example the sulphate and the nitrate, by means of strong mineral acids, particularly nitric acid, is a fact so well known that it is taken advantage of in one of the most reliable methods of volumetric analysis. When strong nitric acid in sufficient quantity is added to a solution of iron nitrate it becomes as clear and colorless as water; upon standing for a time, however, it assumes a very pale yellow color, and then resembles precisely a very dilute solution of iron nitrate alone. Again, the explanation of this phenomenon is furnished by the above hypothesis : when nitric acid is added to a colored solutipn of iron nitrate it at once combines with the hydroxide of iron in the solution to form iron nitrate, which in the undissociated con- dition is colorless. In the course of time, however, the newly formed nitrate will undergo a slight dissociation, and as a result there will again be formed a small quantity of iron hydroxide. This now, finding itself in a relatively large quantity of water, is converted into the yellow hydroxide peculiar to very dilute solu- tions of ferric salts. It is believed that these considerations apply with equal force to the explanation of the facts observed in the case of the colored solutions of other salts. For example, as above cited, the following facts have been observed concerning the color of various copper salts in water : 336 Kastle. 1. The color of copper salts in solution is either blue or green or a mixture of these two colors. 2. These solutions become lighter upon dilution. 3. These solutions become darker on heating. 4. That of four of these salts — viz., the sulphate, nitrate, chloride and acetate — the acetate, at corresponding dilutions, is much the darkest and strongest in color of the four. The acetate was found by Vernon' to be about eight times more densely colored, whereas between the chloride, nitrate and sulphate the difference in color was but slight. Now, on the supposition that two soluble hydroxides of copper exist in solutions, we have a ready explanation of the first of the above general conclusions, viz., that these solutions are either green or blue. Secondly, the hydration of each of these hydroxides upon dilu- tion of the solution, would certainly account for the fact of their becoming lighter in color on dilution. Thirdly, that these solutions become darker on being heated is, in the light of this hypothesis, to be accounted for upon the sup- position that these highly hydrated and lighter hydroxides lose some of their water upon being heated. And, lastly, the fact that the acetate forms a much more deeply colored solution than either of the other salts, is not only expli- cable by, but furnishes strong support for, this hypothesis. We may suppose that, as with other salts, the copper acetate dissoci- ates ; the acet ion, however, is so weak that it very readily unites with the water present to form acetic acid. The setting free of the hydrogen of the water, however, and the formation of acetic acid involve a corresponding formation of the base, and that, too, in larger quantities than it is formed in the case of stronger acids; and hence, owing to the larger quantity of copper hydroxide, the solution of copper acetate is darker than either the chloride or the sulphate. In other words, according to this hypothesis, whatever changes occur in the molecule of copper acetate when it goes into solution, these same changes take place with the other salts of copper, only to a greater or a less extent, depending upon the nature of the acid with which the metal finds itself in combination : and all of these changes, by virtue of which a salt produces a colored solu- > Chem. News 66, 152. The Color of Salts in Solution. 337 tion, are alike in kind and differ only in degree. For example, if we have in the case of copper acetate not only a dissociation but also a subsequent and dependent formation of the base copper hydroxide ; so in copper chloride or sulphate we have not only a dissociation of the molecule into its ions, but also a subsequent decomposition of a part of the water present by means of one or both of these ions, whereby not only ions of copper and acid residues find themselves present in the solution, but also some free base and some free acid. It is upon this decomposition and its extent that the kind of color and also its intensity depend in the case of salts in solution. The only essential difference between the color of a solution of copper acetate and that of a solution of the sulphate would depend upon the extent to which one or the other of these processes had been caiTied. In the case of copper acetate the dissociation is small, whereas the conversion of copper ions into the hydroxide is relatively great ; on the other hand the dissociation in the sulphate is large, and the conversion of copper ions into the hydroxide is relatively very small. It would seem that this view that ail salts which give colored solutions have not only been dissociated, but partially decomposed, by water, derives a certain support from the fact that of the many salts forming colored solutions but few, if any, are neutral. Most of them are acid in reaction. According to the advocates of the dissociation theory, a solution of potassium carbonate is alkaline for the reason that a base and an acid are both formed in solution, and of these two substances the base is much the stronger and hence gives to the solution its own reaction. It would seem that the converse of this ought to be true, viz. that the non-neutrality of a solution of a substance shows that not only has dissociation taken place, but also the subsequent formation of the acid and base from the salt. The marked acidity of a solution of copper sulphate, for example, ought to show that this salt has been decomposed by water, giving rise to a small quantity of the base Cu(OH)2 and the acid H2SO4. Of these two substances, however, the sulphuric acid is the stronger, and hence the acidity of the solution; on the other hand, the acid being colorless and the base blue, the latter manifests its existence by imparting to the solution its characteristic color„ Again, in support of the view that we really have free bases in the solutions of colored salts may be cited the fact that many if 338 Kastle. not all of these metals tend to form basic salts, in which the pres- ence of the hydroxide has been proven beyond a doubt; and it is interesting to note in this connection that in the case of copper salts, those which are basic have, as a rule, a green color, whereas the normal ones are blue. It is certainly difficult to see how these basic salts could have been formed from a solution of the normal salts, if prior to their formation the hydroxide in some form or other had not existed in. the solution. It may be argued that this view concerning the existence of the free acid and base in solutions of colored salts is inadmissible, on the ground that such solutions show none of the general proper- ties of an acid or an alkali, such as characteristic taste, smell, etc. To this it may be replied, that the quantity of an acid or base, either directly or indirectly concerned in the production of a color, may be altogether too small, and too much obscured by the properties of the salt as a whole, to be recognized by either of these senses. As a matter of fact it is well known that the quantity of an acid or of a base which will affect the color of an indicator is as a rule entirely beyond the reach of either taste or smell. Looked at from another standpoint, it is very much to be doubted if there is anything like absolute neutraliiy ^.mox\ Keiser : This Journal 14, 285 (1892). -' Bull. Soc. chiin. [2] 5, 191 (i866). • See page 370. Oxidation and Chemical Properties of Gases. 343 being passed through boiling alcoholic potash solution and absorbed by ammoniacal cuprous chloride. The yellow precipi- tate resulting was washed and afterwards decomposed by hydro- chloric acid, allylene being then set free in a pure state. I. — Reactions in Solution. Reagents. Reactions. Palladium chloride Uark-brown precipitate.whichmay be preserved without decomposition. Closely resembles the precipitate produced by acetylene. Platinum chloride Unchanged. Gold chloride Slowly reduced. The color of the precipitated gold is very dark. Silver nitrate A lo-per cent, solution quickly coag- ulates to a white, curdy mass. The precipitate dissolves on boil- ing or on addition of ammonia, A very delicate reaction. Ammoniacal silver nitrate Unchanged. Iridium chloride Unchanged in the cold.at 100° iridium is precipitated. Rhodium chloride Unchanged. Potassium rutheniate.. Slowly reduced. Black precipitation of metallic ruthenium. Cerium dioxide dissolved in dilute Unchanged, sulphuric acid. Potassium permanganate. Quickly turns brown. Potassium permanganate in dilute Quickly bleached, sulphuric acid. Potassium permanganate crystals in Prompt oxidation to carbon dioxide, concentrated sulphuric acid. Mercuric chloride Dense white precipitate. Very del- icate reaction. Potassium bichromate acidulated Unchanged in color, with dilute sulphuric acid. Osmic acid Reduced. Metallic osmium is depos- ited as a black powder. Ferric chloride Decided reduction to ferrous chloride. Calcium hypobromite Allylene is oxidized to carbon diox- ide. The fluid grows milky. Lime-water and hydrogen peroxide. Unchanged. Potassium ferricyanide Unchanged. Iodine dissolved in potassium-iodide Unchanged, solution. Cuprous chloride in excess of ammo- Canary-yellow precipitate, changing nia. slightly to greenish-yellow on con- tact with air. Soluble in acids, with liberation of allylene. Mercurous nitrate White precipitate. 344 Phillips. 2. — Reactio7is at High Temperatures. Experiments were tried in the reduction of certain metallic oxides, but the results are not of sufficient importance to be detailed here. Commetiis. — The reactions of allylene closely resemble those of acetylene. As regards intensity, scarcely any difference can be found. The colors of the palladium compounds of acetylene and allylene do not differ materially. Towards ammoniacal cuprous chloride the two gases exhibit very characteristic differences as regards the color of the resulting compound. The copper allylide is easily soluble in dilute hydrochloric acid. Ammoniacal silver nitrate yields a gelatinous precipitate with acetylene, but is not changed by allylene. Oxidation of allylene to carbon dioxide, as in the case of acetylene, is not easily effected except by the most powerful oxidizing agents, such as calcium hypobromite or potas- sium permanganate in concentrated sulphuric acid. Although the allylene copper compound is rapidly formed in an ammoniacal cuprous-chloride solution, the absorption of the gas is singularly incomplete. Agitation with the solution is .quite necessary in order to insure complete absorption. Wagner' has shown that the higher acetylenes, like the olefines, are converted by neutral potassium-permanganate solution into hydroxyl compounds. The various classes of hydrocarbons of the fatty series possess in common a high resistance to destructive oxidation by oxidiz- ing agents, yielding in some cases hydroxyl compounds, but rarely carbon dioxide. This is true also of benzene, which is changed by potassium permanganate into oxalic acid and formic acid.' Sulphur Compounds. Carbon Oxy sulphide. This gas was prepared by the method of Klason/ To a cold mixture of 290 cc. sulphuric acid and 400 cc. water, 50 cc. of a saturated solution of sulphocyanide of potassium was gradually added, the mixture being warmed to 30°. The gas was evolved in a steady stream and was purified by passage (i) through 20- per cent, potash solution, (2) through 25-per cent, solution of aniline in alcohol, (3) through broken ice. 1 Ber. d. chem. Ges. 31, 3343 (1888). - Bernihsen: p; 326. 'J. pmkt. Chem. [2] 36, 64 (1887^- Ber. d. cheni. Ges. 20, Ref. 550. Oxidation and Chemical Properties of Gases. 345 Carbon oxysulphide was also prepared by the action of carbon disulphide on alumina at a high temperature.' The gas, if dry, may be preserved over mercury. Contact with water causes a change into carbon dioxide and hydrogen sulphide. Caustic soda solu- tion is changed into a mixture of sodium sulphide and carbonate. The constant tendency to decomposition renders it impossible to preserve the gas over water without loss. In trying its reactions, it was found necessary to conduct the gas immediately before use into some substance specially adapted to absorb hydrogen sulphide. For the absorption of this gas Fresenius' recommends pumice saturated with copper-sulphate solution and dried. In a series of trials with this and other absorbents, precipitated oxide of mercury was found to answer best. Dampened absorbent cotton is coated with the yellow powder by rubbing with a large pestle. This preparation, used dry in a long glass tube, completely removes hydrogen sulphide, but exerts no action upon carbon oxysulphide. I. — Reactions m Sohdion. Reagents. Reactions. Palladium chloride Prompt precipitation. Precipitate is brownish-black and flocculent. Platinum chloride Black precipitate. Gold chloride Rapidly darkens. An olive-brown precipitate collects. Copper sulphate Black precipitate, which forms very slowly. Ammoniacal copper sulphate Black precipitate, forming promptly. Silver nitrate Voluminous, brownish-black preci- pitate, Ammoniacal silver nitrate Prompt precipitation. Cadmium chloride Precipitation is slow and incomplete. Cadmium chloride in excess of am- Rapid and complete precipitation ; monia. bright yellow. Arsenious chloride Yellow precipitate, forming very slowly. Potassium permanganate acidulated Rapidly bleached. The solution then by hydrochloric acid. precipitates barium chloride, so that oxidation to sulphuric acid has occurred. No separation of sul- phur is observed. Lead acetate Black precipitate. Bromine-water Prompt oxidation to sulphuric acid. No sulphur is separated. » Gautier : Compt. Rend. 107, 911. ^Quant. Analyse, 6te Auflage. Vol. XVI.-26. 346 Phillips. Mercuric nitrate Turns milky white and darkens grad- ually to black. Nickel hydrate in water Darkens slowly to black. Ferric chloride Decided but incomplete reduction to ferrous chloride. Potassium ferricyanide Traces of reduction to ferrocyanide. Osmic acid Rapidly reduced. Potassium rutheniate Rapidly blackened. Cerium dioxide in dilute sulphuric Rapidly bleached, acid. Yellow oxide of mercury Litharge Precipitated carbonate of copper These are all unchanged if dry. Sealed in a glass tube filled with the gas for two months, no change "White lead i had occurred. If spread upon cot- I ton, these substances may be used L to remove sulphuretted hydrogen from a mixture of the two gases. Silver foil Is unattacked if dry ; under water it is quickly blackened. Comments. — Carbon oxysulphide, by reason of its ready change- in presence of water into hydrogen sulphide and carbon dioxide, gives, in the main, the reactions of hydrogen sulphide, so that by the reagents usually employed for the detection of the latter gas these two sulphur compounds are not distinguishable. In fact, towards ammoniacal cadmium-chloride solution, silver nitrate and palladium chloride, its reactions are characterized by somewhat greater promptness than in the case of hydrogen sulphide. It is to be noted that, should hydrogen sulphide and carbon oxysulphide occur in a gas mixture, the latter would, by ordinary analytical methods, be mistaken for the former, and equal volumes of the two gases would yield the same weight of the sulphides of silver, cadmium, copper, etc. The principal metallic sulphides as usually formed by hydrogen sulphide could be, in many cases, more rapidly produced by carbon oxysulphide ; and, so far as I have been able to observe the reactions, there is little or no tendency to separation of free sulphur such as is common in precipitations by hydrogen sulphide. The absorption of carbon oxysulphide in ammoniacal cadmium solution is so complete that on passing the gas through this solu- tion none will escape unabsorbed to cause precipitation in a second solution. An analysis seemed desirable in the case of the silver compound obtained when carbon oxysulphide was passed through Oxidation and Chemical Properties of Gases. 347 a solution of ammoniacal silver nitrate, as the precipitate appeared much more flocculent and of a more brownish color than ordinary, precipitated silver sulphide. Accordingly, determinations of silver and of sulphur were made with the following results : Calculated for AgjS. Ag 87.06 S 12.94 Found (I) (2i 91-15 90.95 8.70 8.75 100.00 99-85 99-70 Hence the compound consisted of silver sulphide, with a small quantity of silver thrown down by the carbon monoxide present in the gas. Yellow mercuric oxide forms an excellent means of separation of the two gases, and after the removal of the hydrogen sulphide by this reagent used in a dry state, the production of a precipitate in ammoniacal cadmium-chloride solution would indicate that this cadmium sulphide has been caused by carbon oxysulphide. The presence of a little carbon monoxide in the carbon oxysul- phide made from potassium sulphocyanate and sulphuric acid, is liable to mislead in the reaction towards palladium chloride, causing a black precipitate of palladium resembling the sulphide. Methyl Hydro sulphide, CHs.SH. This gas may be produced by several typical reactions : (i) When methyl chloride (bromide or iodide) is heated with potassium hydrosulphide in alcoholic solution, the reaction being CHsCl -h KSH = KCl 4- CH3SH. Methyl chloride gas was conducted into a boiling alcoholic solu- tion of potassium hydrosulphide contained in a tube of the shape here shown. The long limb of the tube (length, thirty inches) was connected with a reversed condenser and was heated over a small flame. The gas as it escaped was passed through a long glass tube containing cotton coated with red oxide of mercury, which absorbs any pos- sible traces of hydrogen sulphide and some of the mercaptan. The gas was passed through broken ice. As the reaction above men- tioned is rather incomplete, the gas contains much unaltered methyl chloride. The same is true when the bromide and iodide Oxidation and Chemical Properties of Gases. 349 are used. Methyl chloride is the best suited to the purpose, since it may be conducted into the liquid as a gas. The iodide, being a very volatile liquid, is not easily added without danger of tumul- tuous boiling. Formation of difficultly soluble potassium chloride or iodide causes clogging and greatly interferes with the process, even when large delivery-tubes are used. (2) Methyl sodium sulphate and potassium hydrosulphide, on being brought together and warmed, yield the following reaction: NaCHsSO. + KHS = KNaSO* + CHsSH. The reaction may be carried out in aqueous solution by the very excellent method of Klason.' He directs as follows : 800 grams potassium hydroxide are dissolved in water; the solution is satu- rated by sulphuretted hydrogen. It is placed in a large flask and sodium methylsulphate, made from 500 cc. methyl alcohol, added in small portions. On gently warming, a mixture of the vapors of methyl sulphide and hydrosulphide is evolved. The vapors are passed through an empty bottle and then into a second bottle containing soda solution, which should be cooled. The methyl hydrosulphide is completely absorbed by the soda, forming sodium methyl sulphide, or mercaptide, NaCHsS. The sodium mercap- tide so produced is very stable. Methyl sulphide condenses to a liquid which floats on the soda solution, but does not combine with the soda. A separation is therefore easily effected. The methyl sulphide may be driven off" by warming the bottle contain- ing the soda solution, the mercaptide being unaffected by the heat. The methyl sulphide may thus be used in vapor-form to produce its reactions. After expulsion of the methyl sulphide, the soda solu- tion may be placed in a flask and decomposed by dilute sulphuric acid, and methyl hydrosulphide then expelled as a gas. Some lead acetate is added to the solution in order to bind sulphuretted hydrogen during the decomposition of the sodium mercaptide by acid. Klason advises a further purification ; but, by the process described, the two sulphur compounds may be obtained of suffi- cient purity for the study of their reactions. Methyl hydrosul- phide is a gas above 6° C. (Klason). It is remarkable for its penetrating odor, which adheres most tenaciously to all surfaces, glass not excepted, for months. All work with the gas should be > Ber. d. chem. Ges. 20, 3407 (1887). 350 Phillips. done out of doors. It is somewhat soluble in water, to which it imparts its properties. The compounds produced by the action of methyl hydrosul- phide upon metallic oxides are the true mercaptides. Typical of these is the mercury mercaptide, which results as follows : HgO + 2CH3SH — (CH>S).Hg -f H2O. The yellow lead compound has the formula (CH3S)2Pb; the copper compound, CHsSCu ; the silver compound (CH»S)Ag.' Methyl hydrosulphide combines also with metallic chlorides : HgCh + CHsSH = CHsSHgCl + HCl. According to this reaction, numerous metallic compounds are formed. These compounds change more or less readily, on exposure to air, into metallic sulphides (Klason). Reactions. The vapor was caused to bubble through various solutions with the following results : Reagents. Reactions. Palladium chloride Cinnamon-colored, flocculent precipi- tate in strong or weak solutions. Insoluble in hydrochloric acid, nitric acid, sulphuric acid, aqua regia, ammonia and caustic soda, in the cold or at 100°. Extremely delicate reaction. Platinum chloride Yellowish-brown, flocculent precipi- tate in dilute or concentrated solu- tion. Insoluble in the strong acids and alkalies, and in this respect similarto the palladium compound. Iridium chloride Yellow precipitate, resembling in appearance the platinum com- pound. Gold chloride Light yellow, very voluminous pre- cipitate, changing gradually to white as the passage of the gas is continued, and finally redissolving to a clear solution. Mercuric chloride White, flocculent precipitate. Dark- ens slightly on exposure to air and light. Extremely delicate reaction. ' See Klason: loc. cit.\ and Richter : Org. Chem., trans, by Smith, p. 143. Oxidation and Chemical Properties of Gases. 351 •Copper sulphate Straw-yellow precipitate, insoluble in ammonia. Darkens rapidly. Soluble in hydrochloric acid. The hydrochloric-acid solution of the precipitate contains cuprous chlo- ride. Ammoniacal copper sulphate Yellow precipitate like the preced- ing. Soluble in hydrochloric acid to cuprous salt. Rapidly darkens. •Cuprous chloride White, flocculent precipitate, chang- ing to crystalline needles. More stable than the preceding com- pound. Silver nitrate Yellow precipitate, resembling in appearance the copper compound. Insoluble in ammonia. Rapidly blackens. Ammoniacal silver nitrate Yellow precipitate, similar inappear- ance and properties to the preced- ing. Ammoniacal cadmium chloride White precipitate in flocculent masses, somewhat soluble in the reagent and in water. Permanent, if protected from the air. By oxidation is converted readily, in the cold, into yellow cadmium sul- phide. Arsenious chloride in dilute hydro- The fluid grows milky from floating chloric acid. oil drops, which gradually collect as a very heavy oil at bottom. Zinc sulphate in excess of caustic Unchanged, soda solution. Potassium permanganate, 6-per cent. Rapidly bleached. Becomes heated solution acidulated by hydrochloric from the intensity of the reaction, acid. No sulphuric acid is produced. Lead acetate. Straw-yellow precipitate, insoluble in acids and alkalies. Rapidly blackens. ' Potassium bichromate acidulated by Promptly reduced to green chromic hydrochloric acid. chloride. No sulphuric acid is produced. Bromine-water Rapidly bleached. No sulphuric acid is produced. Mercurous nitrate Grayish-black precipitate. Bismuth nitrate Slowly forming black precipitate Kickel hydroxide in water Slowly blackens. 352 Phillips. Ferric hydroxide in water Unchanged. Yellow oxide of mercury Turns slowly gray and finally black. Ferric chloride Rapidly reduced to ferrous chloride. No sulphuric acid is formed and no sulphur liberated. Potassium ferricyanide Reduced to ferrocyanide. Osmic acid Rapidly blackened. Potassium rutheniate Extremely slow and incomplete re- duction. Hydrogen peroxide Nooxidation tosulphuricacid occurs. Cerium dioxide in dilute sulphuric Quickly bleached, acid. Litharge and white lead Quickly changed, yielding a volumi- nous yellow powder. Copper carbonate The resulting mercaptide is similar in appearance to that obtained in the preceding reaction. The mer- captides of lead and copper are very stable. Silver foil Is not changed, dry or in water. After three months the silver appeared slightly darkened in color. Comments. — In the remarkable diversity of its reactions, methyl hydrosulphide probably excels every other known gas. The stability of many of its metallic compounds is often nearly as great as that of the corresponding sulphides. The reagents employed include many substances of high oxidizing power. It was not possible, however, to detect in any case a trace of sulphuric acid. Under the influence of oxidizing agents, the tendency of the mercaptans is to produce the sulphonic acids. CHaSH + 03 = CH3S03H. Hence the failure to form sulphuric acid. The following experiment illustrates the remarkable stability of methyl hydro- sulphide: * The gas was passed in slow stream through a glass combustion- tube containing a fused mixture of sodium carbonate and potas- sium bichromate, but no sodium sulphate was produced. More- over, the gas escaping from the tube possessed the characteristic odor of the mercaptan. The same experiment was tried with a mixture of sodium carbonate and sodium nitrate with a similar result. On account of its numerous^ reactions towards the various metallic salts, a separation of methyl hydrosulphide and sulphur- Oxidation and Chemical Properties of Gases. 353 etted hydrogen is a difficult matter. As it acts slowly and incom- pletely upon yellow mercuric oxide, this substance may be used to absorb sulphuretted hydrogen. Although methyl hydrosul- phide attacks and combines with the mercuric oxide, sulphuretted hydrogen gradually expels it, the yellow color gradually changing to black, owing to the formation of sulphide of mercury. The yellow copper compound changes into copper sulphide. The same is true of the yellow compounds of lead and silver and the white cadmium compound. This change into sulphide is in every case promoted by exposure to air, especially in presence of ammonia. In an aqueous solution of methyl hydrosulphide con- tafning neither acids nor alkalies, the various compounds are more stable. The mercaptides are easily produced in many cases by the action of a solution of the mercaptan in water upon the oxides, hydroxides or carbonates of the metals, and when so formed they are easily preserved unchanged. It is of importance to note that sulphuretted hydrogen expels methyl hydrosulphide from many of its metallic compounds. The reactions of the latter towards gold chloride and arsenious chloride are especially remarkable. In the former case the pro- duction of a precipitate, gradually changing from yellow to white and finally disappearing, distinguishes this gas from sulphuretted hydrogen. The formation of an oily liquid insoluble in water, in the case of arsenious chloride, also serves to distinguish between the two gases. Met/iyl Sulphide. This compound may be prepared by the action of methyl iodide (or, preferably, methyl chloride) upon potassium sulphide in alcoholic solution. Gaseous methyl chloride may be led into an alcoholic solution of potassium sulphide contained in a flask heated over a water-bath and connected with a reversed condenser. The vapor of methyl sulphide thus formed may be freed from sulphuretted hydrogen by oxide of mercury or by passage through warm soda-solution. In the process of Klason, already described, methyl sulphide is produced simultaneously with the hydrosulphide. The process yields, in fact, a larger proportion of the former than of the latter. It may be readily separated, as already stated, by means of soda- 354 Phillips. solution, which absorbs and combines with the mercaptan but exerts no action upon the sulphide. On warming the soda-solu- tion, therefore, the sulphide, condensed and floating upon its surface, may be expelled in vapor-form. In the following experi- ments methyl sulphide was used, prepared by the action of methyl chloride upon potassium sulphide and also by the method of Klason. Methyl sulphide is a colorless liquid, boiling at 37° C. The compound formed by mercuric chloride with methyl sul- phide is (CHs)2SHgCl2. The yellow precipitate produced in platinum-chloride solution is PtS Compt. Rend. 113, 71. «Ber, d. chem. Ges. 34:, 3940 (1891). 358 Phillips. nitride.' The combination of nitrogen and magnesium could only prove of interest as a gas-reaction in case the formation of the compound is not interfered with by the presence of such gases as Are not readily removable from a mixture. Sulphur and oxygen compounds would naturally be decomposed by magnesium. In experiments with natural gas, as supplied to Allegheny in October, 1892, it was found that magnesium heated to redness in a stream of the gas for one-half hour yielded a strong odor and the usual reactions characteristic of ammonia on moistening. The compound produced exhaled ammonia on exposure to air. Nitro- gen also unites directly with lithium and potassium. Ouvrard^ obtained a nitride of lithium containing 50.28 per cent, of nitrogen. Oxygen — Reactions. The presence of oxygen in very small quantities in a gas- mixture is easily recognized by the change of color produced in a precipitate of ferrous ferrocyanide or manganous hydroxide, or in a solution of pyrogallol in soda, or indigo solution previously bleached by zinc dust. All of these substances absorb oxygen and at the same time undergo a change of color. The most sen- sitive of these is the mixture of pyrogallol and alkali. A very sensitive reagent for free oxygen is found in precipitated mangan- ous hydroxide in water, which, by reason of very complete oxygen absorption, changes into manganic hydroxide, Mn^Os, its color changing at the same time from white to brown. The following process is a modification of Winkler's method^ for the determina- tion of dissolved oxygen in water. Two bottles of about two ounces capacity are connected as shown in the sketch. The gas stream enters by A and bubbles through soda solution and manganous chloride consecutively. After complete expulsion of air by the gas current, the tube B is pushed down so that some of the soda solution is forced over into the manganous-chloride solution, causing a precipitation of man- ganous hydroxide. This precipitate remains white in the absence of oxygen. If oxygen be present, it gradually darkens in changing to manganic hydroxide. On adding now a little iodide of potassium solution and then sulphuric acid, by the tap-funnel, •Aslanoglou: Chem. News 63, 99 (1890 V 'Compt. Rend. 113, 120 (1891); Ber. d. chem. Ges. 25, Ref. 104. 5 Ber. d. chem. Ges. 33, 1764 (1889) ; Ztschr. ang. Chem. 1891, p. 105. Co w 1 T =^ 36o Phillips. the oxideof manganese redissolves, liberating iodine, recognizable by its color even when the most minute traces only of oxygen are present. The same apparatus can be used for pyrogallol and soda. The method above described is very satisfactory in testing for oxygen in presence of paraffins, olefines, acetylene, allylene, carbon monoxide, or carbon-disulphide vapor. Sulphuretted hydrogen and carbon oxysulphide must be absorbed by ammo- niacal cadmium-chloride solution, or other suitable reztgent, before the test can be applied. If an ammoniacal cadmium- chloride solution is thus used, vapors of ammonia must be absorbed by dilute sulphuric acid before the reaction is tried. No difficulty occurs in applying the same reaction in testing a limited volume of gas. Instead of conducting the gas in a stream through the solutions, a eudiometer may be used. The reaction is, however, far less satisfactory, as the solutions are liable to hold atmospheric oxygen dissolved. By means of a standard hyposulphite-solution the free iodine may be estimated ; and, as 254 parts of iodine correspond to 16 parts of oxygen, the latter element is easily determined. Von der Pfordten' proposes the use of chromous chloride in presence of sodium acetate for the quantitative determination (by absorption) of oxygen in a gas-mixture. According to this author, the change from colorless to greenish blue renders the solution a suitable reagent for the recognition of oxygen. Very minute quantities of oxygen can be detected. The preparation of the reagent (reduction from chromic chloride by zinc in presence of hydrochloric acid) is effected in a Woulfe bottle, through which the gas is already passing, and the air is thus removed previous to the test. Fuming sulphuric acid is said to dissolve oxygen.^ Nothing need be said here concerning the detection of oxygen when occurring in large quantities in a gas-mixture. The general study of gas reactions has not yet attracted the attention it deserves. The majority of articles bearing upon the subject have referred only incidentally to reactions by which a particular gas, or group of gases, may be recognized. Every effort has been made to cite references to all published statements ' Ztschr. anal. Chem. 26, 74 (1887). 2 Lean and Bone : J. Chem. Soc. 61, 879 (1892). Oxidatio7i and Chemical Properties of Gases. 361 concerning reactions which I have detailed. It is probable, how- ever, that many have been overlooked. In conclusion, I have to express my thanks to Messrs. R. B. Carnahan, Gustav Miller and Henry Phillips, for assistance in the work, and especially to Mr. Henry T. Weed. HI. — Substitution-products of the Action of Chlorine UPON Methane. Natural gas was used in the following experiment. The gas was collected in June, 1888, from main conveying gas directly from Murraysville to Pittsburgh. A steel cylinder provided with thoroughly tested valves was filled from the gas main under a pressure of eighty pounds. An analysis of the gas showed it to have the following composition : Methane 95-40 Carbon dioxide 0.20 Carbon monoxide o. Ethylene O. Hydrogen o. Oxygen trace. Nitrogen 4.40 The following is an outline of the process of treatment : Chlorine was generated in a large flask and washed in B before drying in Q. Natural gas was freed from carbon dioxide by caustic soda, and then dried by sulphuric acid. Q and C served at the same time to regulate the flow. ^ is a glass combustion- tube which was filled with bone-black previously washed by hydrochloric acid. It was sought to produce the reaction : CH.+ 2C1=CH3C1 + HC1. The furnace D was kept at the lowest possible temperature necessary to cause the color of the chlorine to disappear. After passing this tube, the gases were conducted through several bottles of ferrous-chloride solution to remove any excess of chlor- ine, as well as hydrochloric acid. A reaction occurred at once in E : the chlorine disappeared. Too high a temperature caused a pale flame to appear in the combustion-tube, which invariably led Vol. XVI.-27. 362 Phillips. to a deposit of carbon. It was found necessary to maintain a very low temperature in the combustion-tube. Slight condensation of a clear liquid occurred in O. Whether an excess of chlorine, or an insufficient quantity, or the theoretical quantity for the above reaction was used, there was formed continuously the tetrachloride of carbon, which collected in oily drops in i^and G. From experiments in using different proportions of chlorine and methane, employing higher and lower temperatures, and when sand or asbestos was substituted for the bone-black, and in using an empty combustion-tube, I am led to the conclusion that the tendency is always to form methyl chloride and carbon tetrachlor- ide ; but that the intermediate products, dichlormethane CHiCls, and chloroform CHCls, are formed only in relatively small quantity. The manufacture of chloroform from natural gas, so far as these experiments indicate, is likely to prove a matter of difficulty. The gas escaping from O has the odor of methyl chloride from methyl alcohol, is readily soluble in water and in alcohol, and burns with a green flame. The gas, after leaving P, passed into a solution of potassium hydrosulphide in R and then on into a solu- tion of mercuric chloride in X. An immediate and copious pre- cipitation occurred in X. Methyl chloride from methyl alcohol, as is well known, is char- acterized by the property of forming a solid crystalline hydrate when conducted into ice-water. The gas, prepared by the method above described, was passed through the bottle P containing broken ice while the ice was slowly melting, but no trace of a crystalline hydrate appeared. It was not attempted to analyze the gas, for the reason that an analysis of a mixture of methyl chloride with some unaltered methane and traces of the higher chlorides would lead to very uncertain results. The odor, the solubility in alcohol, the green color of the flame and the reaction with potassium hydrosulphide, all tend to show that it was methyl chloride. The failure to pro- duce the crystalline hydrate with ice-water I cannot explain. It has long been considered a settled fact that only one methyl chloride is possible, Berthelot having shown' that when methyl chloride from methane and chlorine is treated with potash, sapon- ification results with production of methyl alcohol, just as in the 1 Compt. Rend. 45, 916. Oxidation and Chemical Properties of Gases. 363 case of methyl chloride from wood-spirit and hydrochloric acid. Von Baeyer' states that methyl chloride prepared from methyl alcohol and hydrochloric acid is different from the methyl chloride obtained by the action of chlorine on methane in the fact that the chloride from the latter source fails to form a crystalline hydrate when led into ice-water, and that there are, therefore, two com- pounds isomeric, but not identical, having the formula CHsCl. Roscoe and Schorlemmer^ explain the failure to form a crystal- line hydrate by the methyl chloride from methane, on the ground that other chlorinated substitution-products occur with the methyl chloride. My experiments lead me to think that this does not satisfactorily explain the difference. Dichlormethane and chloro- form do not occur except in traces in the gas which was produced, while carbon tetrachloride was easily condensed in /^and G, as it boils at 78°. It cannot therefore contaminate the methyl chloride. The methyl chloride formed in the apparatus above described was caused to pass through a second combustion-tube heated in a furnace, and through a side tube a stream of chlorine was passed directly into this second combustion-tube. The methyl chloride supposed to have been formed by the action in the first furnace received, therefore, an additional quantity of chlorine before pass- ing through the combustion-tube in the second furnace. It seemed possible that in such a case the formation of higher chlor- inated derivatives might thus be better controlled. The equation CH4-f-2Cl = CH3Cl-f- HCl probably represents the reaction occurring in the first combustion-tube. The gases were then passed through water to remove hydrochloric acid. They were then dried by sulphuric acid and received the additional volume of chlorine, as above mentioned, before entering the second heated combustion-tube. This reaction might then occur : CH3Cl-f2Cl = CH.Ch + HCl. In the second tube the results were hardly different from those originally obtained. The methane tends constantly to produce methyl chloride or carbon tetrachloride, and there is little or no probability of obtaining intermediate products except in relatively very small proportions. > Ann. Chem. (Liebig) 107, 269 (1858); see also Watt's Die. (ist Ed.), Vol. Ill, p. 987. a Vol. Ill, Pt. I, p. 203. 364 Phillips. IV.— Preparation of Halogen Compounds of Alkyls AND Olefines. The alkyl iodides serve as the most convenient source for the preparation of the paraffins by the Gladstone and Tribe reaction, to which reference has already been frequently made, e. g. : CH3I + Zn + H2O = Zn0. With the expectation of obtaining a similar compound, the chloride of orthosulphobenzoic acid was treated with potassium hydrosulphide. In this case, however, potassium hydrosulphide acts as a reducing agent: sulphur is thrown down, sulphuretted hydrogen evolved, and a compound is formed whose analysis cor- CH" responds to the formula for sulphonphthalide, C6H4<^gQ^'>'0. The chloride which was used for this purpose was prepared by treating the acid ammonium salt of orthosulphobenzoic acid with phosphorus pentachloride.^ 10 grams of the oil thus obtained were added, a drop at a time, to a solution of potassium hydrosulphide which represents 10 grams of caustic potash in 150 cc. of water. At ordinary temperatures the substances do not react, but on warming sulphur begins to separate, the oil passes into solution, and at the same time there is an evolution of sulphuretted hydrogen. As the action proceeds heat is produced sufficient to boil the liquid and thus cause undesirable results. The vessel containing the materials should therefore be placed in cold water after the reaction has started, and should be shaken continually in order to prevent inclusion of the oil by the separated sulphur. At the end the materials are warmed on a water-bath. The liquid is now filtered through an asbestos filter on the top of which has been placed some flowers of sulphur. This process removes nearly all of the sulphur. The yellow filtrate is diluted with 1 Ber. d. chem. Ges. 17, 1175. 2 This Journal 11,332- A Reduction- Prodiut of Orthosulphobenzoic Chloride. 367 about 2 liters of water, heated to boiling and acidified with hydro- chloric acid. A copious precipitate is produced which consists of sulphur and the compound sought ; the greater part of the latter, however, remains in solution and separates as a poorly crystalline mass when the filtered solution cools. A further yield may be obtained if the material on the filter is warmed with potassium hydrosulphide and the process repeated. The com- pound which is precipitated from hot water will not dissolve on heating with the sanie solution from which it was precipitated. If, however, sulphuretted hydrogen is passed in, the compound dissolves and is again precipitated on cooling. The product was crystallized from alcohol and analyzed, with the following results: I. 0.2234 grani gave 0.4057 gram COsand 0.0709 gram H2O. II. 0.2419 gram gave 0.4382 gram CO2 and 0.0780 gram H2O. III. 0.2193 gram gave 0.3009 gram BaSO*. IV. 0.1994 gram gave 0.2740 gram BaSO^. Theoretical for Found. CHeSOj. I. II. III. IV. C 49.38 49.52 49.40 H 3-54 3-52 3-59 S 18.85 ... ... 18.85 18.88 The formation of a compound of this composition under the conditions stated would suggest the treatment of the chloride with simple reducing agents. The action of nascent hydrogen was studied and the compound obtained was found to be identical with that formed by the use of potassium hydrosulphide. 5 grams of the chloride were dissolved in ether and placed in an Erlenmeyer flask which had been filled nearly to the top with granulated zinc. The flask is placed in ice-cold water and a few drops of dilute hydrochloric acid added from time to time during 12-16 hours. The products of the reaction are allowed to stand in a cool place for 24 hours. As the ether evaporates the product deposits on the zinc. The acid solution is poured oflf and dilute alkali placed in the vessel. The organic compound dissolves and, after filtering from the zinc, is precipitated by acid. Dissolv- ing in an alkali, filtering and precipitating with an acid, is repeated several times to remove traces of zinc chloride, and the compound is finally crystallized from alcohol. I. 0.1872 gram gave 0.3401 gram CO2 and 0.0595 gram HsO. II. 0.2003 gram gave 0.3619 gram CO« and 0.0638 gram HsO. 368 Jones. III. 0.3013 gram gave 0.4150 gram RaS04. IV. 0.2885 gram gave 0.3960 gram BaSO*. Theoretical for C,H,S03. I. Found. n. c 49-38 49-54 49.22 H 3-54 3-53 3-54 S 18.85 18.92 18.86 The two substances prepared by these two methods of reduction are identical in all their chemical and physical properties. The compound melts and decomposes at 287^-289°. It is insoluble in water, soluble in alcohol and in ether. The crystals obtained from alcohol are difficult to dissolve again ; but if dissolved in ammonia or alkali and precipitated by an acid, the material can easily be dissolved in alcohol. It reduces permanganate solution, and is oxidized to orthosulphobenzoic acid by the action of boiling concentrated nitric acid. The substance obtained by dissolving the crystalline compound in alkali and precipitating with an acid can be oxidized almost instantaneously by warming with dilute nitric acid. When this precipitated material has been dried it requires concentrated acid and continued heating for its oxidation. When warmed with resorcin and concentrated sul- phuric acid a compound is formed whose alkaline solution shows a fine eosin fluorescence. This reaction suggests the thio-anhy- dride, but as the body which is formed in the reaction with zinc and hydrochloric acid also produces a fluorescent compound, and as the fluorescent compound cannot be produced without the assistance of sulphuric acid, the conclusion follows that oxidation takes place before a phthalein is formed. All attempts to unite the compound with ammonia or alkaline sulphites failed to produce crystalline bodies. An attempt was made to prepare an ammonium salt by allowing a slightly alka- line solution in ammonia slowly to evaporate ; ammonia gradually escaped and the original compound was deposited. By boiling a solution in ammonia with barium carbonate, ammonia is driven out, but as the filtered solution is slowly evaporated, barium hydroxide crystallizes and the original substance is deposited. It has been shown' that the chloride of orthosulphobenzoic acid made by treating the acid ammonium salt with phosphorus pentachloride consists of two substances, one of which separates 1 Ber. d. chem. Ges. 26, 2634. A Re dtidion- Product of Orthosidphobenzoic Chloride. 369 in fine crystalline form when the mixture is allowed to stand at the ordinary temperature; and that the difference between these two isomeric chlorides is probably expressed by formulae similar to those which have in turn been offered for phthalyl chloride. It was at first supposed that the compound which is the subject of this paper is a mixture of two isomeric substances resulting from the two isomeric chlorides. This explanation of several apparent anomalies noted above is insufficient, for it was found that the crystalline chloride which separates from the oil on standing, produces with potassium hydrosulphide a compound which is soluble in water and dilute acids. Moreover, there is in this case no separation of sulphur. This reaction is now under investiga- tion. It would seem that a compound whose composition is expressed by the formula CvHeSOs, formed by the reduction of one of the chlorides of orthosulphobenzoic acid, must be represented struc- turally by one of the two formulae : I. C.H<™:>0 II. C.H Recueil trav. chim. 7, 343 ; Ber. d. chem. Ges. 23, Ref. 296. -Ann. Chem. (Liebig) 144, 114; 146, 107. 3 Ann. chim. phys. [4] 17, 103. Addition- Products of the Aromatic Isocyanides. 373 was shown by Net',' who studied exhaustively the phenyl and ^-tolyl isocyanides. Nef was the first to grasp the fact that the most characteristic property of the isonitriles is their power of adding to the carbon atom of the isocyanogen radicle, two monovalent atoms or groups, thus proving the presence of bivalent carbon. The work which follows was carried out under the direction of Dr. Nef and consists chiefly of a general study of the addition- products of /-tolyl isocyanide and a study of some of the reactions of one of the addition-products of phenyl isocyanide. As is to be expected, the reactivity of/>-tolyl isocyanide is about the same as that of the other aromatic isonitriles and its addition- products are similar. Besides the addition of chlorine, which was accomplished by Nef," the author has effected the following addi- tions ; 1. Of sulphur; form ing/>-tolyl mustard-oil: OH7N : C -h S = CvHiN : C : S. 2. Of sulphuretted hydrogen, and mercaptans; giving thio- form-/)-toluide and its alkyl ethers : H SH' H OHvN : C 4- H2S = C7H7N : C< i OHvN : C + RSH — CvH^N : C-toluidine : C,H,N:C + 4H = OH,NHCH3. Of amines, giving amidines: OH,N : C-f RNH2=OH,N : C<5;Jj^j^. Of organic-acid chlorides, giving imide chlorides: CvHvN : C + R.COCl = OHiN : C<^q j^. Preparation of Paratolyl Isocyanide.* This substance was made according to Hofmann's method for the preparation of the isonitriles. The yield is always poor, the 1 Ann. Chem. (Liebig) 270, 267. "-Ibid.^lQ, 321 SNef : Ann. Chem. (Liebig) 270, 320. 374 Smith. best being obtained as follows: A solution of 210 grams caustic potash in 800 cc. alcohol (95 per cent.) is warmed to about 50* and a solution of 100 grams /-toluidine in 190 grams chloroform is slowly added by means of a dropping-funnel. Reaction at once sets in, accompanied by the separation of potassium chloride and the evolution of heat. It is necessary to cool con- tinually, the temperature being best kept at a little over 50°. The reaction is much less violent toward the end and is never com- plete. After standing a short time, the alcohol and unchanged chloroform are distilled off on the water-bath, and the residue, after the addition of a quantity of water sufficient to dissolve the potassium chloride, is distilled with steam. A mixture of/»-tolui- dine and /)-tolyl isocyanide goes over as an oil of slight yellow color. This is taken up with ether and washed with 120 grams hydrochloric acid (sp, gr. 1.20), previously divided into two portions and largely diluted. It is best to do this washing as rapidly as possible, for the isocyanides react very easily with the amine hydrochlorides.' To this end it is well to dilute the hydrochloric acid as much as the size of the separatory funnel used will allow, and to warm it to about 30° ; for otherwise the di-/-tolylformamidine, which is always formed, gives trouble- some emulsions, as it is only slightly soluble in cold water. As soon as the amidine hydrochloride is all removed by washing with water the ethereal solution is washed once with dilute alkali and dried over caustic potash. After distilling off the ether, the isonitrile is distilled at reduced pressure. Paratolyl isocyanide boils at 99°* at 32 mm. pressure (" 99" at 36 mm." [Nef]), and has a specific gravity of 0.96 at 24*^ (Westphal). On cooling it solidifies to a crystalline solid which melts at 21°. It» odor does not differ much from that of the other aromatic isonitriles, being less disagreeable than that of phenyl isocyanide. Its taste is extremely bitter, as may be noticed by inhaling a small quantity of its vapor at ordinary temperatures. No unpleasant physiological effects were experi- enced by the author in working with this substance. When first distilled it is colorless, but immediately begins to take on a green color and finally becomes brown. It is best kept in the dark at ' See p. 380. -The temperatures given in this paper were measured by a thermometer which was care- fully compared with a standard set of Gerhardt's short thermometers. At no point between o and 200° was there a variation of one degree. Addition- Products of the Aromatic hocyanides. 375 a temperature below its melting-point. Under these conditions it may be kept for several weeks with only slight change. On heating polymerization takes place, interfering with such reactions of this compound as require a high temperature for their accom- plishment. Action of Sulphur. — 5 grams /-tolyl isocyanide and the calcu- lated amount of sulphur dissolved in carbon disulphide were heated in a sealed tube for six hours at 1 20^-140°. On opening the tube no smell of isocyanide was left. After the carbon disulphide had been distilled off, the mustard-oil was separated from the polymerized isonitrile by distilling with steam. The distillate was extracted with ether and the solution dried with chloride of calcium. After distilling off the ether the mustard- oil boiled at 242°-243° and melted at 26°. Hofmann' gives the boiling-point 237° and the melting-point 26°. Sulphur was determined by the method of Carius. 0.1538 gram substance gave 0.2409 gram BaSOi. Calculated for CgHiNS. Found. S 21.47 21.55 Paratolyl isocyanide seems to have a great tendency to take up sulphur from the compounds of sulphur with the halogens and form mustard-oils. A solution of bromine in carbon bisulphide when treated with isocyanide gave considerable quantities of mustard-oil. The odor of mustard-oil was noticed on vulcanized rubber which had been exposed for some time to the vapor of isonitrile. Mustard-oil was the only compound that could be isolated from the substances obtained by the reaction of thio- phosgene on isocyanide. Sulphur monochloride and /-tolyl isocyanide react with great violence. Several experiments were carried out, using carbon disulphide as a diluent. Five grams isocyanide and 3 grams sulphur monochloride were used in each experiment. The only products that could be identified were polymerized isonitrile and mustard-oil, the latter being obtained in quantities varying from 2.5 to 4 grams in the several experiments. 5 grams />-tolyl isocyanide and 2.2 grams sulphur dichloride were mixed in carbon-disulphide solution, moisture was carefully excluded, and the mixture was cooled by means of salt and ice. ' Ber. d. chem. Ges. 1, 173. 376 Smith. On boiling off the carbon disulphide and distilling with steam, 3.5 grams of an oil were obtained which, after taking up with ether and drying, boiled from 225°-240°. This was proved to be a mixture of/-tolyl mustard-oil and/>-tolylimido carbonyl chloride by treating with an excess of /toluidine, which gave 4.5 grams sulphocarbtoluide and i.i grams «-tri-/>-tolylguanidine. This reaction may be explained in two ways : The sulphur dichloride may first act simply as a chlorinating agent, giving the imido- carbonyl chloride and sulphur monochloride, which reacts with more isonitrile, giving mustard-oil ; or the sulphur dichloride may add to the isonitrile, giving ;5-CH3.C6H4.N = C-tolyl isocyanide. Paratolyl isocyanide gives the corresponding para derivative. 50 cc. of gS-per cent, alcohol were saturated with sulphuretted hydrogen at 0° ; 2 grams of/>-tolyl isocyanide were added, and the mixture was heated in a sealed tube for eight hours at 100°. On cooling, brown needles separated out, and more were obtained from the mother-liquor by evaporation, giving in all i.i grams crude product. This was recrystallizecJ from ligroiin (boiling-point 70°-8o°), using animal charcoal ; and twice from benzene, giving nearly colorless, flat needles that melted at 175°-! 76°. Dried in vaczio over sulphuric acid and analyzed by the method of Carius : 0.1573 gram substance gave 0.2405 gram BaS04. Calculated for CsHsNS. Found. S 21.19 21.04 Senier* obtained thioform-/>-toluide from form-/)-toluide and phosphorus pentasulphide. He describes it as a yellow substance melting at 173.5°. 1 Sell and Zierold : Ber. d. chem. Ges. 7, 1228 ; Nef : Ann. Chem. (Liebig) 870, 284. 2Ber. d. chem. Ges. 10, 1095. 3 Ann. Chem. (Liebig) 270,313. * Ber. d. chem. Ges. 18, 2295. Additio7i- Products of the Aromatic Isocyanides. 377 Paratolylimido thiojormic ethyl ester, CH3.C6H4.N = ^<^Q„Y{,'— 5 grams/>-tolyl isocyanide and 2.5 grams ethyl mercaptan were heated in a sealed tube for two and one-half hours at 100°. The reaction was not complete at this point, for on opening the tube the odor of each of the substances could be distinguished. The product was heated on a water-bath to drive off the unchanged mercaptan and then distilled. On the second distillation 3 grams of a slightly yellow oil were obtained, which boiled at 25o''-252°. On analysis : 0.2080 gram substance gave 0.1355 gram H2O and 0.5101 gram CO.. 0.2712 gram substance gave 18.6 cc. N at 17.5° and 744.7 mm. 0.1659 gram substance gave 0.2133 gram BaS04 (Carius). Calculated for C,oH,3NS. Found. c 67.04 66.88 H 7.26 7.24 N 7.82 7.81 S 17.87 17.69 The odor of this ester is disagreeable, reminding one of the imido ethers and also of sulphur derivatives. By treatment with aqueous hydrochloric acid it is converted into di-/-tolylform- amidine hydrochloride.' Wallach and Wusten^ have obtained /^-tolylimido-thioacetic ethyl ester (boiling-point, 27i°-273'') and phenylimido-thioformic methyl ester (boiling-point, 230°-240°). These with hydrochloric acid give corresponding amidines. Paratolylimidoformic ethyl ester, CH3.C6H4.N = Crams) was redistilled, boiling at 23i°-232° at743mm. On analysis: 0.2072 gram substance gave 0.1488 gram H2O and 0.5572 gram CO». 0.2375 gram substance gave 19. i cc. N at 23° and 736.1 mm. Calculated for C,oH,3NO. Found. C 73-62 73.33 H 7-97 7-98 N 8.59 8.82 This imido-ether is a pleasant-smelling oil of a slightly yellow color. On cooling to about 0° it solidifies to a crystalline solid which melts at 8°. By treatment with aqueous hydrochloric acid (dilute or concentrated), it could be converted into di-/'-tolyl- formamidine hydrochloride, which gave the free base melting at 140°. Action of Meihy I Alcohol. — 5 grams/>-tolyl isocyanide and a solu- tion made by dissolving i gram sodium in 10 cc. methyl alcohol were heated in a sealed tube for three hours at 130°. On cooling, quite a quantity of crystals, 2 grams, separated out. After recrystallization from alcohol, they melted at 140° and were iden- tified with di-/>-tolylformamidine obtained by other methods. The filtrate was poured into an excess of water. The oil which sepa- rated out was taken up with ether, and the ethereal solution, after washing several times with water, was dried over calcium chloride. The ether was then evaporated and the residue distilled, giving about a cubic centimeter of a slightly yellow oil which boiled between 215° and 220°. This was undoubtedly />-tolylimido- formic methyl ester (boiling-point, 2i6°-2i8°),' for it resembled the ethyl ester in odor and appearance, and on treatment with dilute hydrochloric acid gave di-^-tolylformamidine. The above experiments make plain the reason for the failure of former attempts to obtain the alcohol addition-products of the isonitriles. The treatment has been too energetic, and the unstable imido-ethers have gone over at the temperature of the reaction to formamidines. Reduction of Paratolyl Isocyanide. — Nascent hydrogen con- verts ^-tolyl isocyanide into monomethyl-/>-toluidine. 8 grams isonitrile were dissolved in 70 cc. amyl alcohol ; the solution was heated to boiling in a flask with reversed condenser attached, and ' Comstock and Clapp : This Journal 13, 527. Addition- Products of the Aromatic Isocyanides. 379 7.5 grams sodium gradually added. As soon as the sodium had all disappeared, the solution was cooled and washed with water. The smell of isocyanide had all disappeared. The solution was then partially dried with lime and distilled. After the amyl alcohol had distilled out, 4 grams of slightly colored oil came over between 180° and 215°, 3,5 grams between 215° and 225°, and 3 grams between 225° and 250°. These fractions were then treated with dilute hydrochloric acid. The first two gave no precipitate, but the third gave 0.4 gram of di-^-tolylformamidine hydrochloride, probably due to the decomposition of /-tolylimidoformic amyl ester, which could very easily be formed during the reaction. The acid solution was heated on the water-bath till all odor of amyl alcohol had disappeared. It was then treated with sodium nitrite, and the nitrosamine extracted with ether. After repeated crystal- lizations from alcohol and ether, it melted at 48°-49°. This nitrosamine was reduced with tin and hydrochloric acid. The base was then set free, distilled with steam, and the distillate extracted with ether. After drying with caustic potash, distilling off the ether, and distilling the residue, 2.5 grams of monomethyl-^-tolui- dine were obtained which boiled at 208°. On treatment with acetic anhydride, methylacet-/>-toluide' was obtained, which, after several crystallizations from ligroin, melted at 79°, Actio7i of Amines. — Weith^ has shown that phenyl isocyanide unites with aniline to give diphenylformamidine. This observa- tion has been confirmed by Nef,^ who also obtained di-■ Ibid. 270, 282, 313, 321. 'Ann. chim. phys. (4] 17, 222 and 239. ■•Ann. Chem. (Liebig) 370, 303. ''Ibid. 370, 286 and 315 et seg. ^ See reaction with chlorides of sulphur, page 375. Addition- Products of the Aromatic Isocyanides. 381 seem to be possible to isolate the addition-products, since they immediately decompose, giving the original chloride and isocya- nide which in this nascent state entirely polymerizes. With the organic-acid chlorides, however, much more tangible results were obtained. OH Mesoxalparatoluide hydrate, cHsGh! N = C^^<^OH- ~ Phosgene reacts very energetically with p-\.o\y\ isocyanide. A smooth reaction was obtained as follows : 9.5 grams ^-tolyl isocy- anide were mixed with a little more than an equal volume of abso- lute ether and cooled in a mixture of salt and snow. 10 cc. of phosgene, previously cooled in a freezing-mixture, were added, and the mixture was allowed to warm slowly to the temperature of the room, moisture being excluded by a chloride-of-calcium tube. After standing for thirty-six hours, it was heated for a short time on the water-bath to drive off the excess of phosgene. The reaction-product was a thick oil, colored dark by polymer- ized isonitrile, but was without a trace of isonitrile odor. It was not attempted to purify this imide chloride directly, but instead it was poured into an excess of cold water. It became solid slowly, without any marked evolution of heat. After pulverizing and treating thoroughly with cold water, the solid was filtered off and dried. From this crude product, mesoxal-^-toluide hydrate was obtained in a pure state by boiling with water, or more rapidly, by digestion with dilute alkalies. It is almost insoluble in cold water, and a litre of boiling water dissolves only about 0.25 gram, which crystallizes out on cooling in the form of fine colorless needles. These, pulverized and dried in vacuo over sulphuric acid, gave the following figures on analysis : 0.1841 gram substance gave 0.0969 gram H2O and 0.4346 gram CO=. 0.1897 gram substance gave 0,4508 gram CO-j (HqO lost). 0.1830 gram substance gave 14.2 cc. N at 9° and 757 mm. Calculated for C^HigN-jOi. Found. C 64.97 64.39 64.80 H 5-73 5-85 N 8.92 9.31 If heated rapidly the hydrate becomes yellow at a little above 100°, and melts to a sticky resinous mass at i20°-i30°. If heated 382 Smith. more slowly, the water goes off gradually and the residue melts at 187°. Mesoxal-/-toluide hydrate has acid properties. It does not color litmus, but it dissolves in sodic hydrate and is precipi- tated unchanged by hydrochloric acid. It dissolves readily in most organic solvents, but on heating these solutions they become yellow, showing the conversion of the toluide hydrate into the toluide. Both the pure hydrate and the crude product mentioned above are readily soluble in alcohol, being converted thereby more or less completely into Mesoxalparatoluide alcoholate, ch'"c6H4 N — C^^^OC-H ' — This was obtained chieHy from the crude product by dissolving in alcohol and precipitating with water and then recrystallizing from dilute alcohol and dilute acetone. In order to be sure that the hydrate was all converted into the alcoholate, it was finally dissolved in absolute alcohol and boiled for an hour in a flask, to which a reversed condenser was attached. Then the alcohol was partly distilled off and, on cooling, the alcoholate crystallized out in colorless microscopic needles. Pulverized and dried over sulphuric acid m z;ac«^, they gave the following analytical figures: 0.1429 gram substance gave 0.0897 gram H2O and 0.3491 gram CO2. 0.2439 gram substance gave 16.6 cc. N at 9° and 752 mm. Found. 66.61 6.97 8.12 In its physical properties the alcoholate resembles the hydrate very closely. By treatment with water a small quantity dissolves, being probably converted into the hydrate. On heating, the alcoholate loses alcohol quantitatively, as the following experiment proves : 0.3965 gram alcoholate heated 3 hours at 95° lost 0.0432 gram, 3 hours at 105°-! 15° " 0.0023 " 8 hours at 95°-i30° " 0.0045 " 3 hours at iio°-i20° " 0.0021 " 3 hours at iio°-i20° " o.ooio " 3 hours at I io°-i20° " 0.0000 " 23 hours at 9o°-i30° " 0.0531 " Calculated for C19H, c 66.66 ' H 643 N 8.19 AdditioJi- Products of the Aromatic Isocyanides. 383 Calculated for loss of C^HjO. Found. 13-45 13-49 The yellow powder left behind was Mesoxalparatoluide, p,^ 'p tt*\j ZZ p^C::=0. — It was ana- lyzed without further treatment : 0.1402 gram substance gave 0.0714 gram H2O and 0.3542 gram CO2. Calculated for CnH.eN^O. Found. c 68.92 68.87 H 5-41 5-66 Its melting-point was found to be 187°. It is soluble in the ordinary organic solvents, giving yellow solutions in such as con- tain no water or alcohol ; but if these are present the solution is colorless and the substance is converted into the hydrate or alco- holate. It reacts with phenylhydrazine, but the reaction-product was not investigated. OH Benzoyl/ormoparaioluide, CH3.C6H4N. = ^^Cpo C H -""Ben- zoyl chloride and p-io\y\ isocyanide react slowly on each other. 8 grams isonitrile and 9.6 grams benzoyl chloride were heated on a water-bath in a flask closed by a chloride-of-calcium tube, for ten hours. The smell of isonitrile had all disappeared, but poly- merization had taken place to a considerable extent. In order to remove the polymerization-product, the tany mixture was boiled out with several volumes of ligroin (boiling-point, yo^-So") and filtered hot. After distilling off the ligroin, water was added to decompose the chlorides, and the mixture was heated on the water-bath till all smell of benzoyl chloride had disappeared. The resultant solid was dissolved in ether and the solution washed with sodium carbonate to remove benzoic acid. After drying with calcium chloride the ether was partially distilled off, and the benzoylformic-/>-toluide was allowed to crystallize out. After recrystallization from alcohol, ether, and finally from ligroin, fiat yellow needles were obtained which melted at iii°-ii3°. On analysis : 0.1350 gram substance gave 0.0718 gram H2O and 0.3741 gram CO2. 0.3738 gram substance gave 20.25 cc. N at 25° and 749 mm. 384 Smith. Calculated for C,sH,3NO,. Found. c 75-31 75-54 H 5-44 5-91 N 5.86 5-99 This substance is very slightly soluble in hot water, but is quite soluble in hot sodic-hydrate solution (i : 30), from which it crys- tallizes unchanged on cooling and may also be precipitated by acids. Pyruvic paratoluide, CH^.C^H^.N = C < ^oCHs- ~ Acetyl chloride reacts with /-tolyl isocyanide much more readily than benzoyl chloride. If a mixture of these two substances be incau- tiously heated, the reaction is so violent that the whole mass chars. The addition was accomplished as follows: 9.5 grams of/-tolyl isocyanide and 7 grams acetyl chloride were mixed in a flask at the ordinary temperature. A reversed condenser closed by a chlor- ide-of-calcium tube was attached and the flask was placed in a cold-water bath, which was then slowly heated to the boiling- point. After the water had boiled four minutes the flask was removed. As soon as the contents of the flask had cooled they were poured into 200 cc. of ice-water. After the mass had become solid it was pulverized and again treated with water, then filtered off and dried. 10 grams of crude product were obtained. This was recrystallized from 2.5 liters boiling water and finally from alcohol, giving 5 grams colorless scales which melted at 108°. On analysis: 0-I793 gram substance gave 0.1040 gram H:0 and 0.4460 gram CO2. 0.2314 gram substance gave 16.7 cc. N at 13.5° and 754 mm. Calculated for C,oH,,NOj. Found. c 67.79 67.83 H 6.21 6.44 N 7.91 8.47 It is easily soluble in boiling alcohol or ether, from which it crystallizes on cooling. It dissolves readily in cold lo-per cent, sodic-hydrate solution; on adding an acid, however, the original compound is not precipitated, but a substance that seems to be a mixture of a hydrate and a polymer; for if it is heated it becomes soft at 6o°-70°, and if warmed with water becomes sticky and a Addition- Products of the Aromatic Isocyanides. 385 part (the hydrate) goes into solution while the polymer remains behind. The same change seems to take place to some extent when pyruvic /-toluide is dissolved in water, for, on allowing the aqueous mother-liquor from which the toluide was originally crystallized, to stand for some days in the cold, quite a quantity of the polymer separated out. If the sodic-hydrate solution of pyruvic ;{»-toluide is made very dilute, the addition of acid pro- duces no precipitate, but, on extracting with ether, a clear ethereal solution is obtained, from which, on standing in the cold, the poly- mer very slowly separates out in the form of colorless microscopic prisms. This polymer melts at I93°-I94°. It is very slightly soluble in water, alcohol, and ether, and somewhat more readily in chloroform. On analysis: 0.1 1 45 gram substance gave 0.0664 gram H2O and 0.2783 gram CO2. 0.1311 gram substance gave 0.0818 gram H2O and 0.3275 gram CO2. 0.2087 gram substance gave 14.6 cc. N at 22° and 745.6 mm. Calculated for tCi„H,,N02™). Found. c 67.79 66.28 68.12 H 6.21 6.44 6.93 N 7.91 7.85 Action of Phenylhydrazine on Pyruvic Paratohcide. — On mixing in the cold an ethereal or alcoholic solution of pyruvic /-toluide with the calculated (i mol.) amount of phenylhydrazine, a sub- stance quickly begins to separate out in the form of microscopic needles. This precipitate is very bulky, soon filling the entire solution when in the proportion of i gram pyruvic /-toluide to 50 cc. solvent. The reaction never seems to be complete and the calculated quantity is never obtained. This reaction-product is pyruvic-/>-toluide phenylhydrazone hydrate, but it loses water so readily that it could not be analyzed. It was prepared for analysis by drying on a clay plate, being pressed between filter- papers, and standing in vacuo over sulphuric acid for one hour. Analyses of three different preparations gave varying figures, all between the hydrazone and the hydrazone hydrate. On heating the freshly precipitated hydrazone hydrate in a capillary tube it partially melts at about 100°, then becomes solid again and finally melts at about 200°. On heating 0.409 gram in an 386 Smith. open dish for five hours at 115° in an air-bath it lost 0.0202 gram and did not change in weight by the further heating. If it had been pure hydrazone hydrate it should have lost 0.0259 grani. CH..C6H7.N = C-OH Pyr2ivic-paratoliiide phenylhydrazone , \ CcH..NH-N = C-CH3 — The product obtained by heating the hydrate was crystallized once from alcohol. It then melted at 204°. On analysis : 0.167 1 gram substance gave 23.5 cc. N at 16° and 746.3 mm. 0.1335 gram substance gave 0.0777 gram H2O and 0.3497 gram CO2. Calculated for C,7H,7N30. Found. C 71.91 7144 H 6.37 6.47 N ■ 15.72 16.15 This hydrazone can be obtained more simply by heating the hydrazone hydrate in ethereal or alcoholic solution. The proof that the above bodies are derivatives respectively of mesoxalic, benzoylformic, and pyruvic acids is unnecessary, as it has already been given by Nef ' in the case of corresponding anilides and orthotoluides. Action of Formic Acid. — Paratolyl isocyanide reacts with formic acid in the same way as do the other isocyanides that have been carefully studied.' 4 grams /-tolyl isocyanide were added to the calculated amount (two molecules) of crystallized formic acid at 0°. A lively evolution of gas took place at once. The residue was distilled, giving 3 grams of an oily substance which on crystallization from ether and ligroin gave colorless crystals melting at 54° and identical with formo-/>-toluide as described by Tobias.' Molecular Rear rang eme7it. — The molecular rearrangement of the isocyanides by heat was first noticed by Weith* in the case of phenyl isocyanide. Hofmann'* obtained the rearrangement of the isocyanides derived from pentamethylamidobenzene and tetra- methylamidobenzene. Nef* has obtained ^-tolyl cyanide from ^-tolyl isocyanide. 5 grams p-\.o\y\ isocyanide were heated in a sealed tube for three hours at 2io°-225°. On opening the tube 'Ann. Chem. (Liebig) 370, 293, 302, 320. -Gautier: Ann. chim. phys. [4] 17,223,241; Nef: Ann. Chem. (Liebig) 370, 278, 310. 3 Ber. d. chem. Ges. 15, 2446. « Ibid. 6, 213. '/(JjV. 17, 1914; 18,1824. 'Ann-Chem. (Liebig) 270, 311. Addition- Products of the Aromatic Isocyanides. 387 the isonitrile smell had entirely disappeared, but much polymer- ization had taken place. The product was taken up with dry ether to get rid of the polymerized isocyanide. After distilling off the ether the residue was distilled, yielding 3 grams oi p-X.o\y\ cyanide boiling at 21 3^-2 14° and melting at 27°. Paterno and Spica' give as the boiling-point 2i7°-2i8°, and as the melting- point, 28.5°. Its identity was further proved by saponification with sulphuric acid, giving />-tolylic acid^ melting at 180°. Paratolyl Isocyanide and Silver Cyanide. — If /-tolyl isocya- nide and silver cyanide are mixed in the cold, a noticeable evolu- tion of heat ensues, and the mass on being thoroughly mixed becomes solid. 2 grams /-tolyl isocyanide and an equal quantity of silver cyanide were mixed in a mortar and rubbed till a dry powder ensued. This was boiled out with 2 liters of water, and on cooling 0.6 gram colorless needles crystallized out. These were filtered off, dried on a clay plate, pressed between filter-papers and analyzed : 0.1537 gram substance gave 0.051 1 gram H2O and 0.2751 gram CO2. 0.1544 gram substance gave 15.5 cc. N at 21° and 750.8 mm. 0.2165 gram substance gave 0.1055 gram AgCl (Carius). 0.1092 gram substance gave 0.0380 gram H2O ando.1916 gram CO2. 0.1288 gram substance gave 0.0629 gram AgCl (Carius). Calculated for ] ^gNC^^^ fc, | ^^AgNC^^ P„^„j c 43-03 50 48 H 2.79 3.39 Ag 43-03 34-84 N in6 11.29 It melts at 118° with decomposition, residue with as much more water more material is obtained, but this does not seem to be as pure and is not so well crystallized. If the clear solution first obtained be boiled for a few moments it becomes turbid from the silver cyanide which separates out. If the crystalline product be washed with alcohol or ether it is decomposed and the residue is much poorer in carbon and hydrogen, as was proved by analysis. The substance is decom- posed by heating with potassium-cyanide solution, drops of > Ber. d. chem. Ges. 8, 441. 2 Fischli : Ber. d. chem. Ges. 18, 615. 48.81 48.96 3-69 3.87 36.67 36.76 11-34 ... On boiling out the 388 Smith. isonitrile separating out. It is also decomposed by hydrochloric acid, giving silver cyanide. The product and its solutions always smell of isonitrile. The above experiments agree very well with the supposition that a simple double salt is first formed and that this is then decomposed by the water from which it was crystallized. This is in accordance with the results of Meyer,' who obtained a double salt AgNC.C'iHsNC, which was decomposed by alcohol, ether, potassium cyanide, dilute acids, and even by standing in the air. Hofmann" states that phenyl isocyanide easily unites with other cyanides, the compound with silver cyanide crystallizing especially well. In the case of/>-tolyl isocyanide I have been unable to obtain evidence of reaction with any of the metallic cyanides at my disposal, except with silver cyanide. As was shown by the experiments of Nef^ the isocyanides easily add two atoms of chlorine, forming alkylated imidophos- genes. These products can be more easily obtained, however, by the action of chlorine on the mustard-oils. From the consti- tutional resemblance of these compounds to phosgene I was led to investigate the action of sodium alcoholates upon one of them, and found that the analogy of the reactions was very striking. When phenylimidophosgene is treated with sodium alcoholates under the proper precautions, it gives phenylimidochlorformic esters, only one atom of chlorine reacting: C6H5.N : C<^} + NaOR = CsHs.N : C Ber. d. cheni. Gcs. 19, 864. 392 Smith. Calculated for CsHsNOCI. Found. c 56.64 56.56 H 4.72 5-02 N 826 8.1 1 CI 20.94 21.17 Phenylimidocarbonic dimethyl ester, C6H5.N=:C Hofmann : Ber. d. chem. Ges. 4, 249. 2 Merz and Weith : Zeit. f. Chem. 1868, 513 ; Hofmann^Ber. d. chem. Ges. 8, 453. 394 Review. classes are generally unsatisfactory. Few students acquire, in the time usually devoted to this study in schools and colleges, a sufficiently good knowledge of the subject to enable them to analyze a mineral or an industrial product of some complexity without the book before them which gives the scheme of analysis to which they have been accustomed, and to which they are slavishly attached. Either this condition of affairs is inherent in the subject itself, or else the usual method of teaching fails to give the student a comprehensive view of qualitative analysis which will enable him to think and work independently of books, and to adapt his methods and procedures to each particular case which may arise. The large number of manuals on qualitative analysis which have been published express, apparently, the dissatisfaction which each teacher has felt with those already existing, and his desire to teach the subject in his own way. Dr. Noyes' book is the outgrowth of his teaching of qualitative analysis in the Massachusetts Institute of Technology, and embodies the systematic class-room instruction in the subject which accompanies the laboratory course. The idea of the book may be thus briefly described : Under each group of metals there is given the " procedure," expressed both in text and in tabular form, giving the necessary directions for adding the different group reagents and the purpose of this addition. Then follows — and this is the most valuable feature of the work — a series of " notes" which state concisely and clearly the necessary conditions for the successful separation and identification of each metal, the complications which may arise from a faulty procedure (such as adding too much or too little of a reagent, failure to get rid of an excess of sulphuretted hydrogen or nitric acid and the like), the characteristic appearances caused by the presence of different metals — in short, all that is necessary to know and observe in conducting intelligently an ordinary qualitative analysis. The same idea is carried out in the detection of the acids, and in this especially difficult part of qualitative analysis the author has been very successful in giving clear and precise directions which make their identification easy. In the short limit of eighty pages Dr. Noyes has condensed the greater part of the material found in large treatises, such as Fresenius' Qualitative Analysis, and by his admirable arrange- ment of this material has made a book which can be put into the hands of beginners, and, as experience has shown, with the hap- piest results. Thomas M. Drown. Vol. XVI. [Junf, 1894.] No. 6. AMERICAN CHEMICAL JOURNAL, THE MENTHOL GROUP. By Leo C. Urban and Edward Kremkrs. Two years ago, in a preliminary note to this Journal,' the dis- covery of a nitrosyl-chloride addition-product of menthene was announced. Since then we have published further work in con- nection with this subject, in the Proceedings of the American Pharmaceutical Association." As these Proceedings have not a wide circulation among readers of this Journal, a short review of the work reported upon may serve as an introduction to this article. The compounds which have been described are : 1. Meniheyie^ CioHis. — Prepared by the use of acid potassium sulphate as a dehydrating agent for menthol ; boiling-point, 165° C. under pressure of 744 mm.; specific gravity at 20°, 0.8134; [«],=-f 31.83°. 2. Menihene nitrosochloride,' CxoWn^OCi. — Melting-point, 113°; [a]o-=.-\- 13.76°. The melting-point (128°) of the nitrosochloride which was optically inactive, has since been observed.* v. Baeyer'' has recently stated that the hydrocarbon obtained on dehydrating the tertiary menthol yielded a nitrosochloride melting at 146° "wie J This Journal 14, 23a. 3 Proc. Am. Pharm. Ass. 1892, 273; 1893, 185. ^liid. 1892, 273. Ber. d. chem. Ges. J85, 143. sproc. Amer. Pharm. Ass. 1892, p. 273. ' Ibid. 1893, p. 185. 398 Urba7i a7id Kreniers. colorless needles in the upper part of the tube. A more rapid sep- aration of the nitrosomenthene from the dark-colored by-products resulting in small quantity upon decomposition of the nitrosochlor- ide in this manner, is effected by distilling the product with water- vapor. The nitrosomenthene passes over slowly, solidifying in the receiver to a white mass, which, when spread on a porous plate and dried, melts at 67° without further purification. This method of purification has also been employed to separate it from the thick viscid mass remaining from the mother-liquor obtained upon recrystallization of the nitrosomenthene prepared by heating the nitrosochloride with alcoholic potassa.' Reduction of 7iitroso77ienihene. — Nitrosomenthene, melting at 66°-67°, was reduced by means of acetic acid and zinc-dust, and the resulting amido compound separated in the form of nitrate, as indicated in the previous work.' This base had been looked upon as a saturated compound, since from the method of reduction employed such a compound might be expected. Upon treatment with nitrous acid we had therefore anticipated the formation of a menthol C10H29O. Alcohol, CioHisO. — 20 grams of the nitrate were dissolved in acetic acid and 8 grams of sodium nitrite gradually added. An oily liquid separated slowly. When the reaction was completed the solution was made slightly alkaline and distilled with water- vapor. The oily liquid, which had a peppermint odor, was sepa- rated and exposed to cold. After some time small white crystals separated which melted at 89°-90° and were soluble in acidulated water. As this showed the presence still of basic matter, the oil was shaken with an aqueous solution of oxalic acid and again distilled with water-vapor. On attempting to fractionate the oil after drying, nitrous fumes were evolved. The heat was immedi- ately removed and the substance heated for some time with a solu- tion of potash to saponify any nitrous acid ester, and again driven over with water-vapor. The resulting product, separated and dried, yielded a main fraction passing over between 2io°-2i5'' ; a little passed over below 210*^, and some above 215°. Fraction 2io°-2i5° was analyzed with the following results: I. 0.1298 gram gave 0.3705 gram C02 = o.ioi045 gram C, and 0.1349 gram 1^20=0.014988 gram H. ' Proc. Amer. Pharm. Ass. 1892, p. 273. - Ibid. 1893, p. 185. The Menthol Group. 399 II. 0.1779 gram gave 0.5078 gram €02 = 0.13849 gram C, and 0.1832 gram H2O =0.020355 gram H. Calculated for C,oH,„0. C.oHjsO. c 76.92 77.91 H 12.82 11.68 77.84 77.84 11.54 11.44 Although the percentage of hydrogen is found rather low, the figures, as will be seen, correspond very well with those calculated for a compound CioHisO. An ethereal solution of the substance readily absorbed bromine, indicating that it is an unsaturated compound. When treated with chromic-acid mixture it formed a dark-colored chromic compound which liquefied upon raising the temperature, A secondary alcohol is thereby indicated. If the alcohol is an unsaturated compound, then theamido compound from which it is obtained must also be unsaturated ; the formula C10H17NH2 may therefore be assigned to it. All of these experi- ments are to be repeated, and a special study of the amido com- pound with regard to its saturation will be made. Inactive menthojte,^ CioHisO. — In the preparation of the amido compound from nitrosomenthene there separated as a by-product an oily liquid having the odor of menthone. This was separated from the solution of zinc acetate and amidoacetate by distillation with water-vapor. The oily liquid was separated, dried and fractionated ; fractions boiling between 204° and 206° and between 206° and 208° were obtained. Most of it passed over between 204° and 206°, and a small quantity remaining above 208°. Upon combustion fraction 204^-206° yielded the following results : I. 0.2064 gram gave 0.5883 gram €02 = 0.160445 gram C, and 0.2198 gram H2O = 0.024422 gram H. II. 0.1690 gram gave 0.4781 gram 00-2=0.13039 gram C, and 0.1821 gram H2O =0.020233 gram H. III. 0.1727 gram gave 0.4925 gram 002 = 0.134318 gram C, and 0.1850 gram H2O =0.020555 gram H. Found. Calculated for C,oH,sO. I. 11. III. C 77-91 77.73 77.15 77.77 H 11.68 11.83 ii-97 11-90 The figures found leave no doubt as to its composition. Its • A notice of the isolation of this compound was given in a paper read before the Wisconsin Academy of Sciences, Arts and Letters, December 27, 1893. 400 Urban and Kremcrs. boiling-point agrees very well with that of menthone. Its specific gravity at i8° C. was found to be 0.9071, It is optically inactive. Likelaevogyrate menthone it is readily converted into the oxime. It was heated for ten minutes in alcoholic solution with hydroxyl- amine hydrochloride and an excess of sodium bicarbonate in the proportions given by Beckmann' for the preparation of laevo- menthoxime. The solution was filtered while hot and set aside to crystallize. Masses of needle-shaped crystals separated, which when spread on a pdrous plate and dried melted at 78°. The recrystallized product melted at 82°. When hydrochloric acid gas is passed into a solution of the oxime in dry ether, a separation of the white crystalline hydro- chloride takes place. An excess of the acid causes the precipitate to redissolve. The ketone-character of the compound CioHisO cannot be doubted. Beckmann and Pleissner' have obtained a menthone yielding an oxime melting at 84°-85°, from pulegone, a body having the composition CioHicO. This menthone they obtained by treating the hydrobromic acid addition-product of pulegone with zinc-dust. It was laevorotatory, whereas pulegone is dextro- rotatory. The menthone regenerated from the oxime and again treated with hydroxylamine, they found to yield an oxime melt- ing several degrees lower, i. e. at 78°-82°. It is therefore very probable that the menthone obtained as a by-product in the reduc- tion of nitrosomenthene is the optically inactive modification of that obtained by Beckmann and Pleissner. MenihoL — Fraction 204*^-206° of the menthone was treated according to Beckmann's method^ for the reduction of ketones. An ethereal solution was heated for some time with metallic sodium, the solution was then poured off from the excess of sodium and shaken with water. It was then removed from the water, the ether allowed to evaporate, the product again dried, dissolved in dry ether and treated a second time with metallic sodium. The resulting product was distilled with water-vapor and the oily liquid after separation exposed to cold. After some time it con- gealed to a crystalline mass. The crystals dried on a porous plate melted at from 24° to 26°. They were exposed to air for some time and then placed over sulphuric acid, but the melting- 'Ann. Chem. (Liebig) 250, 329. -^Ibid. ^62, 25. 3 IHd. 362, 25; Ber. d. chem. Ges. 3'i, 912. The Menthol Group. 401 point could not be brought above 29*'-3i°. The quantity obtained was too small to admit of purification by recrystallization or other means. The crystals had the pungent and cooling taste and odor of menthol, which is quite different from that of menthone. Larger quantities will be prepared and its properties more accu- rately determined. The formation of menthol from menthone according to this process is represented by the following equations : 2C.oH,80 + Na2 = C,oHnONa + CioHi90Na. C.oHnONa + C.oH.sONa -f- 2H2O = CioH.sO -f C,oH,,OH + 2NaOH. C.oH.sO + C.0H19OH + NaQ == 2CioH,90Na. 2C..H.90Na 4- 2H2O = 2C10H19OH + 2NaOH. The formation of menthone, CioHisO, a ketone, as a by-product in the reduction of nitrosomenthene in acid solution, indicated that nitrosomenthene is an oxime ; that a partial hydrolysis of this oxime took place, hydroxylamine being split off, and the ketone CioHioO, which was probably formed, reduced by the nascent hydrogen to the saturated ketone CioHisO. It was therefore expected that the ketone CioHieO could be isolated and nitroso- menthene regenerated on treating it with hydroxylamine. Nitrosomenthene is comparatively stable towards sulphuric and acetic acids, but when warmed with hydrochloric acid, hydroxyl- amine is split off with great readiness. Ketone CioHieO from nitrosomenthe7ie. — 18 grams of nitroso- menthene were dissolved with the aid of a gentle heat in twice its weight of hydrochloric acid (sp. gr. 1.20), and diluted with an equal volume of water. The solution was warmed for a short time on a water-bath and then distilled with water-vapor. A colorless oil was carried over which became yellowish on exposure for a short time. It had a mild peppermint odor, reminding one some- what of pulegone. It was dried with caustic potash and fraction- ated. Nearly all passed over between 210° and 212°, a small amount between 208° and 210°, and some above 212°. The acid liquid remaining in the flask on distillation of the hydrochloric- acid solution immediately reduced Fehling's solution, showing the presence of hydroxylamine. Fraction 2io°-2i2° had a specific gravity of 0.9150 at 20° and is optically inactive. The following results were obtained upon analysis : 402 Urbari and Kremers. I. 0.1676 gram gave 0.4818 gram €02=0.13140 gram C, and 0.1603 gram H2O = 0.017811 gram H. II. 0.2072 gram gave 0.5967 gram C02=r 0.162736 gram C, and 0.1994 gram H20=: 0.0227 gram H. Found. Calculated for CoH.eO. I. II. C 78.94 78.40 78.54 H 10.52 10.62 10.95 Oximido derivative, CioHieNOH. — To convert it into its ox- imido derivative, 3 grams of the ketone were dissolved in two and a half times its weight of alcohol and heated together with 2 grams hydroxylamine hydrochloride and an excess of sodium bicar- bonate, for two hours on a water-bath, in a flask connected with a reflux condenser. The solution was filtered while hot and set aside. After the evaporation of the alcohol a thick viscid mass remained which was purified by distillation with water-vapor. The oxime congealed to a white crystalline mass in the receiver. When spread on a porous plate and dried it melted at 67°. Crys- tallized from methyl alcohol it was obtained in crystals having the characteristic form of nitrosomenthene and melting at 67°. The identity of the oxime with nitrosomenthene cannot therefore be doubted, and the presence of the oximido group in nitroso- menthene must be assumed. Conclusions. It will be seen from the above experiments that the nitrosyl- chloride addition-product of menthene has the structure H,Cs — C — CI H,C3 — C — CI H.Cr'^CHNO ^j. H=Cj"^^C = NOH Hsct^^JcH. ^' H^Cv^CH. CH— CH. CH — CH= the formation of the ketone, CioHuO, of menthone, CioHisO, and of menthol, C10H20O, the corresponding secondary alcohol, leaving no doubt as to the position of the nitroso group or the oximido group, and consequently as to that of chlorine. Accord- ing to Markownikow,' when an unsaturated body combines with a compound YZ at a comparatively low temperature, the more negative part will add itself to the carbon atom connected with 1 Beilstein (3te Aufl.) B.I. I, 94. The Menthol Group. 403 the least number of hydrogen atoms, or that carbon which is already connected with a negative element or group. Nitrosyl chloride would seem, then, to follow this rule. There are two possibilities as to the splitting off of hydrochloric acid from the nitrosochloride in the formation of nitrosomenthene: the chlorine going out with hydrogen so as to form a double bond connecting the side chain, or a double bond in the nucleus, as C=C(CH=)= C — CsHt CH — CH. CH — CH3 The latter, however, must be considered as the correct formula, the tendency being always towards splitting off with hydrogen in the nucleus, v. Baeyer,' in his experiments with menthene and carvomenthene, finds that on splitting off hydriodic acid from the tertiary iodides, menthene and carvomenthene are regenerated, and that it therefore seems impossible in this manner to obtain a tetrahydrocymene in which the double bond is partially or wholly in the side chain. The formation of nitrosomenthene from the nitrosochloride is entirely analogous. Semmler^ has shown that in pulegone the double bond must be assumed as being in con- nection with the side chain; consequentl}'^ if the first formula were assumed for nitrosomenthene, the ketone obtained from it should be identical with pulegone; but this is not the case. The oximido group in the nitrosomenthene has been shown. Whether the rearrangement of the nitroso group with the hydro- gen into the oximido group already takes place in the nitroso- chloride is still an open question. Adopting the nomenclatureofWallach," there are possible four yS-ketones of the formula CioHuO having the double bond in the nucleus, namely, C— CsH: CH — GHt HCf^'^CO CHr<^CO HscI^CHj Hcl^^CH. CH — CH. CH — CH. > Ber. d. chem. Ges. 26, 2269. - Ibid. 25, 3515. 'Ann. Chem. (Liebig) 277, 105. Wallach distinguishes between a- and ^-ketones of the hydrocymenes: the a-ketones are tliose in which the carbonyl group is in the ortho position with the carbon atom connected with the methyl group, while ^-ketones are those in which it is nearest to the isopropyl group. CH — GH H^cr '^CO Hct ^CH. C-CHs 404 Urban and Kremers. CH — CaH, H^cr-^co H.ck^CH ' C— CH, The ketone from nitrosomenthene is the first of these, and so far the only one known. Pulegone, a constituent of the oils of Mentha pulegium^ and oi Hedeonia ptilegiodes^ is also a /^-ketone of the formula CioHieO, but has the double bond connected with the side chain. The alcohol CioHisO from the amido compound CioHnNHs must be the same as that to be expected upon reduction of the ketone CioHieO. Upon dehydration this may be expected to yield a terpene C — C3H7 HCf^^jCH H^ck^CH CH — CH» which would be the first body of this class the structure of which could be definitely stated.' Pharmaceutical Laboratory, Univfrsity of Wisconsin. KETONES FROM PINENE DERIVATIVES. By Leo C. Urban and Edward Kremers. Wallach^ states that nitrosopinene is not affected by acids ; that even when warmed with concentrated sulphuric acid it is again separated unchanged upon diluting with water. This, he says, can only be explained by assuming that in isomeric nitroso ' Ann. Chem. (Liebig) 263, 25. - Proc. Wis. Pharm. Assoc. 1893, 51. 'Since writing this manuscript a paper by v. Baeyer has appeared (Ber. d. chem. Ges. ST, 442) in which it appears that formulse of terpinolene and of dipentene must be considered as C = C(CH3)j C-CH(CH3)2 CH HjcLJcH H.cL^CH C-CH3 C-CH, * Ber. d. chem. Ges. 34, 1546. Ketones from Pinene Derivatives. 405 compounds the group =NOH is contained, whereas in nitroso- pinene the group — N = is present. Nitrosomenthene, which undoubtedly contains the oximido group, is also very stable towards sulphuric acid, but when warmed with hydrochloric acid is easily decomposed with formation of a ketone. When nitrosopinene is treated in like manner with hydrochloric acid it readily decomposes into an oily body, probably a ketone, CmHuO, and hydroxylamine, the presence of which is readily shown in the acid liquid by its reducing power on Fehling's solution. This body is now undergoing investiga- tion. Wallach has observed a ketone CioHisO among the decompo- sition-products of pinylamine hydrochloride when heated, and which he admits' may come from an impurity in the salt. This body may possibly owe its origin to a reaction during the reduc- tion of the nitrosopinene with zinc and acetic acid, similar to that taking place upon reduction of nitrosomenthene causing the for- mation of menthone. The peppermint odor of this ketone, as well as that obtained upon oxidation of the secondary alcohol from pinylamine, CioHmO,* would seem to indicate that they are /^-ketones, and may be con- sidered as an argument in favor of the first formula set up by C — C3H7 HC<:^CH H2C^^'CH CH — CHs HiCs— C — CI Wallach for pinene : The nitrosochloride then ^ould be HC H2C CHNO CH CH — CH C— GH7 CfS^CNOH H^cl^CH CH " , and the nitrosopinene CH; rearrange itself into C — C3H7 >r HCKT^CNOH H^C'^J^CH C — CHs C — C3H7 HCi^^^CNOH H^cl^^CH C — CH. J Ann. Chem. (Liebig) 268, 310. 2 Wallach : Ann. Chem. (Liebig) 277, ijo. which may possibly The peppermint odor of 4o6 Phillips. the ketones obtained by Wallach would appear to speak for some such rearrangement of the bonds in these ketones, since, if the para bond or bonds remained intact, the bodies might be expected to have a camphoraceous odor. Pharmacbuttcal Laboratorv, University of Wisconsin. RESEARCHES UPON THE PHENOMENA OF OXIDA- TION AND CHEMICAL PROPERTIES OF GASES. By Francis C. Phillips. ^Concluded from page 365.] V. — Composition of Natural Gas. The gas used in the following trials was that supplied to Alle- gheny by the Allegheny Heating Company, and is the product of wells scattered over a considerable gas-producing area. It may be said to represent the average composition of an enormous volume of gas. No important differences have been observed during the period from 1886 to 1S92 in the heating or illuminating power of the gas as supplied to the city, although the odor of petroleum [i. e. of higher paraffins) has been occasionally stronger. Tests have also been made of gas from various localities in Pennsylvania, New York and Indiana, and Vancouver, British Columbia, and also at Cleveland, Ohio. In all cases where possible the tests were made at the wells. When this could not be done it was necessary to use samples brought in glass vessels to the laboratory. In such cases the samples were examined for oxygen before being subjected to the tests. As a leak in a sample- vessel invariably causes an interchange of air and gas, so that air enters in proportion as an escape of gas occurs, much dependence is to be placed on the presence or absence of air in a gas-sample as a criterion of its purity. Hydrogeii. — Hydrogen is almost always mentioned in the pub- lished analyses of natural gas. I have made the following Oxidation and Chemical Properties of Gases. 407 chemical tests: The natural gas, as supplied to Allegheny by the Allegheny Heating Company, was caused to flow through a solu- tion of palladium chloride for periods varying from ten days to three months ; 500 cubic feet have been used in a single experi- ment. Similar tests have been repeated at various times between January, 1886, and May, 1892, but in no case was a trace of pre- cipitation observed in the palladium-chloride solution. Natural gas was found likewise to be without action upon solutions of plat- inum chloride and ammoniacal silver nitrate. A stream of natural gas has been passed through pure djy palladium chloride. This extremely delicate test has failed to show the presence of hydro- gen even in traces, although tried repeatedly during the period from January, 1886, until May, 1892. As already stated, the results of my study of gas-reactions show that palladium chloride produces very diflTerent effects according as it is used dry or in solution. Palladium chloride dry is reduced promptly by dry hydrogen when the gas is used in a free state. The same salt in solution is slowly and incompletely reduced by hydrogen, although it is rapidly reduced by olefines and car- bon monoxide. Similar tests with palladium chloride, both dry and in solution, made at the wells, in the cases of all the localities mentioned in the table No. I, from i to 14, have led to similar results. Natural gas from Vancouver and from Kokomo, Indiana, could not be tested at the wells. Tests made in the laboratory, of the samples received from those localities, gave the same nega- tive results. Another method of testing for hydrogen has been employed: As is well known, a jet of hydrogen is immediately ignited by platinum asbestos. Natural gas under similar conditions is not ignited, even when the gas-jet and the platinum sponge are mounted in an oven kept at a temperature approaching 300°. In order to ascertain the effects of different proportions of hydrogen and natural gas, a gasometer containing the gas-mixture to be tried was connected with a tip in the form of a drawn-out glass tube, above which some platinum asbestos was fixed. The gas pressure could be so regulated as to produce a pointed flame one inch long. By momentarily shutting off the gas by pinching the hose the flame could be extinguished, and the gas, being turned on again, played against the platinum asbestos. The length of the flame when the gas stream from the jet was ignited was there- 4o8 Phillips. fore a measure of the gas flow. The gas was ignited by the plati- num asbestos or not, according as the proportion of hydrogen in the natural gas was greater or less. The ignition of the gas was also dependent upon the temperature of the oven in which the jet and the platinum asbestos were fixed. Mixtures of hydrogen and natural gas produced glowing of the platinum and ignition of the gas at the following temperatures, when the experiment was made in a large iron oven whose tem- perature could be readily measured. The gas pressure was the same in all trials. Temperatures of the Oven at which the Gas Proportion of Hydrogen and Natural Gas. inflames as it strikes the Platinum Asbestos. Naturalgas 95 | 40°-50° Hydrogen 5 ' Natural gas 97.5 I 8o°-90° Hydrogen 2.5 ) Natural gas 99 ) 180° Hydrogen i ) Natural gas 99.5 ) 2io°-220° Hydrogen 0.5 ) Natural gas 99-75 > 270° Hydrogen 0.25 I Pure natural gas 270°-290° The observed temperatures naturally vary with the pressure, size of jet, etc., but trials with different pressures showed greater con- stancy than could be anticipated from a method so rough in appearance. The results corroborate those obtained by the more delicate tests. Experiments have also been tried with mixtures of air and natural gas which were exposed to palladium asbestos contained in glass tubes heated in the iron oven described under " Oxidation-temperatures of Hydrocarbons." It has been repeat- edly shown that, under such conditions, moisture is produced only at temperatures approaching or higher than the melting-point of cadmium iodide. The absence of free hydrogen has interfered with the use of natural gas in gas engines. The prompt, sharp explosion of coal- gas, so necessary for these motors, cannot be produced in the case of natural gas, which requires a higher temperature for its ignition, and explodes with less suddenness owing to the absence of hydrogen. The electrical devices for the igniting of coal-gas jets in dwell- ings by the spark of an induction-coil have not been so successful Oxidation and Chemical Properties of Gases. 409 where natural gas is used, owing to the higher temperature of igni- tion of a gas consisting of paraffins and containing no hydrogen. In laboratories where natural gas is the fuel, chemists have experi- enced the inconvenience that Bunsen burners and blast-lamps do not produce the high temperature easily obtained when coal-gas is used. Ordinary glass combustion-tubing cannot be softened by the employment of natural gas in a Berzelius blast-lamp. A coal-gas flame owes it steadiness and "stiffness" to the hydrogen which the gas contains. Natural-gas flames are much less steady and more easily extinguished by air currents. During May, 1892, a change occurred in the composition of the natural gas supplied to Allegheny City. The gas since that time and up to November, 1892, has been found to contain hydro- carbons which reduce dry palladium chloride. These hydrocar- bons are removed completely by digestion with fuming sulphuric acid, so that the gas after this treatment does not reduce palladium chloride. The nature of these hydrocarbons I have been unable yet to determine. Olefines. — Palladium chloride, iridium chloride, cerium dioxide in sulphuric acid, osmic acid, all remain unchanged by natural gas, cold or at 100°. Potassium permanganate is attacked with extreme slowness. Bromine-water has been repeatedly tried. The solution was in some cases cooled by ice to check evaporation of the bromine, and in others the bromine was added slowly, drop by drop, to compensate for its evaporation. In no case were any oily drops produced. / Prof. Sadtler, of Philadelphia, has in one instance obtained a considerable amount of heavy oil by the action of bromine on natural gas. My experiments seem to prove the absence of ethylene, propy- lene, isobutylene and trimethylene from the gas supplied to Alle- gheny. The same is true of gas from the localities mentioned in the table from i to 17. Tests could not be made at the wells in the case of gas from Kokomo, Indiana, and Vancouver, British Columbia, but samples brought to the laboratory gave similar results. The very low illuminating power of the natural gas of Western Pennsylvania is a further evidence of the absence of olefines, which, as is well known, are remarkable for the brilliancy of the light which they produce. By the kindness of Mr. J. W. Patterson, gas 4IO Phillips. inspector of Allegheny county, I am able to give the following data as to illuminating power: The gas supplied to Pittsburgh by the mains of the Philadelphia Company, November 30, 1892, pos- sessed an illuminating power equal to 10^/^ candles per five cubic feet of gas burnt per hour (mean of ten determinations). On the same date the illuminating power of the natural gas supplied by the People's Natural Gas and Pipeage Company was lo-^^g^ candles. Mr. Patterson's tests were made with a 36-hole Argand burner, having a chimney 7 inches long. Acetylene and Allylene. — Palladium-chloride solution is un- changed, as already stated. Cerium dioxide, mercuric chloride, gold chloride, silver nitrate, ammoniacal cuprous chloride and osmic acid are all unchanged. Hence, in the gas I have tested it may be said that no hydrocarbons of the acetylene series occurred. I have found no reference to acetylenes in any published analyses to which I have had access. Carbon Monoxide. — In the published analyses of natural gas, carbon monoxide is nearly always stated to occur. In my experi- ments, palladium chloride, gold chloride, silver nitrate in ammonia, iridium chloride, rhodium chloride, osmic acid, all used in solu- tion, were unchanged. Experiments have been made with Allegheny City natural gas in the following way : Gas has been caused to bubble for five weeks through ammoniacal cuprous-chloride solution. This solu- tion was then largely diluted with water and boiled. The gases expelled were collected and tested by palladium-chloride solution, but no carbon monoxide was found. It is true that, since the absorption of carbon monoxide in cuprous chloride has been shown to be a case of mechanical solution rather than chemical union, and that the absorbed monoxide can be expelled by a stream of other gases, the use of cuprous chloride for the absorp- tion and recognition of carbon monoxide cannot be implicitly depended on. Still, the direct tests above named lead me to the conclusion that no carbon monoxide occurs in our natural gas. Paraffins. — That the lower paraffins occur in natural gas needs no proof. Methane is the chief constituent. Small quantities of higher paraffins are usually present. Suipimr Compounds. — Pennsylvania natural gas does not con- tain recognizable quantities of either carbon oxysulphide, methyl hydrosulphide or methyl sulphide. Towards the western boun- Oxidation and Chemical Properties of Gases. 411 dary of the State it is possible that minute traces of sulphuretted hydrogen occur. The quantities of all such compounds are far too small to allow of their being easily identified, even in the case of large volumes of gas. The extreme delicacy of the reaction of methyl mercaptan towards palladium chloride would render it possible to detect exceedingly minute quantities of this compound should it occur. I have not had an opportunity to test the gas from the Western Ohio territory, which is said to contain sulphur compounds in considerable quantity. Nitrogen. — Natural gas, dried by calcium chloride and phos- phorus pentoxide, was passed over strongly heated magnesium powder. The magnesium was partly converted into a nitride, easily recognized by its reaction towards moisture, yielding ammonia in considerable amount. Repeated trials have been made of natural gas in the following way : A measured volume of gas was passed over ignited oxide of copper contained in a porcelain tube, the entire apparatus having been previously filled with pure carbon dioxide, which was caused to flow in a continuous stream for several hours in order to expel all traces of air. The escaping gas was collected in a eudiometer over mercury and the carbon dioxide absorbed by soda. There was left invariably a residue of gas unabsorbed by the soda and having no action upon palladium-chloride solu- tion. This residual gas was evidently nitrogen.' In the gas found in an artesian boring at Middlesborough, England, nitro- gen was found in large proportion.' Oxygen. — By the use of pyrogallol and soda, and by the oxida- tion of manganous hydrate in water, I have frequently been able to detect traces of oxygen, although on other occasions no oxygen could be found. It has only been recognized when the gas had bubbled continuously for many hours or days through the reagent. It cannot be said that oxygen is a constant constituent, although it does unquestionably occur in much of natural gas in minute traces. Carbon Dioxide is present in all natural gas, as is easily proved by its action upon lime-water. Ammonia. — In the case of a gas-well near Canonsburg, the following result was obtained : Gas was caused to bubble directly 1 See Table of Analyses, p. 422. Vol. XVI.-31. 412 Phillips. from the main at the well through water for several hours. On applying Nessler's reagent to the water, a feeble reaction was obtained. Ammonia was not found elsewhere in the trials I have made. Mr. S. A. Ford, of the Edgar-Thompson Steel Works, reports a very interesting case where masses of solid ammonium carbonate were blown out from a gas-well by the pressure of the gas. Natural gas appears to consist chiefly of methane, with traces of higher hydrocarbons of the paraffin series. Nitrogen is probably always present, together with a little carbon dioxide. The absence of free hydrogen, of olefines and of carbon monoxide is, I believe, clearly shown in the case of the natural gas I have examined. If natural gas as found in the wells of any one gas-region is derived from one vast subterranean reservoir, approximate uni- formity in composition should be looked for. It is often noticed, however, that gas from adjacent wells possesses a different odor. A carbon-dioxide determination was made in the case of samples of gas from six wells near Tarentum, Pa. These wells were situated nearly on a straight line less than one mile in length. The samples were all taken within an interval of three hours. The determinations were made with a y-per cent, soda solution in a eudiometer over mercury. Well. Carbon Dioxide. Well. Carbon Dioxide." No. I, 0.42 per cent. No. 4, 1.47 per cent 2, 1.25 5, 1.28 3. 0.25 6, 1.28 The differences in the proportion of carbon dioxide — a con- stituent determinable with great precision — would be difficult to explain if the gas flowing from these different wells is derived from one subterranean reservoir. VI. — Quantitative Analysis of Natural Gas, It is not possible to determine the proportion of the individual paraffins in a gas-mixture by the Bunsen method of explosion with oxygen unless it can be positively asserted that only two paraffins occur. This may be readily shown by an example. If a mixture of one volume each of marsh-gas, ethane and propane is burnt, the volumes of oxygen required, carbon dioxide and steam pro- duced will be as follows : Oxidation and Chemical Properties of Gases. 413 Oxygen Required. CO3. H^O Vapoi I vol. methane, 2 vols. I vol. 2 vols. I " ethane, 3^ 3 I " propane, 5 3 4 loj 6 9 Three volumes of ethane require for combustion ten and one- half volumes of oxygen, and yield six volumes carbon dioxide and nine volumes of steam. Hence a mixture of three gaseous paraffins could not be distinguished, in the case of a volumetric analysis, from the intermediate paraffin. Moreover, the heat of combustion of three volumes of the intermediate paraffin is almost exactly equal to that of a mixture of one volume each of the three. From this fact it follows that, as regards the calorific value of a mixture of paraffins, an exact determination of the character of the individual paraffins is not required. A saving of time, the possibility of using a larger volume of gas, the avoidance of a volumetric determination of water-vapor, are some of the advantages gained by a combustion over copper oxide. The application of gravimetric methods for the examination of gas is not new. Winkler' has described such a process for the analysis of mine-gas. Description of Method. — The process employed was, with some slight modifications, the same as described in the Annual Report of the Geological Survey of Pennsylvania for 1886 : Glass cylin- ders having stop-cocks at both ends, accurately calibrated by mercury and of 300-400 cc. capacity, were filled with natural gas. Where possible, this was done at the well. Before filling with gas, finely drawn-out threads of glacial phosphoric acid were inserted through the stop-cock into the vessel." After twenty- four hours the gas sample could be considered dry. The cylin- der was then connected with a porcelain combustion-tube, C, con- taining copper oxide. The general arrangement of the apparatus is shown in the accompanying sketch. Before the communication was made between the tube C and the glass cylinder A, air was expelled from C by pure nitrogen 1 Handbook of Technical Gas Analysis, p. 87. 2 Glacial phosphoric acid, on softening in the flame, may be readily drawn out like glass into rods of almost hair-like fineness. The quantity required was not sufficient to cause error in the gas measurements, inasmuch as the gas as it flows from the wells is in most instances remarkably dry. 414 Phillips. dried in the tubes M. The combustion-tube was intensely heated during the passage of the nitrogen. After expulsion of air by nitrogen, the natural gas was caused to flow over the copper oxide previoualy heated for some time. The movement of the gas through the combustion-tube was controlled by means of mer- cury, which flowed from the funnel D into the gas-cylinder, and was so regulated that two hours were required for complete com- bustion. Experiments showed that there is no danger of produc- tion of carbon monoxide or unsaturated hydrocarbons when the gas-stream is slow. After the gas had been expelled from the cylinder A, it was rinsed by lowering the mercury funnel so that nitrogen passed down into the cylinder, to be again driven out by raising the funnel. After the gas had been fully burned, air (purified by the lower system of dryirig-tuoes in the sketch) was passed through the apparatus till the nitrogen and moisture had been fully displaced and the process was then complete, the carbon dioxide and water being determined by weight. The method, as is seen, gives merely the proportions of carbon and hydrogen. As the exact percentage of the parafiins in the gas-mixture cannot be ascer- tained by analysis, an approximation alone is possible. The composition by weight of some of the lower (gaseous) paraffins is as follows : Paraffins. Carbon. Hydrogen. Methane, 74'97 per cent. 25.03 per cent. Ethane, 79-96 20.04 Propane, 81.78 18.22 Butane, 82.72 17.28 In the following table, the calculated composition by weight of various mixtures of methane and ethane is given (the atomic weight of carbon being 11.97) • Mixture of Methane. Ethane Carbon. Hydrogen. I vol. I vol. 78.22 per cent. 21.78 per cent. 3 2 77-73 22.27 2 I 77.38 22.62 5 2 77.11 22.89 3 I 76.89 23.11 7 2 76.70 2330 4 I 76.56 23-44 5 I 76.30 23.70 6 I 76.15 23.86 9 I 75.82 24.18 -f 2hirof^ N«.OH 1 • r mm^ Apparatus for Quantitativ 1 ' J I '■:■ n|]yy 1 /. i¥^ k^ Apparatus for Quantitatiye Analysis of Natural Gas. Oxidation and Chemical Properties of Gases. 415 From a gravimetric analysis of natural gas it is easy to deter- mine the relative proportions by weight of carbon and hydrogen in unit volume, and from these the composition may be stated volumetrically in terms of ethane and methane, by the use of the preceding table, and with a fair approximation to the truth. It is probable that minute quantities of propane and perhaps higher paraffins occur, but these cannot be identified. The nitrogen and carbon dioxide being determined, the volume of CH. + OH« + C3H8+ is obtained as a difference. The error involved in such a method may then be exactly defined as follows : The hydrocarbons may consist of methane with traces of propane, or of methane with ethane or butane, but the analysis will be stated volumetrically in terms of methane and ethane only. As regards the question of fuel value, I have endeavored to show' that the above method will give closely approximate results when certain factors relating to available heat of combustion of paraffins are used. The gravimetric method affords at the same time a means of control, for it is not only true that in a given volume of a particular paraffin, or of a mixture of paraffins, the hydrogen and carbon will occur in definite quantity, but the ratio ^ is a constant, and will be greater as the proportion of higher to lower paraffins is greater. These considerations will serve to show the limits of accuracy of the method. Nitrogen was determined by passing a measured volume (100 cc.) over ignited copper oxide contained in a porcelain tube, and then into a eudiometer containing soda solution. By means of a stream of carbon dioxide continued for several hours, the air was expelled from the apparatus previous to the combustion of the gas. In presence of a large excess of carbon dioxide, combus- tion by copper oxide is greatly retarded, and the process must be conducted very slowly to effect complete oxidation. Oxygen, as already stated, occurred in too small proportion to allow of a quantitative determination. Carbon dioxide was determined by soda solution in a eudior meter over a mercury trough. iSee Rep. Geol. Surv. Penna. 1886. ■» fO " "■ N t^ •Kiquinion • ^spug 'jPAiioauE^ " vdd 0000^10 00 1 TVO i-.ro VO ° Tf 10 •|1»M 'l«S ^'!n ^usMSailV VO t^ d N d Vl Tf Ovj 1 t>> N °t ^. 1 1 Tfvo •P"I 'OUI03105I ►C' VO d ro 1 d »A Tf ON 1^ r^ Mm rv 00 N ro Tf N Tf u^ •«d '"'P«a ■^ N d t^ d vd fo " 00 00 -. Ov •Bj 'qSinqsjiij n ^ "T ° V u^ rr 'IPAV s.-oO ^ "luiBj " d d oc V 8 VO Tf 'd- VO VO Cr 00 N •ej 'uomSpio N roc Ovd d vOtJ- a r^ N 00 ^v8 ro N "■ 1 •O •puBpA3,3 ^ \0 d r' ■> 6 VO ro a r^ N ON 00 ro VO u « TfVO •Bj 'spnnojo uoijiqinxg qSjnqsjiij r-^ d 2 «^ d V •OTf 00 -> VOVO T '^. "^ «00 •=- i 00 8 vd ro " ., ., ^, N M 00 q^s ' s s Tf m •BJ '3133^3 UOODOEH r^ Ov 2 2 2 C • i vd ro r< 00 VO Tf M 8 '^ OvO •«d '9n!AS'^"niV 'un^jsuoXq VO N c t* t> i V l>. (^ « vri rj — =~CTr VOO 1> T »00 ■8d 'itJiD I!0 "=" 'XiqDsads u^ ■^ d n! vr 1 d t^ « i <7 V f-, N >- 11 00 00 OS d « d •Bd '03 uBa;H3I^ 'xoouas. TT d ro i a t^N 0^0 -, - t^ ro r-.N JJ t>, N •Bd -03 UB95[0IV 'auB5i fo dv d rt d vd ro i ^ V t-. N ^^ J. vS 0\ rt C ■ § S^ •BJ '-03 U3JJBAV 'PPHJ3MS C) 0" VO ro V t^N Tf « ,> "■ 1 TfVO ' "^t y ° ►- 00 •A -N '«!uop3jj « dvd 2 d d C« j;j i; c V c X c o'c c E ■-£ c u b > c 3 rt The paraffi these ga the foll( tion by b C •a > : Oxidaiion and Chcfnical Properties of Gases. 417 VII. — Origin of Natural Gas and Petroleum, Soon after the early discoveries of oil and gas in Pennsylvania, geologists proposed an hypothesis to account for the origin of these remarkable substances: Remains of the marine vegetation of the Devonian inland sea, as they were gradually buried under the later accumulations of sediment and exposed to gentle heat from below, underwent a slow process of destructive distillation. In this way all the varieties of petroleum and natural gas were produced. This view, adopted from a purely geological stand- point, seemed so plausible that for a long period no other was thought of. Mr. J, F. Carll, of the Second Geological Survey of Pennsylvania, has discussed the hypothesis very exhaustively in his various official reports. If this view is correct, oil and gas are probably stored products, and are not being continuously gener- ated at the present time. Opposed to this view is the more strictly chemical hypothesis of Mendel6eff, who, in 1876, expressed his belief that petroleum and gas are of igneous origin. On account of the high value assigned by astronomers for the mean density of the earth as compared with that of the surface rocks, it follows that the heavy metals are mainly accumulated at great depths where a tempera- ture of fusion may be assumed. Many of these metals combine readily with carbon to form carbides. Iron, in form of a carbide, when exposed to steam at high temperatures, is rapidly oxidized, the hydrogen of the water then combining with the carbon set free and producing hydrocarbons. Citing experiments of Cloez, who produced mixtures of hydro- carbon oils by the action of hydrochloric acid upon ferromanganese, Mendel^eff concluded that such reactions have occurred at great depths below the earth's surface by the contact of steam with incandescent metallic carbides. " During the upheaval of moun- tain ranges, crevices would be formed at the peaks with openings upward, and at the foot of the mountains with openings down- ward. Thus there was opportunity for the water to penetrate to great depths and for the hydrocarbons to escape. The situation of naphtha at the foot of mountain chains is the chief argument in my hypothesis." ' According to this view, oil and gas are being continuously generated, for there is no reason to suppose that the masses of ' Mendeleeff : Principles of Chemistry, Vol. I, p. 365, 4i8 Phillips. metallic carbides in the earth's interior are exhausted; such, in fact, seems to be Mendeleeff's view. He points especially to the absence of large quantities of nitrogen compounds in petroleum as an argument in favor of the hypothesis. The objection has been urged against this hypothesis that petroleum, if thus produced, should be .abundant in the primary rocks from which it is usually absent. The original heated condition of these rocks would have prevented the condensation of oil, however, and although the vapors may have passed through the earlier rocks, there is no reason to expect thai con- densation should have occurred before reaching much higher strata. While on geological grounds difficult to prove or disprove, it meets with one fatal objection : the composition of natural gas in Pennsylvania does not justify the supposition that superheated steam and carbon have been concerned in its formation. We should certainly look, in such a case, to find natural gas composed mainly of free hydrogen containing small quantities of paraffins, defines and carbon monoxide. When it is considered that paraffins alone cannot under any known circumstances be produced from the oxidation of carbide of iron by steam, the hypothesis does not seem to be tenable. It is true that varying conditions of temperature might have produced a great variety of hydrocarbons, but no evidence has yet been obtained that paraffins alone result from such a reaction. In an experiment made with ferromanga- nese and dilute sulphuric acid, the gas evolved was found to contain 6 per cent, of olefines.' It is further to be noticed that this hypothesis requires that water should take part in the process, yielding up its hydrogen, while, according to the older geological hypothesis, the water may have served mainly to cover and give protection from atmospheric oxidation, if it has been concerned at all in the reaction. Water contains dissolved oxygen, and in descending to the iron carbides, must have given off its dissolved oxygen long before reaching the region at which actual formation of hydro- carbons could occur. Hence, on this hypothesis, oxygen should be found in natural gas in larger quantity than the chemical tests indicate. In fact, in rocks of moderately high conducting power, a wide interval would exist between the depth at which water ' Experiments by F. C. P. Oxidation and Chemical Properties of Gases. 419 boils and the much greater depth at which water-vapor could oxidize metallic iron in quantity. It is doubtful whether water could have traversed this interval so as to reach the latter depth at all. Engler' has published the results of interesting investigations upon the distillation-products of menhaden fish-oil. By conducting the distillation at a high pressure (25 atmospheres), this author produced a mixture of hydrocarbon oils from which a large number of normal parafifins was obtained, compounds not found elsewhere in nature than in petroleum. This has led to the revival of an older theory as to the origin of petroleum and gas, i. e. that they have resulted from the distillation under pressure and at low temperatures of the accumulated remains of marine life buried under the sediments of the ancient Devonian seas. Much has been written in support of the hypothesis of Engler, and it may be said to have gained very general acceptance in Europe. Ochsenius'' has summarized many of the arguments usually adduced in support of the hypothesis: This author says, " Concerning the origin of petroleum there is now no doubt that, with a few exceptions, animal remains (mainly of marine life) have yielded the raw material." Originally the opinion was held that it was derived from vege- table matters, because the accumulation of animal remains suffi- cient to account for its formation by any distillation-process in the rocks could not be explained. Distillation of vegetable matters would, however, have left greater deposits of coal (as a residue in the Devonian rocks). But petroleum occurs in rocks of marine formation where coal is uncommon. Rocks in which plant remains are found do not contain bitumen (petroleum). If animal remains are associated with those of plants, then bitumen is usually found. The objection urged against the hypothesis of Engler, that nitrogen does not occur in petroleum, is easily over- come by the fact that the nitrogen of animal tissues tends finally to produce ammonia, and this in the case of petroleum may have been carried away in solution by water ; hence the absence of nitrogen compounds. From Engler's experiments it appears that animal fats are the chief source of petroleum. ■ Ber. d. chem. Ges. 31, 1816 (i883) ; 33, 592 (1889). -Chem. Ztg. 1891, 956. 420 Phillips. It is true that fatty matters do not ordinarily sink in water, although Von Gucmbel, in the voyage of the Gazelle, found fat globules in dredgings from the bottom of the Atlantic Ocean, in water 15,000 feet deep. Putrefactive changes would tend to yield considerable quanti- ties of ammonia and carbon dioxide." These in presence of salt water would produce alkali bicarbonate and ammonium chloride. Hence, alkaline waters might be looked for in the neighborhood of petroleum. The petroleum at Pechelbronn is associated with water containing 0,5 per cent, of alkaline carbonate.' Such alka- line waters are not known in archaean rocks, and are not, there- fore, likely to be derived from greater depths than the rocks in which they are found. Probably no cases can be cited where fatty tissues alone of buried animals have yielded oil or gas. The presence of strongly saline water is apparently needed. Great differences occur in the chemical character of petroleum. Caucasian oils are mainly composed of olefines or substances related to the olefine group. The German oils are mixtures of par- affins and olefines, while the American are chiefly paraffins. Such differences may be attributed to the character of the rock in which the distillation has occurred. Sandstones would probably prove without action ; while limestones, by reason of their basic charac- ter, would tend strongly to influence the products. Such are some of the arguments of Ochsenius in favor of Engler's hypothesis. If this view is accepted, it follows that the generation of petro- leum and gas must be considered as a finished process, so far as all existing productive gas and oil regions are concerned. Engler has analyzed^ the gas evolved (i) when menhaden oil and (2) when oleic acid is distilled («) under atmospheric pressure and {b) under a pressure of 25 atmospheres. Menhaden Oil. Oleic Acid. I atmos. 25 atmos. i atmos. 25 atmos. Methane 25.2 38.3 9.3 4.36 Olefines 11.4 7.8 12.5 2.9 Carbon dioxide 26.7 17.4 37.2 26.0 Carbon monoxide 34.9 34.5 38.6 25.5 Incombustible residue, 1.8 2.0 2.4 2.0 » In Western Pennsylvania many cases are known of water having a decided alkaline reac- tion in the neighborhood of gas-wells. In the Murrysville gas territory, water of alkaline reaction was so abundant as seriously to interfere with gas development. — Note by F. C. P . 2Ber. d. chem. Ges. 3S, 592 (1889). Oxidation and Chemical Properties of Gases. 421 The liquid distillates produced at the same time that these gases were evolved were rich in the normal paraffins and their isomers. 100 parts of menhaden oil yielded 8.9 parts of gas and 63 parts of liquid oils. A strong argument in support of the Engler hypothesis is found in the fact that by distillation offish-oils, besides methane, several of the lower paraffins are produced in large quantity. Hydro- carbons of the paraffin series are not obtainable in such propor- tions by the distillation at high temperatures of other organic material under ordinary conditions. It should be noted as a fact of much interest as regards the results of Engler's researches, that in the distillation at higher pressures the proportion of olefines contained in the gases evolved is considerably less. This is also true of carbon monoxide when oleic acid was distilled. It is to be regretted that Engler's experiments were not repeated at still higher pressures, in order to ascertain whether these same constituents of the evolved gases diminish progressively with increased pressure. Engler was the first to show clearly that the problem of the origin of oil and gas must be studied from the chemical rather than the geological standpoint. The hypothesis advanced by this author has been very generally accepted. Nevertheless, my examinations of natural gas have led me to doubt some of his con- clusions, well founded as they seem. The most careful tests, carried on during a period of six years, have failed to show the presence of either olefines or carbon monoxide in the natural gas of Western Pennsylvania. Some of the constituents of gas are soluble in water. This is notably the case with carbon dioxide, butane, hexane, etc. If ethylene and carbon monoxide have been produced in the rocks even in much smaller proportion than Engler finds in menhaden- oil gas, these substances would now occur in the natural gas of Pennsylvania. Ethylene would give to the gas such illuminating power that there would be no occasion for the use of coal-gas in any town in the Western Pennsylvania gas-region. As a matter of fact, natural gas is almost useless as an illuminant, its light being equal to 5 to 11 candles per five feet of gas consumed per hour. Mr. Robert McKinney, formerly gas inspector of Allegheny county, found as a mean of forty trials of natural gas supplied to Pittsburgh an illuminating power of 6.5 candles. Mr. J. W. 422 Phillips. Patterson, the present gas inspector of the county, states that the illuminating power of natural gas as supplied to Pittsburgh in November, 1892, is a little less than 11 candles per five feet per hour. The reason for this is that natural gas, as found in Pennsyl- vania, does not contain olefines. If carbon monoxide occurred in gas there w^ould have been innumerable cases of poisoning among workmen at gas-wells. It is common to find such leaks of gas about the majority of gas-wells that no one could strike fire at a well without risk of fatal consequences. Although inhaling the escaping gas for much of a lifetime, a gas-well driller will usually maintain that no bad effects to health come from exposure to the gas. Air containing 0.2 per cent, of carbon monoxide is known to produce dangerous effects upon health. According to Wyss,' air containing o.i per cent, of water-gas is poisonous to breathe. It is hardly probable, moreover, that carbon monoxide or ethylene occurring in gas could have been absorbed or removed at low temperatures by any natural process in the rocks. Unlike carbon dioxide and anmionia, their slight solubility in water would preclude the supposition that they had been dissolved away. Muck'' cites analyses of fifty-seven samples of gas from coal mines and of gas occluded in coal. In only one case is carbon monoxide mentioned, but it is distinctly stated that its occurrence was not proved. Ethylene is mentioned in six cases, but Muck states that more recent analyses have failed to demonstrate its presence usually in gas from coal. The absence of hydrogen in all the analyses is especially noticeable. In the case of gases from the Caspian region, the presence of ethylene and carbon monoxide is to be anticipated, as, from all accounts, subterranean heat has been concerned in their production." > Ztschr. ang. Chem. 1888, p. 465. - Grnndzijge und Ziele der Steinkohlenchemie (1S81). 3 Constituents. i. 2. 3. 4. 5. 6. 7. 8. Carbon monoxide.... . 00 000000 Carbon dioxide 0.95 2.18 3.50 o 2.47 4.44 o 03 Olefines 4.11 3.26 4.26 0000.. Methane 92.49 93.07 92.24 95.39 97.57 95.56 1.90 Hydrogen 0.94 0.98 o o o o o Nitrogen 2.13 0.49 .. .. .• .. 9^-57 96-8 Oxygen .. .. .. .. .. 153 29 100.62 99.98 100.00 .. 100.04 100.00 100.00 100.00 Nos. 1, 2, 3, 4, 5 and 6: natural gas from the Caspian region. Communicated by letter from Mr. M. Beliamin of Nobel Bros., St. Petersburg. No. 4 is the result of a partial analysis. Nos. 7 and 8, gas obtained by deep borings at Middlesborough, England (Bedson : J. Soc. Chem. Ind. 7, 662 (i838)). Oxidation and Chemical Properties of Gases. 423 Thomas' gives analyses of fourteen samples of gas occluded by coal, and also of gas from blowers in coal mines in New South Wales. The analyses showed the presence of methane, nitrogen, carbon dioxide and oxygen ; but no carbon monoxide, hydrogen or ethylene was found. Franke^ gives analyses of mine-gases, according to which only carbon dioxide and methane were found. Winkler^ found no hydrogen in nine samples of mine-gas. Many similar statements might be cited, all tending to prove that hydro- gen, ethylene and carbon monoxide do not occur in gases occluded in coal. The occurrence of gas consisting of nearly pure nitrogen, such as that obtained at Middlesborough, England' — in a region there- fore where gas similar to Pennsylvania natural gas might be looked for — may perhaps be explained by the action of subterranean water upon deposits of coal or bituminous shale. The dissolved air in such waters, by causing slow oxidation, might lead to the production of carbon dioxide and the consequent removal of oxygen from the water. The carbon dioxide produced would lessen the solubility of the water for nitrogen, by causing the water to dissolve carbonate of lime, etc. Gentle heat from below would tend still further to the expulsion of the nitrogen, and thus a con- siderable but limited quantity of nitrogen might be obtained as a sudden outburst from a drill-hole. It may be said that varying conditions of temperature and pressure, and kind of rock, have modified the products, so that perhaps the carbon monoxide and ethylene resulting from a labo- ratory experiment have, in nature's workshop, given place to paraffins. But, if the chemistry of the reaction supposed to occur is to be considered at all, the fact that distillation- experiments have produced from fish-oil certain bodies found in natural gas (paraffins) should not count more forcibly as geological evidence than the other fact that such distillation yields bodies which are foreign to natural gas as usually found in Pennsylvania. I have failed to find any data tending to show that organic matter can be subjected to destructive distillation in such a manner as not to yield carbon monoxide and considerable quantities of olefines, together with hydrocarbons of still less saturated character. As a rule, the acetylenes and members of the benzene series appear. 1 Watls' Die. (ist Ed.), 3d Suppl. 529. 2 j. prakt. Chem. [2] 37, 91, 113 (1888). ' Jsb, 1882, 1063. 424 Phillips, Engler's hypothesis involves the supposition that a process of dis- tillation has occurred at moderately high temperatures and at pressures measured by great depth of rock-strata. The carbon dioxide evolved in this destructive distillation must have come continuously into contact with the vast quantities of carbon which in its various stages of transformation from vegetable tissue to anthracite is so widely distributed throughout the rocks. The reaction C03-j-Ci= 2CO, which proceeds rapidly at a strong heat and also slowly at lower temperatures, would then probably have occurred wherever the temperature was sufficiently high. Prolonged contact of carbon dioxide with the carbonaceous residue of the distillation would perhaps be sufficient to increase considerably the final yield of carbon monoxide. According to I. L. Bell,' the reduction of carbon dioxide to carbon monoxide by carbon in the form of soft coke begins at 427° C. This is about the temperature at which Engler's distillation-experiments were conducted (36o°-42o° C). Engler has shown that distillation of animal fats at very high pressure (25 atmospheres) may yield gas containing less of carbon monoxide and olefines than when the process is conducted under atmospheric pressure. No data are at hand as to results at still higher pressure. If it is conceded that the proportion of carbon monoxide and ethylene in the gas evolved during destructive dis- tillation decreases progressively with increase of pressure, and that these two constituents vanish altogether at sufficiendy high press- ures, it would still seem necessary to suppose that the pressure must have been at least twice as great when the process occurred in the rocks, as in the case of Engler's experiments. Taking the specific gravity of the rocks to be about 2^, it may be assumed that twelve feet of rock-strata represent a pressure of one atmos- phere ; six hundred feet of solid rock would then be required to produce a pressure of 50 atmospheres. This would be consider- ably less than the depth of the same quantity of rock material in the form of loose sediment before its consolidation. No case can be cited in recent times where sediment six hundred feet deep has been so suddenly accumulated as to bury unchanged the vast quantities of animal remains necessary to account for the produc- tion of oil and gas upon Engler's hypothesis, that oil and gas have 1 Chemical Principles of the Manufacture of Iron and Steel, p. loi.] Oxidation and Chemical Properties of Gases. 425 resulted from the action of pressure and moderate heat upon animal matters. There is probably no reason to suppose that the gaseous ole- fines have, under the influence of pressure, given place to others of higher boiling-point by a process of polymerization. Should the possibility of such a change be proved, the absence of olefines from natural gas and their presence in petroleum might be explained. The possibility of secondary reactions among the con- stituents of a complex gas-mixture at high temperatures and under pressure adds difficulty to the problem, and caution is needed to avoid the error of overestimating the importance of any given reaction. It is generally true, however, that under such conditions secondary changes are probable, and that unsaturated compounds — olefines, acetylenes, carbon monoxide — are likely to result, especially when water-vapor and carbon dioxide are present. It is a well-known fact that when petroleum is distilled, con- siderable quantities of unsaturated hydrocarbons are produced which did not exist in the original crude oil. This is shown by the bromine-absorption of the different products. The process of "cracking" or breaking-up by heat of the hydrocarbons in petroleum into simpler and less saturated compounds is familiar to all oil-refiners. Chemically speaking, " cracking " means the production of unsaturated hydrocarbons. The fact that Engler has, in his extremely interesting and im- portant researches, produced by distillation of animal matters so great a variety of paraffins constitutes by far the strongest argu- ment in favor of his hypothesis. Sorge, in an article which has been reproduced in numerous journals,' has stated that a strong resemblance exists between Pennsylvania natural gas and gas manufactured from Westphalian coal. Similarity in composition between natural gas and coal-gas would greatly simplify the problem of origin, and the fact of such similarity would prove of great interest. In this connection the following analysis of gas from Westphalian coal, carried out in the laboratory of the Westphalian Berggewerkschaftskasse in Bochum, will be of interest. I am indebted to Mr. Bergassessor E. Krabler, of Bochum, for the figures which he has kindly com- municated by letter. » SUM ttnd£isen, 7, 93; Jahrb. Min. 1887, 8, Ref. 318; J. Chem. Soc. 54, 31 (1888). 5 4 45 35 4o 50 5 5 I 7, 4 3 426 Phillips. Hydrocarbons, C^H, Methane Hydrogen Carbon monoxide Carbon dioxide Nitrogen The large percentage of hydrogen and the proportion of carbon monoxide in this gas illustrate at once the results of high temper- ature in the production of the coal-gas, but a similarity between this coal-gas and natural gas can hardly be said to exist. When vegetable remains are buried under water, as is well known, decomposition occurs, yielding gas in considerable quan- tity. Tappeiner' has studied the products of this change very exhaustively. Pure cellulose (filter-paper) was found, under the influence of a microbe which was supplied with nutritive fluids, to dissolve in water, yielding gas-mixtures of two different types. Carbon dioxide "I Hydrogen sulphide J Under water of neutral reaction. At beginning. At end. 85.48 76.98 Under slightly alka- line water. 55.39 Hydrogen 0.0 0.0 42.71 Methane 11.86 23.01 0.0 Nitrogen 2.73 0.0 1.90 From these experiments it appears that, by the action of a microbe, either methane and carbon dioxide (neutral fluid), or hydrogen and carbon dioxide (alkaline fluid) may result. Hoppe- Seyler'' found that gas evolved in the decay of cellulose under the influence of a microbe (marsh-gas fermentation) contained: Carbon dioxide 50 Methane 45 Hydrogen 4 Popoff finds in a gas from decaying vegetable matters : Marsh gas 68.56 Carbon dioxide 3i«44 Berthelot states that hydrogen is produced in the vinous fer- mentation of mannite. In very careful experiments which I have 'Ber. d. chem. Ges. 16, 1734 (1883). "^Ihid. 16, 122 (1B83). Oxidation and Chemical Properties of Gases. 427 tried I have failed to find hydrogen in the gas evolved during the fermentation of 200 grams of sugar. Chemical changes of this type are not likely to be of importance, however, as regards the hydrogen question. Gases from Sea-weeds. The following experiments were tried in order to study the nature of the gases evolved in thedecay of sea-weeds : A quantity of a large fucus kind from Santa Barbara, California, was used. 50 grams of the air-dried plant were soaked in water and then introduced into a flask filled with water, which had been previously boiled (in order to expel air) and cooled. The flask was con- nected with a bell-jar over a mercury-trough. After setting up the apparatus, no gas appeared until the third day ; then a strong evolution of gas began and continued in slowly diminishing quan- tity for ten days, when the process ceased. In all, 803 cc. of gas were collected. Analyses were made (i) of the first portion of 300 cc. and (2) of a second portion of 300 cc, and (3) of the last portion of 203 cc. The results are tabulated below : First Portion, , Second Portion. Third Portion. Carbon dioxide 18.23 32.47 5344 Carbon monoxide ... ... Ethylene ... ... ... Methane .30 .28 .08 Hydrogen 62.24 48-97 42.02 Nitrogen 19.23 100.00 18.28 4.46 100.00 100.00 Carbon dioxide was determined by soda solution over mercury ; hydrogen by palladium asbestos, using a Hempel apparatus. The absence of carbon monoxide and ethylene was proved by palladium-chloride solution. Methane was determined by com- bustion with air, using a red-hot platinum tube. The carbon dioxide produced in the combustion was absorbed by baryta solu- tion of known strength, and the excess of baryta determined by standard oxalic acid. The following facts are of especial interest: I. The carbon dioxide increases towards the end of the decay. 2. The hydrogen steadily diminishes. 3. Methane occurs only in traces, 4. Nitrogen occurs in such considerable quantity as to render it probable that this gas is set free in the process of decay. Vol. XVI.-3a. 428 Phillips. The same apparatus was kept in position for two and a half years after the above experiments were finished. During that time a continuous production of gas was observed, but it was so slow that at the end of this period only about 30 cc. of gas col- lected. This was found to consist of methane. I have examined the gases produced in swampy ground in many different places. Samples were taken from streams having muddy bottoms and in which vegetable matter had collected. Samples of gas have also been taken from salt marshes on the coast of Maine. Gas has also been collected from the very deep accumulations of mud and decaying vegetable remains found in some parts of Lake Chautauqua. The general result of examina- tions of these gas samples may be stated to the effect that the gas occurring in shallow swamps and streams consists of methane, carbon dioxide and nitrogen. In some of the much deeper swamp waters, where masses of vegetable debris of greater thickness are found (as in Lake Chautauqua), hydrogen occurs in very small quantity. Great difficulty is experienced in taking samples of gas from localities of the latter type. Tappeiner observes that the marsh-gas fermentation is probably a very important source of methane in nature. The fact that buried vegetable matters may, after a brief period of rapid gas-evolution, pass into a condition of extremely slow decay, adds greater force to the original theory of petroleum and gas. The occurrence of so large a proportion of free hydrogen among the gases evolved by vegetation in process of decay is a matter of great interest, as it suggests the existence of an impor- tant source of hydrogen wherever deeply submerged plant- remains occur. Frankland and Jordan ' found that grass left to decay under water (air being excluded) evolved in three days gas of the following composition: Carbon dioxide, 84.63 Oxygen, 0.13 Hydrogen, 6.90 Other combustible gases, 2.51 Nitrogen, 5.83 Vegetable tissue, after the somewhat sudden and tumultuous evolution of gas, seems to be capable of relapsing into an 1 J. Chem. Soc. 43, 295 (1883). Oxidation and Chemical Properties of Gases, 429 extremely slow and long-continued process of decay. After the first decomposition, such remains might become accumulated and buried deeply under sediments before the tissues are materially altered. The generation of gas might then proceed in the cold. It seems hardly possible to ignore this probable source of natural gas in discussing any theory as to its origin, especially when it is considered that no other process in nature has been found to yield a gas at all similar in composition to that found in the rocks. Of the three hypotheses which have been proposed to account for the production of oil and gas, two are open to a serious objec- tion. I. — The chemical changes supposed by Engler to have been the cause would probably yield gas different in composition from the natural gas now being obtained in such large quantity in Western Pennsylvania ; and if the gas originally contained ethy- lene and carbon monoxide, it is not easy to explain their complete disappearance in the natural gas I have examined from wells scat- tered over so large a region. 2. — The hypothesis of Mendel6eff would be much more difficult to reconcile with the facts as regards composition. The total absence of hydrogen could not be easily explained. The only process in nature which is known to yield gas similar in its con- stituents to natural gas is that which occurs in swamps and decay- ing masses of submerged vegetable remains. The important fact that the solid plant tissues may be preserved for long periods after the preliminary gas-evolution has ceased shows that the remains are likely to become slowly buried, to undergo the " fermentation " changes leading to the production of methane. Animal tissues can suffer no such arrest of decom- position. Decay once set in is carried rapidly onward to com- plete destruction without intermission. The contrast between the conditions in which animal and plant-remains occur in the rocks seems to justify this statement. If chemical evidence shall count in the discussion, it is difficult to find a more satisfactory explanation than the older hypothesis which the geologists advanced, although in their treatment of the subject the strictly chemical arguments were neglected. 430 Walker. THE CONDENSATION-PRODUCTS OF AROMATIC HYDRAZIDES OF ACETACETIC ETHER.— INDOL AND PYRAZOL DERIVATIVES. By C. Walker. In a former paper I showed that the product of the reaction between acetacetic ether and concentrated sulphuric acid is ethylic a-methylindol-/3-carboxylate.' With the permission of Dr. Nef, I have continued this work, applying the reactions to homologous compounds. As considerable diversity of experience with acetacetic-ether phenylhydrazide has been recorded by different authors,' 1 shall briefly give my own experience in working with the hydrazides. In my earlier preparations of these substances, especially in the case of oxalacetic-ether phenylhydrazide, I experienced difficulty in preventing the splitting off" of alcohol and consequent forma- tion of pyrazolone. If instead of adding the hydrazine to the acetacetic ether the ester be added to the hydrazine, no pyrazolone is formed. Operating in this manner, only a very small volume of absolute ether is necessary, and in the prepara- tion of all hydrazides except oxalacetic-ether hydrazide, the ester may be poured slowly into the hydrazone. When about half of the acetacetic ether (or substituted acetacetic ether) is added, a turbidity due to the separation of water is noticed. The addition of the ester is continued, accompanied by vigorous shaking all the while. The mixture should not be cooled. After standing 3 to 4 hours the water is removed by means of a separatory funnel, and the ethereal solution dried with calcic chloride or anhydrous sodic sulphate. This was the method by which all hydrazides experi- mented with were prepared, except when special purity was desired, as in the numerous attempts to obtain ethoxypyrazols from them by means of acetyl chloride. The hydrazides were then washed in ethereal solution with alkali and acid. Oxalacetic- ether phenylhydrazide cannot be washed in ethereal solution with dilute sodic hydrate (i : 20) without almost complete conversion into pyrazolone. All other hydrazides, however, can be thus treated with only partial conversion into pyrazolone. All of the 1 This Journal 14, 576. s Ibid. 14, 516. The Conde7isation-produds of Certain Hydrazides. 431 hydrazides (including oxalacetic hydrazide) can be washed in ether solution with dilute sulphuric acid with only a slight forma- tion of pyrazolone. The tolylhydrazides are much more stable towards alkalies than phenylhydrazide. That the hydrazides are purer when washed with alkali and acid is shown by the following observations, which were often made. The washed and unwashed hydrazide were placed side by side in the same desiccator ; after a few days' standing the unwashed hydrazide was almost com- pletely changed into pyrazolone, while the washed hydrazide resisted this change for some weeks. In their deportment towards concentrated sulphuric acid the hydrazides have been found to behave quite differently. Acet- acetic-ether phenylhydrazide when treated with strong sulphuric acid at low temperatures gives an indol derivative; acetacetic- ether orthotolylhydrazide and acetacetic-ether paratolylhydrazide also lose ammonia, giving indol bodies; methyl- and benzylacet- acetic-ether phenylhydrazides, when similarly treated, give pyra- zolonesulphonic acids : ethylacetacetic-ether phenylhydrazide, under the same treatment, loses ethylamine, with consequent for- mation of an indol derivative; while oxalacetic-ether phenyl- hydrazide loses water and gives rise to an ethoxypyrazol. The different compounds obtained will be described in the order of indol, pyrazol and pyrazolone derivatives. I. — Indol Derivatives. Ethylic paraiolyl-a.-methylindol-^-carboxylate. — The acetacetic- ether paratolylhydrazide from which this body is derived, was prepared by the method already indicated, viz. by adding acetacetic ether to paratolylhydrazine partly dissolved in a small volume of absolute ether. The oily hydrazide was dropped slowly into about five volumes of concentrated sulphuric acid at — 15°. Not more than 10 grams of the hydrazide should be used in one operation. After standing 10-15 minutes the sulphuric- acid solution was poured slowly upon ice contained in a funnel, the drippings being received in a beaker containing ice. The solution was then extracted with ether, the ethereal solution washed repeatedly with alkali and dried with calcic chloride. The yield of indolcarboxylic ester was a little less than 30 per cent. After recrystallizing from dilute alcohol and drying over sulphuric acid in a vacuum, it was analyzed : 432 Walker. I. 0.1475 gram substance gave 0.3877 gram CO? and 0.0917 gram H2O. II. 0.2171 gram substance gave 0.5692 gram CO2. III. o. 1797 gram substance gave 10.5 cc. N at 23° and 757.5 mm. Found. Theory for C,,H,5NO,,. I. II. c 71.88 71.69 71-50 H 6.91 6.90 6.96 N 6.45 6.71 The acetacetic-ether paratolylhydrazide, therefore, has lost one molecule of ammonia; that the splitting off of ammonia has taken place in precisely the same manner as in the case of acetacetic- ether phenylhydrazide when similarly treated will appear presently : OiH hICCOOR ^^ CCOOR I I ^C.CH3=r CHf Y )CCH3 + NH.. Inh— Inh \-^"^nh Ethylic paratolyl-a-methylindol-i5-carboxylate (or B3, Pr2-di- methylindol-Pr3-carboxylate) crystallizes from dilute alcohol in regular octahedra and tetrahedra melting at i63°-i63.5° without decomposition. It is readily soluble in alcohol, ether, benzene and acetone. Like the indol derivative from phenylhydrazide, it gives the indol splinter-reaction. It is very stable towards alcoholic caustic potash : 0.4 gram was heated with an excess of potash with return-cooler two days without change. It can be heated in a sealed tube with a slight excess of alcoholic caustic potash at 125°-! 30° for several hours without undergoing change. Heated, however, at 150° with a little less than an equal weight of potassic hydrate dissolved in alcohol, it is saponified. The alcoholic solution, after being filtered from the potassic carbonate which had separated, was evaporated to dryness and the oil thus obtained was finally crystallized from dilute alcohol. This indol derivative melted at 11 2"-! 14°, and agreed in all its proper- ties with B3, Pr2-dimethylindol obtained by Raschen.' As a further means of identification a nitrogen determination was made : 0.1401 gram substance gave 12.2 cc. N at 20° and 750 mm. 'Ann. Chem. (Liebig) 339, 227. The Condensaiion-produds of Certain Hydrazides. 433 Calculated for C,oH,,N. Found. N 9,66 10.04 Ethylic orthotolyl-a-methylindol-^-carboxylate. — Acetacetic- ether orthotolylhydrazide was prepared in the usual manner and purified by crystallizing from ligroin. It melts at 95°-97°. It is readily soluble in all organic solvents. Like the corresponding para compound, it is more stable toward alkalies than the phenyl- hydrazide. It may be kept in a vacuum-desiccator for weeks without change. The hydrazide was converted into the indolcarboxylic ether by treating with strong sulphuric acid, pouring upon ice, etc., as in the case of the para compound. The yield of indol was a little less than that from the paratolylhydrazide, being about 22 per cent. After recrystallizing from dilute alcohol and drying in a desiccator it was analyzed. I. 0.2648 gram substance gave 0.6921 gram COa and 0.1744 gram H2O. II. 0.1985 gram substance gave 0.5227 gram COa and 0.1261 gram H2O. III. 0.2438 gram substance gave 14.0 cc. N at 20° and 749.8 mm. Theory for CsH.bNOj. I. Found. II. c 71.88 71.28 71.78 H N 6.91 6.45 7-30 7-05 6.57 From the analyses and the previous results with acetacetic- ether phenylhydrazide and acetacetic-ether orthotolylhydrazide, there can be no doubt that the orthotolylhydrazide, by treatment with concentrated sulphuric acid, has lost ammonia in a similar manner, giving Bi, Pr2-dimethylindol-/?-carboxylic ester, so that the conversion of the ortho-carboxylic indol into the correspond- ing dimethylindol derivative by saponifying and simultaneous splitting off of carbon dioxide was not considered necessary. Ethylic orthotolyl-a-methylindol-/3-carboxylate crystallizes from dilute alcohol in prisms belonging to the monoclinic system. It melts at 173° without decomposition (remelting at the same tem- perature). It is readily soluble in ether and alcohol. Indols from the Substituted Acetacetic Ethers. — It was next attempted to obtain indol derivatives from the substituted acet- acetic ethers by the method above employed. The alkyl acet- 434 Walker. acetic ethers experimented with were methyl-, ethyl- and benzyl- acetacetic ether. The methylacetacetic ether used was prepared according to the method of Conrad and Limpach.' It boiled at 195°-I97° under ordinary pressure with only slight decomposi- tion. It reacts with phenylhydrazine much less readily than any of the substituted acetacetic ethers experimented with, so that in preparing the hydrazide the whole of the methylacetacetic ether may be added at once to the phenylhydrazine in absolute ether. The mixture may be heated on the water-bath at 50°-6o° for several hours without danger of converting into pyrazolone, or it maybe allowed to stand twelve hours at the ordinary temperature,, when the reaction is complete. After drying the ethereal solution with calcic chloride and evaporating off the ether completely, the hydrazide was treated with concentrated sulphuric acid in the usual manner. This hydrazide may lose one molecule of ammonia in two different ways : (I) C.H4 (II) CeH HjCH. H I ^C:C(CH.)COOR = C6H<^^>C : C(CHs)COOR +NH. ,CH^ -^aH4( ■^C.CH(CH3)COOR, or CH.CCOOR .CH« — CCOOR yCCH3=C6H4 II +NHu NHi- NH -NH-CCH, The main product of the reaction, however, is neither of the above compounds, but is a pyrazolonesulphonic acid, as will be shown under the section devoted to pyrazolone derivatives. But reaction I. was also shown to take place at the same time, giving CH. a-Indolpropionic ether, C6H< >C— CH(CH3)COOR.— NH This compound was found in the ethereal extract of the products of the reaction. After washing the ether solution with alkali,, drying and evaporating off the ether, there remained an oil, part 'Ann. Chem. (Liebig) 192, 153. The Condensation-products of Certain Hydrazides, 435 of which was made to solidify. The soHd portion was separated from the oil by absorbing the latter in clay plates. After recrys- tallizing the solid substance from dilute alcohol and drying in a desiccator, it was analyzed. I. 0.1 120 gram substance gave 0.2926 gram COa and 0.0710 gram H2O. II. 0.1323 gram substance gave 0.3471 gram COs and 0.0867 gram H2O. III. 0.1930 gram substance gave ii.occ. N at 24° and 754 mm. Theory for CaHnNOj. I. Found. II. C 71.88 71-34 71-55 H 6.91 7.04 7.28 N 6.45 6.50 a-Indolpropionic ether, or methylindolacetic ether, separates from dilute alcohol in white crystals melting at 1 36°. It is readily soluble in organic solvents and gives the indol splinter-reaction. By saponification and splitting off of carbon dioxide it should yield a-ethylindol, but for lack of sufficient substance this experiment could not be made. Ethylic a-methylindol-^-carboxylate from eihylacetacetic-ether hydrazide. — When ethylacetacetic-ether phenylhydrazide is treated with strong sulphuric acid at — 12° a faint aromatic odor is observed.' The ethereal extract, obtained by the usual method, after being washed with alkali contained an indol derivative which incited at 131" and in all its properties agreed with the indol obtained from acetacetic-ether hydrazide by similar treatment. After crystallizing from dilute alcohol and drying in a desiccator it was analyzed. I. 0.1670 gram substance gave 0.4291 gram CO2 and 0.0990 gram H2O. II. 0.2030 gram substance gave 13.1 cc. N at 24° and 750 mm. Found. Theory for CjHjjNjO. I. II. C 70.90 70.77 H 6.40 6.57 N 6.90 ... 7.30 The formation of ethylic a-methylindol-/3-carboxylate from ethylacetacetic-ether hydrazide necessitates the splitting off of ethyl amine : 1 The ethylacetacetic ether used was obtained of Kahlbaum. 436 Walker. C^\lM\(^ OH. NH- CCOOR CCOOR ■^CCH3=: CeH.^^ ^CCH. + NH.C=H. jNH NH The mechanism of the reaction, therefore, between ethylacet- acetic-ether phenylhydrazide and sulphuric acid is quite different from that of the other indol reactions studied. 11,— Pyrazol Derivatives. CHs.C-C.H II II i-Phenyl-yjnethyl-e^-ethoxypyrazol, N \ / COC2H5. — The N CeHs fact that the condensation-product of acetacetic ether and phenyl- hydrazone by oxidation with mercuric oxide gives phenyl-/?- azocrotonic ether is sufficient proof that this condensation-product is a hydrazide and not a hydrazone.' As a further proof of the hydrazide constitution, Nef treated the hydrazide in ethereal solu- tion with an excess of acetyl chloride, and the oil thus obtained was supposed to be a diacetyl derivative of the hydrazide: CHs — C:=CH — COOR (CH3CO).N.N(COCH thus further proving the presence of two imido groups in the hydrazide. In view of the results obtained with oxalacetic-ether hydrazide," and the high percentage of carbon found by Nef in the supposed diacetyl derivative (two per cent, more than that required by theory),^ it seemed highly probable that acetyl chloride had acted on the hydrazide as a dehydrating agent, with the formation of an ethoxypyrazol. This supposition has been confirmed by experiment, but it will also presently appear that acetyl derivatives of the hydrazides are formed at the same time. Freer has more recently treated acetacetic-ether phenylhydrazide with acetyl chloride,^ but his results are wholly different from my own. The isolation of the ethoxypyrazol was at first attended with considerable difficulty ; I was formerly able to prove its existence only by means of its platinum double salt.' Various dehydrating ■Ann. Chem (Liebig) 260, 74. ^This Journal 14, 576. 'Ann. Chem. (Liebig) 366, 76. •• This Journal 14, 517. ^Ibid. 14, 585. The Condensation-products of Certain Hydrazides. 437 agents have been employed to split off water from acetacetic-ether hydrazide. Acetyl chloride has been found most effective in accomplishing this. Zinc chloride, phosphorus pentoxide and acid potassium sulphate have been tried with negative results. Dry hydrochloric-acid gas passed into a glacial acetic-acid solu- tion of the hydrazide (as in the case of oxalacetic-ether hydrazide)/ gave the desired result, but the yield is much better when acetyl chloride is used. After various modifications, the following method was found most satisfactory : To 4 or 5 grams of the hydrazide dissolved in 50-60 cc. absolute ether, well cooled by salt and ice ( — 15°), is slowly added an equal weight of freshly distilled acetyl chloride with constant shaking. The ether solution sometimes becomes turbid, at other times it remains perfectly clear; in either case the yield of ethoxypyrazol is the same. After standing a few minutes at ordinary tempera- ture, an oil separates at the bottom of the flask. The mixture is then heated a short time on the water-bath at so^-Go", and after- wards diluted with water. The ether solution, after washing thoroughly with acid and alkali, is dried with calcic chloride. Upon evaporating off the ether the ethoxypyrazol remains in the form of yellow needles. If in the above preparation, after the addition of acetyl chloride, the mixture is not heated but is poured on ice at once and subse- quently treated as described, the yield of oil is quantitative, — frequently it is more than quantitative, especially in the treatment of acetacetic-ether orthotolylhydrazide — a sufficient indication of the formation of acetyl derivatives. In this case the ethoxypyra- zol may be separated from the oil by one of two methods : (i) by distilling under diminished pressure, or (2) by allowing the oil to stand twelve hours with concentrated hydrochloric acid to decom- pose any unchanged hydrazide or acetyl derivatives that may be present. Of these two methods the latter is to be preferred. After treatment with hydrochloric acid the products are taken up with ether, washed with alkali, etc., as above. The average yield of the ethoxypyrazol was little less than one gram from four grams of the hydrazide; if two or three times this amount of hydrazide be used in an operation the yield is not increased. After recrystallizing from dilute alcohol and drying in a desiccator it was analyzed. 1 This Journal 14, 580. 438 Walker. I. O.I 541 gram substance gave 0.4007 gram COj and 0.0931 gram HsO. II. 02040 gram substance gave 0.5306 gram COj and 0.1230 gram HsO. III. 0.3173 gram substance gave 39.8 cc. N at 23° and 745 mm. IV. 0.2740 gram substance gave 34.0 cc. N at 20° and 745 mm. Theory for CaHi.NaO. I. Found. II. Ill c 71.28 70.91 70.93 H 6.93 6.71 6.69 N 13-86 I4.I 14.21 i-phenyl-3-methyl-5-ethoxypyrazol is insoluble in water, but readily soluble in all organic solvents. It crystallizes from petro- leum ether, at first in yellow six-sided prisms, finally in needles which are slightly yellow. Although this method of purifying is a slow one, the ease with which the pyrazol separates is quite remarkable; when the last trace of solvent has evaporated on standing, the whole of the ethoxypyrazol remains in crystalline form. In this way prisms fully two centimeters in length were obtained. It is most easily purified by crystallizing from dilute alcohol, from which it separates in white needles melting at 68*'-68.5°. In alcoholic solution with ferric chloride it gives no coloration. It shows Knorr's " pyrazolin-reaction," giving a bluish-green coloration, which was permanent after sixty hours standing. Its platinum double salt has already been described.' When heated for 2-3 hours with one-and-a-half times the calcu- lated amount of alcoholic caustic potash, phenylmethylethoxy- pyrazol is quantitatively saponified into i-Phenyl-ymethyl-e^-oxyPyrazol. — Unhke the ester from which it was derived, this body dissolves readily in alkali hydroxide or carbonate, from which solution it is reprecipitated by mineral acids. After crystallizing from dilute alcohol and drying at 110° it was analyzed. I. 0.1909 gram substance gave 0.4801 gram COs and 0.0973 gram H2O. II. 0.2246 gram substance gave 0.5662 gram COs and 0.1163 gram H2O. III. 0.1322 gram substance gave 19.0 cc. N at 27° and 745 mm. •This Journal 14,585. The Condensation-products of Certain Hydrazides. 439 Found. Theory for C,oH,oN,0. I. 11. c 68.96 68.59 68.75 H 5-75 5.66 5.78 N 16.09 ... ... 15.96 Phenylmethyloxypyrazol (i, 3, 5) is readily soluble in-all organic solvents : ether, alcohol, benzene, ligroin, chloroform and acetic ether were tried. It separates from benzene solution at first in flat plates, later in needles. It is best purified by crystal- lizing from dilute alcohol, from which it separates in orange- yellow needles, melting at 196°-! 98° with decomposition. An alcoholic solution with ferric chloride gave no coloration. It gives Knorr's " pyrazolin-reaction," giving a permanent green coloration. Its salts are far more stable than those of the oxy- pyrazol obtained from oxalacetic-ether hydrazide.* The aqueous solution of the alkali salts can be heated for hours on the water- bath without undergoing change, the oxypyrazol being precipi- tated upon the addition of an acid. It forms calcium and barium salts which are quite readily soluble in water, but unfortunately they were lost by the collapse of a vacuum-desiccator while concentrating their aqueous solution. A silver salt, however, was made from the neutral ammonium salt in the usual manner, and after thoroughly drying, a determination of silver in the salt was made : Theory. Found. Ag 38.43 38.02 III. — Pyrazolone Derivatives. It has already been stated that the chief products of the action of concentrated sulphuric acid on methyl- and benzylacetacetic- ether hydrazide are pyrazolonesulphonic-acid derivatives. \- Phenyl- ■^.\-dimethyl-^-pyrazolone-2. f-sulphonic acid* CH3.C = C.CH. I I V # . — Methylacetacetic-ether phenylhydrazide was N treated with strong sulphuric acid at — 15°. Upon pouring the 1 This Journal 14, 583. 'In these sulphonic-acid formulse the sulphonic group is united to nitrogen; no experi- mental evidence, however, of the position of the sulphonic group is thus far known. 440 Walker. sulphuric-acid solution on ice, an odor very much like that of methylketol was observed ; in the second preparation this odor was not perceptible until after the sulphuric-acid solution had stood a few hours. The solid product of the reaction (which proved to be the pyrazolonesulphonic acid) was filtered off and the filtrate extracted with ether as described on page 434. The relative yield of the three products of the action of sulphuric acid on methylacetacetic-ether hydrazide was: 12.0 grams of the hydrazide gave 7.0 grams crude sulphonic-acid derivative, i.o gram indol derivative, 2.1 grams oil of unknown composition. On treating the hydrazide with sulphuric acid at higher tempera- tures than — 15° the yield of the pyrazolonesulphonic acid is diminished. The brownish-colored sulphonic acid at first possessed a strong indol odor, due to adhering indol. Upon crystallizing from alcohol (95-per cent.) a few times, the odor of indol entirely disappeared. After drying in a vacuum-desiccator it was analyzed : I. 0.1265 gram substance gave 0.2270 gram CO2 and 0.0507 gram H2O. II. 0.1389 gram substance gave 0.2481 gram COs and 0.0551 gram H2O. III. 0.2190 gram substance gave 18.9 cc. N at 24° and 754.1 mm. IV. 0.1368 gram substance gave 0.1141 gram BaSO*. Found. Theory for CHijNjSO^. I. II. III. IV. C 49.25 48.94 48.71 H 4.47 4.45 4.40 N 10.44 ... ... 987 S 11-94 ••• ••• ••• 11.46 Phenyldimethylpyrazolonesulphonicacid separates from boiling water at first in gray, finally in perfectly white, needles. It is best purified from 95-per cent, alcohol, from which it separates in apparently short white needles. (The adhering indol should first be removed from the crude product by treating with a small amount of alcohol, in which the indol is much more readily soluble than the sulphonic acid.) It does not melt at 300° and is difficultly soluble in all solvents. It gives Knorr's " pyrazolin-reaction." \-Phenyl-ymethyl-/^-benzyl-^-pyrazolone-2 ?-stilphonic acid. — Benzylacetacetic-ether hydrazide was treated with strong sul- phuric acid at about — 15". Upon pouring the mixture on ice a faint indol-like odor was observed, and a voluminous separation The Condensation-products of Certain Hydrazides. 441 of a brownish-colored solid took place. This was filtered off and the filtrate extracted with ether. The washing of the ether solu- tion with dilute sodic hydrate removed nearly the whole of the coloring matter. After drying the ether solution with calcic chloride and evaporating off the ether there was left a very small amount of oil ; from 14 grams of the hydrazide only 3-4 drops of oil were obtained. This oil possessed a strong indol odor and gave the indol splinter-reaction. The quantity obtained did not permit of further examination. By carrying out the reaction at temperatures above — 15° the yield of this oil is increased. The crude sulphonic acid is best purified by treating with a small volume of alcohol to dissolve adhering indol and then crys- tallizing from 95-per cent, alcohol. After repeated crystallizations and drying at 110°, it was analyzed. I. 0.1206 gram substance gave 0.2607 gram CO 2 and 0.0520 gram H2O. II. o.iiii gram substance gave 0.2415 gram CO2 and 0.0476 gram HsO. III. 0.3309 gram substance gave 0.7183 gram COa and 0.1455 gram H2O. IV. 0.2683 gram substance gave 18.7 cc. N at 20° and 742 mm. V. 0.1755 gram substance gave 12.7 cc. N. at 19° and 742 mm. VI. 0.15 10 gram substance gave 0.1040 gram BaSOi. VII. 0.2104 gram substance gave 0.1401 gram BaS04. Theory for C„H„N,SO,. I. II. Found. III. IV. c 5930 58.95 59.28 59.20 ... H 4-65 4.78 4.70 4.88 ... N 8.II ... 8.00 S 9-33 ... ... ... 8.29 ••• 945 9-57 Analyses II and III were made with a mixture of lead chromate and potassium dichromate. I -Phenyl-3-methyl-4-benzyl-5-pyrazolone-2-sulphonicacid crys- tallizes from alcohol in exceedingly light, white flakes. It does not melt at 300°. It is much more easily soluble in hot benzene and chloroform than in alcohol. It is more difficultly soluble in water than the sulphonic acid obtained from methylacetacetic ether hydrazide. It forms stable salts which could not be made to crys- tallize. It gives Knorr's " pyrazolin-reaction." Attempts to split off the sulphonic group by heating with con- centrated hydrochloric acid in sealed tubes were unsuccessful, part 442 Walker. of the substance being unchanged and the remaining portion undergoing total decomposition. Attempts to split off the sulphonic group by heating for some time on the water-bath with moderately strong sulphuric acid also failed. Although all attempts to convert the sulphonic acid into the known phenylmethylbenzylpyrazolone were unsuccessful, that the body is a sulphonic derivative of phenylmethylbenzylpyrazo- lone was proved by direct synthesis : Phenylmethylbenzylpyrazo- lone was treated with concentrated sulphuric acid at — 15° to — 20° and the solid products of the reaction washed on clay plates with ether. On crystallizing from alcohol, two substances were obtained, one melting at 210° (perhaps an inden-derivative not further inves- tigated), and a difficultly soluble substance not melting at 300°. The latter was identical with the sulphonic acid obtained from benzylacetacetic-ether hydrazide. i-Phenyl-^-'niethyl-^-benzyl-^-pyrazolone. — This was obtained by the spontaneous decomposition of benzylacetacetic-ether hydrazide at the ordinary temperature. All unchanged hydra- zide was removed by washing with ether, and the pyrazolone further purified by crystallizing from dilute alcohol. After drying in a desiccator it was analyzed : I. 0.1277 gram substance gave 0.3628 gram COa and 0.0731 gram H2O. II. 0.1345 gram substance gave 13.0 cc. N at 24° and 750 mm. Theory for C,,H,eN,0. I. Found. II c H 77.27 6.06 7749 6.44 •• N 10.60 ... II. C Phenylmethylbenzylpyrazolone (i, 3, 4, 5) separates from alcohol in glistening scales which melt at 136". It is insoluble in ether, slightly soluble in boihng water, readily soluble in alkali, chloroform, acetone and acetic ether, less readily in benzene and ligroin. i-Orihoiolyl-Tf-meihyl-$-pyrazolo7ie. — This compound was pre- pared from acetacetic-ether orthotolylhydrazide and purified by the same method as the benzylpyrazolone. It separates from alcohol at first in four-sided prisms, later in white needles which melt at I43°-I44*'. It is insoluble in ether, water, ligroin and acetic ether, but slightly soluble in chloroform. Chbmical Laboratory, Clark Univbrsity, June, 1893. Aciio7i of Chemical Compounds upon Animals. 443 A SYSTEAJATIC STUDY OF THE ACTION OF DEFI- NITELY RELATED CHEMICAL COMPOUNDS UPON ANIMALS. By Wolcott Gibbs, M. D., Rumford Professor {Emeritus) in Harvard University, and Edward T. Reichert, M. I)., Professor of Physiology in the University of Pennsylvania. \Continued/rom p. 370, Vol. 13.] Our study of the physiological actions of definitely related chemical compounds has been brought to a conclusion by want of funds, a large appropriation from the Bache fund and a smaller one from the Smithsonian Institution having been exhausted. The very extended programme of work must therefore be abandoned, though with the earnest hope that investigators will take up the subject from the same point of view, and that in future the study of isolated compounds will assume a subordinate place. It remains for us to discuss the results which have been obtained, and to draw from them such conclusions as may be warranted under the circumstances. These conclusions are not always positive on account of the limited number of cases to be considered. On the other hand, we hope to be able to show that negative conclusions may have a certain value. Phenol. In discussing compounds belonging to the aromatic series, it will be most convenient to start with phenol, CeHfOH, which has been studied with great care as regards its physiological actions. We cannot trace the changes in its actions produced by passing from benzol, C^Hs, to phenol, by the replacement of an atom of hydrogen by a molecule of hydroxyl OH, but we have materials for discussing successive replacements in phenol as in the di- and triphenol derivatives, and we shall consider these first in order. As phenol forms our term of comparison it will be well in this place to recapitulate the results which have been obtained by dif- ferent investigators as to its physiological relations. It will be sufficient to state that it appears first to stimulate and then to depress the spinal centers. In sufficient quantity the drug de- presses the heart and the vaso-motor centers, and thus reduces arterial pressure. Phenol in the first stages of its action increases Vol. XVI.-33. 444 Gibbs and ReicherU the frequency of respiration in a remarkable degree, partly from stimulation of the peripheral pulmonic vagi and partly from that of the respiratory centers. Finally, when injected into rabbits phenol produces a very distinct fall in the bodily temperature, usually but not always coincident with the lowering of the arterial pressure. The formula of phenol is CsHsOH. By the replacement of a second atom of hydrogen we have C6H<(OH)j, and under this general formula we have ortho, meta and para compounds which we may distinguish in the usual manner by writing the formulae <7-OH4(OH)=, w-C6H4COH)3, ;>-C6H4(OH)2. The results of our study of these are as follows : Ortho- dihydric phenol, or pyrocaiechin, ^-C6H«(OH)2. — The action of this compound upon the system is, as we have shown, very much more pov*erful than that of phenol and differs from it by being exerted more energetically on the spinal centers. Con- vulsions are also produced by the action of phenol, but they are very much less marked, and it seems safe to say that the differ- ence in the action of phenol and ortho-dihydric phenol is chiefly , one of degree, and that the replacement of a second atom of hydrogen by a second molecule of hydroxyl results in the pro- duction of a more active agent of the same general character. Mdia-dihydric phenol, or resorcin, w-C6H4(OH)j. — Making due allowance for its less powerful action, resorcin affects the system in a manner very similar to pyrocatechin. The same is true of hydroquinone,/'-C6H4(OH)2, which as regards intensity of action comes next to pyrocatechin and stands above resorcin. In the cases of the two trihydric phenols studied, namely, ^ro- gallol{\, 2, 3) dindphloroghicin (1,3, 5), it appears that the charac- ters of the actions upon the system do not essentially differ from those of phenol or the three dihydric phenols already cited. It is important to note that the intensity of the action of pyrogallol, I, 2, 3-C6H3(OH)3 is less than that of pyrocatechin or the ortho- dihydric phenol, and agrees very closely with that of hydro- quinone when measured by experiments on dogs. The action of pyrogallol is very much more powerful than that of phloroglucin. This may be connected with the fact that pyrogallol possesses a very high reducing power or, better, a very strong affinity for oxygen, and in the same connection we may note that pyro- catechin, like pyrogallol, reduces cold solutions of silver, while Action of Chemical Compounds upon Animals. 445 the (physiologically) less active resorcin reduces silver salts only on boiling. The third trihydric phenol, or oxyhydroquinone (i, 2, 4) should also be studied. We have not been able to procure it for our work. The three tetroxybenzols, C6H2(OH)4, (i, 2, 4, 5 ; 1,3,4, 5; and 1,2,3,4), ^s well as the hexoxybenzol, C6(OH)6, also deserve study, as does the large group of phenols of other types. With respect to the function of the molecule OH in these com- pounds it may be remarked that solutions of hydroxyl according to recent experiments are not toxic when taken into the stomach. We have, further, abundant evidence to show that it is not toxic in inorganic acids, as, for example, in sulphuric acid, S02(OH)3, nitric acid, NO2OH, metaphosphoric acid, PO2OH, phosphoric acid, PO(OH)s, etc. On the other hand Brunton, Bohenham, and others have shown that hydroxylamine, NHjOH, exerts a very powerful action on the animal system, but this result, as we shall show, may be explained without any hypothesis as to the presence of hydroxyl. Cresols. These compounds may be regarded chemically as phenol in which an atom of hydrogen is replaced by a molecule of methyl, the general formula for ortho, meta and para compounds being C6H4(CH3)OH. As already stated all three act as sensory and motor paralysants. In the cases of the three isomers of this series, ortho, meta and para, it appears that there is a marked difference between the action of the ortho and para compound on the one hand and that of the meta compound on the other, the ortho and para compounds acting as stimulants to the inhibitory apparatus of the heart, while the meta compound has no such influence. Considered as cardiac depressants the order of intensity of action is ortho, para, meta, the last being least active. The same is true for the action on the nervous system of the frog. On the other hand metacresol acts more powerfully than the other two upon the vaso-motor system. From this it appears that in speaking of the order of intensity it will always be necessary to state the mode of action, since the order may vary with this and is, therefore, not absolute. It further appears that the replacement of an atom of hydrogen in phenol by a molecule of methyl changes the mode of action on the system, which is not the case, as we have seen, in replacements by hydroxyl, within the limits, at least, to which we have confined ourselves. 446 Gibbs and Reichert. It is of importance to study these replacements from various points of view. Thus in cresol an atom of hydrogen within the phenyl molecule CeHs is replaced by a molecule of methyl, CHj. But in anisol, CeHs.OCHs, the replacement of hydrogen by methyl takes place in the molecule of hydroxyl. Similar replacements occur in all, or nearly all, the di- and trihydroxyl derivatives of benzol. The fact that in cresols and all the higher phenols the hydroxyl OH prevents the oxidation of the alkyl group — for instance, CHs — by chromic-acid mixture is worth noting here. Thus chromic acid does not oxidize CeHiCHsOH. But when the hydrogen in the hydroxyl group is also replaced by an alkyl, the group which replaces hydrogen in the phenyl molecule is oxidized and ethers of the oxy-acids are formed. Thus C6H4.CH3.OCHt yields CeHj ] r^QoH* ^' would therefore seem probable that the compound CeH^.CHs.OCHa would act upon the system in a very different manner from the compound CeH^.CHs.OH, or cresol, if the physiological action depends upon facility of oxidation. Materials for deciding this question are not at present available. Since it appears at lejist probable that the intensity of the action of the cresols is greater than that of phenol, it becomes a question whether the successive replacement of hydrogen in phenol by an alkyl, as for instance methyl, increases toxicity. Di-, tri-, tetra- and pentamethylphenols have been obtained, but their physiological relations have not yet been studied. Of the six possible isomerides with the general formula C6H. Tetrahydric phenols, C6H.(OH>. Pentahydric phenols, C6H(OH>. Hexahydric phenols, a(OH)6. Cresols, CeH.CH^OH. Phenol. CeHsOH Dimelhylphenol, C6H,(CH3)20H. Trimethylphenol, C6H2(CH3)sOH. Tetramethylphenol, C6H(CH3>OH4. Pentamethylphenol, C6(CH3>OH. Amidobenzol, CeHsNHs. Diamidobenzol, C6H4(NH.)2. Phenol. CeHsOH Triamidobenzol, CeH.CNHOa. Tetramidfibenzol, C6H...(NH04. Pentamldobenzol, C6H(NH2)5. Monochlorphenol, C.H4CIOH. Dichlorphcnol, CeHsCLOH. Trichlorphenol, C6H2CI3OH. Phenol. CeHaOH Tetrachlorphenols, C6HCI4OH. Pent achlorph ends, CeCUOH. Anisol, CeHaOCH.. Monometh oxyphenol, C6H4OCH3OH. Dimethcx\ benzols, C6H4(OCH3>. The chinons, oxychinons and hydrochinons deserve attention in this connection. Tables like those which we have given will be found extremely convenient in such investigations as we have undertaken. They enable us to take in at a glance the compounds whose physiolog- ical actions should be compared by placing two or more tables side by side and drawing connecting lines between corresponding terms, as for instance, trichlorphenol, CeHsChOH, and trimethyl- phenol, C6H.H. The physiological actions of these two do not, as we have shown, differ essentially, except in intensity, and are apparently less pow- erful than those of the corresponding phenol compounds. The relatively large proportion of the NO2 in the benzol compounds would lead us to expect a more powerful action on the part of the two nitrobenzols. Nitranilines, C6H4(N03)NH9 {p-, m- and/)-). These three compounds all stimulate the peripheral ends of the vagi in the heart and produce methaemoglobin or otherwise alter the haemoglobin in the blood. The nervous effects are feeble and depend apparently upon the changes in the blood. The para compound is most active ; the ortho compound comes next in order, while the meta compound is least energetic. The action of aniline is exerted mainly upon the nervous system, resembling in some respects that of ammonia, but differing particularly in the fact that true tetanus is never produced.' The introduction of a molecule of NOa completely changes the nature of the action and concentrates its action upon the circulatory system, the nerv- ous phenomena being greatly reduced. ' Brunton-Croonian Lectures. The Nitrites of some Amines. 449 Toluidiries. These maybe regarded chemically as anilines in which an atom of hydrogen in the phenyl is replaced by a molecule of methyl. They differ from methylaniline in the fact that in these last the replacement is in the NH-j molecule. We have shown that the order of intensity of action in the case of the three toluidines is lowest with the ortho and highest with the para compound. All three break down the blood and produce methaemoglobin. All kill by failure of respiration. All lower temperatures, both normal and abnormal, and all decrease reflex activity by acting on the spinal cord. On comparing the actions of the toluidines, it appears that while there is a distinct difference between the intensity of the actions upon the system of the ortho, meta and para compounds, there is no appreciable difference in the nature of the actions. Since toluidines are embraced under the general formula CsH* CHsNHi, they admit of direct comparison with aniline of which the formula is C6H5NH2. We note then the important fact that the replacement of an atom of hydrogen by a molecule of methyl has the general effect of diminishing the actions upon the nervous system and of increas- ing them upon the vascular system. (To be continued^ CuntributioDS from the Chemical Laboratory of the Rose Polytechnic Institute. X.— THE NITRITES OF SOME AMINES. By W. a. Noyes and H. H. Ballard. In a previous article by one of us' the statement is made that the nitrite of 1.4-diaminocyclohexane is decomposed on boiling its solution in water, and that the decomposition and its products would be studied. The results of this study are recorded below. The diamine was prepared according to the directions given in an article by v. Baeyer and Noyes,^ and by a modification of this 1 This Journal 15, 545. * Ber. d. chem. Ges. 88, 2168. 450 Noyes and Ballard, suggested by an article of v. Baeyer's," the details being furnished us privately by the author, though they have since been pub- lished.' As finally adopted the method consisted of the following steps: From succinic diethyl ester to succinosuccinic ester; from this by saponification with dilute sulphuric acid to 1.4-diketocyclo- hexane; from this by treatment with hydroxylamine to the dioxime, and finally by reduction with sodium in absolute alco- hol to the diamine. The base was distilled from the alkaline solution with steam, the distillate being collected in a vessel con- taining hydrochloric acid. The solution thus formed was evapo- rated to dryness on the water-bath to remove the excess of acid. Then just enough water was added to make a hot saturated solu- tion, to which alcohol (four or five volumes) was added and the mass allowed to cool. By this means most of the chloride was precipitated in pure condition. It was then filtered and washed with alcohol. A determination of chlorine gave the following: 0.1 1 27 gram gave 0.1722 gram AgCl. Calculated for CeHieNaClj. Found. CI 37.97 37-88 In the reduction of the oxime it was found best to use small quantities (2-3 grams) of the substance, to boil the solution before adding sodium, and to keep enough alcohol in the flask to retain all solid matter in solution. Without this precaution a precipitate was formed which interfered with the reduction. Decomposition of the Nitrite. — On adding sodium nitrite to a solution of 1.4-diaminocyclohexane chloride there is almost imme- diately an evolution of gas which at ordinary temperatures is very slow, but becomes quite marked on warming the solution, and before the temperature of boiling water is reached becomes very vigorous. At the same time there is perceived an odor which has been shown to be that of dihydrobenzene. For the purpose of observing the rate of evolution and the amount of gas evolved, the solution to be tested, in this as well as in all other experiments described, was placed in a flask as small as could be conveniently used, and this was connected with a nitrometer. The flask was heated sometimes in a steam -bath or a vessel of boiling water and sometimes over a free flame. The 'Ann. Chem. (Liebig) 378, 91. "Ber. d. chem. Ges. 35, 1037. The Nitrites of some Amines. 451 latter method was found more convenient when using larger quantities, and the liquid which distilled was returned by means of a small inverted condenser. In every instance at the end of half an hour about one-fourth of the total nitrogen had been evolved and the rate of evolution had become very slow. The curve plotted from readings made at intervals of one minute appears to be of the nature of a para- bola, the ordinates representing volumes, the abscissae time. At the end of eight hours heating usually 35-40 per cent, of the total nitrogen had been evolved. In many experiments, especially where larger quantities were used, a portion of the diamine was recovered. In two instances careful account was taken of the various quantities, and the follow- ing figures deduced : Chloride of diamine taken, I. 6.45 grams. II. 1. 00 gram. " " recovered, 1.89 .39 Wt. of nitrogen obtained^ .70 .089 Percentage of nitrogen obtained, 10.85 8.88 Percentage of nitrogen equivalent to one atom of nitrogen in diam- ine chloride. 14.96 14.96 Percentage of theoretical yield, 72.5 59-4 Percentage of diamine chloride decomposed, 70.7 61.0 If in the reaction only one amine group is concerned the last two figures should be the same. The fact that small quanti- ties of dihydrobenzene were formed in every decomposition accounts for the somewhat high percentage of nitrogen in I. In another series of experiments a solution of the nitrite of the base was prepared by adding silver nitrite in calculated quantity to the chloride of the base in solution, and shaking the mixture for several hours. At the end of this time the solution was filtered and heated as above described. There was no apparent difference in either the rate of decomposition or the products of the reaction. In a third series, having previously shown that the decompo- sition concerned principally only one amine group, the chloride of the base and sodium nitrite were mixed in the proportion of one molecule of each and heated, in which case the decomposition took place practically as it had in each of the others. The isola- 452 Noyes and Ballard. tion of the products of decomposition was accomplished in two ways, both !?iving the same results : First, the liquid which had been heated was rendered strongly- alkaline with caustic soda, saturated with salt and then extracted several times with ether. To the extract was added an excess of hydrochloric acid and the whole evaporated to dryness. The residue was dissolved in as little water as possible and a solution of chlorplatinic acid added, whereby was caused the almost com- plete precipitation of the undecomposed diamine, the chloro- platinates of the other bases present remaining in the solution, which after filtration was evaporated to dryness in vacuo over sulphuric acid. From the residue by crystallization from strong alcohol were obtained the chloroplatinates of the tetrahydroaniline and 1.4-aminohydroxycyclohexane. By extracting the solution after heating to decompose the nitrite and before adding the alkali, a small quantity of dihydrobenzene could always be obtained and recognized by its odor. The second method of separation was as follows : The solution after decomposition of the nitrite was made strongly alkaline with caustic soda, and the bases thus set free distilled with steam into a receiver containing hydrochloric acid. The solution of chlorides thus obtained was evaporated to dryness on the water-bath and the residue extracted with strong alcohol, whereby an almost complete separation of the undecomposed base was effected, the chloride being left undissolved and in a practically pure condition. The alcoholic extract, evaporated to dryness, left a mixture of the chlorides of tetrahydroaniline and 1.4-aminohydroxycyclohexane, which were separated as chloroplatinates, as before. In the products of several decompositions unsuccessful efforts were made to detect the presence of chinite.' Dihydrobenzene was detected by its odor and the color-reaction with concentrated sulphuric acid in the presence of alcohol.' The amounts obtained were, however, so small that none of the other chemical properties could be identified. In one of the experiments in which a larger quan- tity of diamine chloride was used, the gases, before being collected for measurement, were passed through absolute alcohol cooled by ice. The solution thus obtained smelled very strongly of dihydro- benzene and gave the color very readily on the addition of sul- phuric acid. By observing the liquid condensed in the inverted 'Ann. Chem. (Liebig) 878, 92. » Ibid. 378, 95. The Nitrites of some Amines. 453 condenser it could be seen, however, that the amount of oily matter was very small. It seems, therefore, that in the decomposition of the nitrite of 1.4-diaminocyclohexane one amine group is principally concerned, and that the main products are salts of tetrahydroaniline and 1.4-aminohydroxycyclohexane. It will be shown later that the nitrite of the former base also decomposes on heating. A^- Tetrahydroaniline Chloroplatinate. — The salt is easily soluble in water, very difficultly soluble in hot or cold absolute alcohol, easily soluble in hot alcohol containing a little water. From this solution, on cooling, scaly crystals separate, which when dry have a pale, dull yellow color, and melt with decompo- sition at 210°, beginning to blacken at about 205." Determinations of platinum resulted as follows: I. 0.041 1 gram gave 0.0132 gram Pt. II. 0.0439 gram gave 0.01415 gram Pt. Calculated for Found. {C,H„N),PtCl,. I. II. 32.28 32.12 32.23 \.a^-Aminohydroxycyclohexane Chloroplatinate. — This sub- stance also is soluble in water, insoluble in absolute alcohol, but quite soluble in cold, and much more easily soluble in boiling strong alcohol. When mixed with the tetrahydroaniline chloro- platinate, the two may easily be separated by crystallization from hot alcohol. The crystals from alcoholic solution are flat, oblong plates, regular in form and often in cross-like pairs. When dry the mass has a satiny luster and a bright canary-yellow color, darker than that of the tetrahydroaniline chloroplatinate. Determinations of platinum gave the following : I. 0.0335 gram gave 0.0104 gram Pt. II. 0.0960 gram gave 0.0295 gram Pt. Calculated for (CeH,o Ber. d. chem. Ges. 21, 854. ^ Ibid. 21,iii6. The Nitrites of some Amines. 455 /9- Tetrahydronaphthylamine Nitrite {alicyclic). — The conduct of the nitrite of id'-tetrahydroaniline rendered it desirable to test quantitatively the conduct of the nitrite of /5-tetrahydronaphthyl- amine. To our surprise it was found to decompose on heating its solution to 100°, and its conduct in this regard is almost exactly like that of ^'-tetrahydroaniline under the same conditions. This result is so different from the conduct which the statement of Bam- berger and Lodter ' led us to expect that we are compelled to believe that they did not have a specimen of the pure nitrite, but that the substance with which they experimented (obtained by passing nitrous anhydride into an ethereal solution of the base) consisted chiefly of the nitrate. The analysis reported gives an amount of nitrogen which agrees much better with that required for the nitrate than for the nitrite, when it is noticed that all of the other nitrogen determinations reported in the article give high results, as is usual. Also we find that the nitrite melts with decomposi- tion at 136°, while the nitrate, which we have prepared, melts at 215°. But a mixture of about 3 parts nitrate and one part nitrite melts at about 160° with the phenomena described by Bamberger and Lodter. We prepared the nitrite by precipitation from a strong solution of the chloride by means of a solution of sodium nitrite. The chloride melted at 242°-243°, instead of at 237° as given by Bamberger and Lodter. The nitrite is quite soluble in water and in cold solution shows no immediate evidence of decomposition, but on being heated to the temperature of steam quite an energetic evolution of nitrogen takes place. Using 0.2314 gram nitrite — in the first 18 minutes 6.4 cc. gas was evolved = 21 per cent. " " 32 " 8.2 " " =27 " in 174 " 14.3 " " =47 " so that in the course of three hours nearly 50 per cent, of the nitrite had been decomposed. A colorimetric determination of the nitrous acid in the salt used in the experiments showed the purity of the compound. In a second experiment, to 0.1672 gram of the chloride were added 0.0005 gram of caustic soda and 0.055 gram of sodium nitrite, and the mixture heated as before in steam. There was a 456 Noyes and Ballard. slow evolution of gas which corresponded in rate almost exactly with that of the -^i^-tetrahydroaniline. At the end of seven and one-half hours about two-fifths of the total nitrogen had been collected. In a third experiment sodium nitrite was added directly to a solution of the chloride and the liquid tested to be sure that it was not acid. On heating, the evolution of gas was decidedly more rapid than in the second experiment, but approximately the same as in the decomposition of the nitrite of the base alone. It seems therefore that the nitrite of /5-tetrahydronaphthylamine in water solution, either neutral or very faintly alkaline, decom- poses at the temperature of steam. 2.^-Diaminohexane Nitrite. — To compare the relative stability of a chain compound as nearly as possible analogous to 1.4- diaminocyclohexane nitrite, 2.5-diaminohexane chloride was prepared ' and heated in the presence of an excess of sodium nitrite to the temperature of steam, and the evolved gas collected. At the end of 5 minutes about one-third the total amount had been given off; in 15 minutes about one-half; and at the end of II hours an amount of nitrogen equal to 22.7 per cent, by weight of the chloride used. 22.14 P^r cent, is equivalent to 3 atoms of nitrogen. It appears therefore that the substance is much less stable than the corresponding ring compound, 1.4-diaminocyclo- hexane nitrite. Conclusion. The fact which gave rise to this investigation was that in the decomposition of 1.4-diaminocyclohexane nitrite only one of the four atoms of nitrogen seemed to be given off, and consequently there was a possibility of a condensation within the molecule. This idea has been shown to be incorrect. After one-fourth of the nitrogen has been given off the decomposition becomes very slow, but continues. Furthermore, by an examination of the pro- ducts, the decomposition has been shown to be normal and to concern principally one amine group, the splitting off of which and the establishing of the double union seem to render the resulting compound more stable, but at the same time not per- fectly so. The similarity in the behavior of the nitrites of J*-tetrahydro- aniline and /S-tetrahydronaphthylamine is a slight indication that 1 Ber. d. chera. Ges. 33, 168, aioo. Constituents of the Nodes and Iniernodes of Sugar Cane. 457 the two compounds have a similar structure. We do not attach much importance to analogies of this kind, and especially we do not consider that any valuable conclusion as to the structure of benzene can be drawn from them. In any case, however, the analogies do not favor the centric formula for benzene. With that formula there is no double union in /?-tetrahydronaphthyl- amine where its analogy with the ^'-tetrahydroaniline would lead us to expect one. April 25, 1894. A STUDY OF THE CONSTITUENTS OF THE NODES AND INTERNODES OF THE SUGAR CANE. By J. L. Beeson. It is a belief among sugar-planters that the nodes of the cane stalk (the joints) are the main source of the reducing sugars contained in the expressed juice. In order to throw light upon this and other questions, a study was made this season of the constituents of the nodes and internodes of the cane as follows : Twenty stalks were selected from the field, of as near the same size and number of joints as possible. The nodes were separated from that portion lying between the nodes — the internodes — with a fine saw. These separ- ate portions were each passed twice through hand-roller mills, an effort being made to give as nearly as possible the same pressure and relative extraction of juice from the nodes as the internodes. The two juices thus obtained were markedly different in appearance and constituents. That from the nodes was in every case highly colored, gave a heavy precipitate with a solution of subacetate of lead, and a heavy coagulum on heating to the boiling-point of the juice ; while that from the internodes was clear, light in color, gave only a small precipitate with subacetate of lead solution, and no appreciable coagulum on heating. This difference in coagulable bodies indicates a wide difference in the water-soluble albuminoids in the two juices. In the two following tables are given the analyses of the juices : 458 Beeson. A7ialysis I. — Twenty stalks, purple variety, plant cane. Aver- age weight of stalk, 3 pounds. Total Reducing Solids not Solids. Sugars. Sucrose. Sugars. Topsj^"'^^^ J5-5 0.66 12.7 2.64 I Internodes 16.8 1.20 15.0 1.60 Middles i^"^^" "^•^ °-'° '3-5 2.90 Internodes 17.6 i.oo 15.6 i.oo Nodes 14.2 0.26 1 1.9 2.04 Internodes 17.2 0.89 15. i 1.21 Analysis II. — Twenty stalks, purple variety. Average weight of stalk, 2) pounds. 'Tops I Total Reducing Solids not Solids. Sugars. Sucrose. Sugars. Fiber. Nodes 15.3 0.18 II. 3 3.82 15.6 Internodes 169 1.25 14.3 1.35 8.6 Middles! ^"'^^^ '^'7 0.07 13.7 307 iS-9 I Intel Battsj^"^^^ 15-7 0.15 12.8 2.75 18.2 Internodes 17.7 0.98 16.0 0.72 8.0 Nodes 15.7 0.15 12.8 2.75 18. Internodes 17.7 0.61 16.4 0.69 8. It will be seen that there was found more glucose or reducing sugars in the tops of the stalk than in the butts. This is to be expected, since the tops are less mature. Also that there is approximately twice the percentage of fiber (that portion of the stalk which is insoluble in water) in the nodes as in the contiguous internodes, the butt nodes showing the highest percentage. But the most striking fact is the marked decrease in reducing sugars and a like increase in solids not sugars (which are mainly albu- minoids, gums and mucilages) in the nodes over that of the con- tiguous internodes. This wide difference in content of reducing sugars in the different portions of the stalk is surprising, in view of the readily diffusible nature of these bodies. The cane from which the samples were taken had been blown down, and it was noticed that the eyes had very slightly sprouted, yet their rootlets had not taken hold in the soil. It was thought possible therefore that this low percentage of reducing sugars in the node or joint was due to this fact : that the incipient growth of the eye was at the expense of the glucose in that region, using it up more rapidly than it could difiuse from the adjoining internode — the internode containing the bulk of the juice. This hypothesis, of course, assumes glucose — and not starch, as is maintained by Sachs and others — to be the first visible product of plant-assimilation. Since Constituents of the Nodes and Internodes of Sugar Cane. 459 pure starch is nowhere found in the cane,' it is difficult to see how this body could in this case be a product of plant-assimilation. It would appear more reasonable that a sugar, in the case of cane, supplies the food for the first stages of growth. If the above hypothesis be correct, namely, that the decrease in reducing sugars in the node is due to the incipient growth of the eye, then canes bearing normal eyes should not show this difference, and, on the other hand, canes bearing sprouts, yet whose rootlets had not taken hold in the soil, should show a dif- ference in content of reducing sugars similar to that in cases I and II, with a loiver percetiiagc in the internode, unless some of the sucrose should be inverted by the growing sprouts as a further supply of glucose for food. In order to throw light on this point the following analyses were made: (i) of canes having normal eyes, (2) of canes having large or slightly sprouted eyes, (3) of canes bearing sprouts. Analysis III. Normal eyes, 4 stalks, striped variety, plant ca7ie. Average weight of stalk, 3 pounds. Reducing Sugars. Sucrose. Nodes of whole Stalk 0.40 12.0 Internodes of whole stalk 1.23 15.5 Analysis IV. Large eyes — sample and weight same as in above. Reducing Sugars. Sucrose. Nodes of whole stalk 0.50 12.8 Internodes of whole stalk i.ii 13.4 Analysis V. Sprouted eyes. — Same variety as in above. Upper halves and upper thirds from 4 stalks, sprouts from two to six inches long. Reducing Sugars. Sucrose. Nodes of upper halves and thirds 0.3 12.00 Internodes of upper halves and thirds 0.9 16.00 These canes were selected from the same plat, of as near the same size and number of joints as possible, were analyzed the same afternoon, and as soon as possible after expressing the juice. From analysis III it will be observed that canes having normal eyes show the same wide difference in content of reducing sugars in the nodes and internodes, and this difference cannot be due in the main to the incipient growth of the eye. Analysis of another ' Bull. 10, ist Ser.. p. 320, La. Sug. Expt. Sta. Vol. XVI.-34. 460 Beeson. variety of cane having normal eyes confirmed the resuUs of analysis III. Upon comparingr analysis V with analyses III and IV it will be seen that in the canes bearing sprouts (yet whose rootlets had not taken hold in the soil) there is found less glucose in both the nodes and internodes than in the case of canes having normal, and larger eyes. This difference is augmented when it is remem- bered that in case V the content of reducing sugars in the upper halves and upper thirds of the stalk — which are richer in reducing sugars — is compared against that of the whole stalk in cases III and IV. It would therefore appear that the growth of the sprouts on the stalk was at the expense of the glucose or reducing sugars, and that, in the case of the cane, glucose is the first visible product of plant-assimilation. During germination it is likely that the initial growth is at the expense of the ready formed glucose in the stalk — and the ripest canes have been shown to contain this sugar — and that the sucrose is attacked and inverted as a future supply of food when it is needed by the young plant. If so, it may be that there is permitted by the young plant a certain using up of this ready- formed glucose, a certain glucose minimum, before the sucrose is attacked as a further supply of food. Then by working with canes bearing longer and longer sprouts, the rootlets of which had not taken hold of the soil, it was hoped that this glucose minimum might be established for the sugar cane ; but an untimely frost killed the cane and thus cut short the work. Since the canes having normal eyes showed this marked decrease in content of reducing sugars in the nodes over that in the inter- nodes, and since these bodies are so readily diffusible, it was thought most likely (i) that there were some physiological forces at work in the plant organism connected with the ripening of the cane which were using up the reducing sugars more rapidly in the node than they could diffuse in from the adjoining internode ; (2) that the decrease in reducing sugars and increase in " solids not sugars" in the node mutually account for each other ; (3) that as the cane ripens these simpler sugars are passing by the condensation of the molecules into more complex and less soluble forms — f Whole of nodes. Class B-< ' I " " internodes, 12.77 14.87 0.9s 1.06 9.1 12.2 2.74 1. 41 Class cjW'^^l^^f^^^^^' t " " internodes, 13-62 1.44 8.1 3.08 15.17 1. 00 12.0 2.19 It will be seen that in the three stages of deterioration of the cane there is a gradual increase in the content of reducing sugars Constituents of the Nodes and Internodes of Sugar Cane. 463 in the juice of the node over that of the internode until they are in large excess of the latter. This is the reverse of what was found to be true in the case of ripening cane, analyses I and II showing an average of 0.14 per cent, of reducing sugars in the node and i per cent, in the internode, or a ratio of about i to 7. In these cases it is strange that the glucose, which is so readily diffusible, had not distributed itself equally throughout the whole stalk. But it is probable that the death of the cane brought about a change in the cell walls which prevented diffusion. Nothing short of experiments on the diffusion of glucose in dead cane can answer definitely this question. No such experiments were made. By comparing these results with those (analysis I) of the same variety made just before the frost, it will be seen that there is an increase in the quantities of " solids not sugars" in the juice of both the nodes and internodes, that of the former here, also, being in excess. At first sight it would appear that, since the " solids not sugars " of the juice would most easily yield reducing sugars by the action of ferments during decay, there should be found a decrease in "solids not sugars" in the juice of the node corres- ponding to the increase in the reducing sugars. But it will be remembered that there was found in the "fiber" of the node 13 per cent, of insoluble bodies not sugars which, by the action of ferments, would most likely yield first soluble bodies not sugars before passing into reducing sugars. This process would increase the "solids not sugars" in the juices of both the nodes and inter- nodes, as was found to be the case. Further, if this explanation be correct, this process would be at the expense of what is estimated as " fiber " in the cane, and canes which had deteriorated should not only show an increase in "solids not sugars" and in reducing sugars, as was found to be the case, but also a decrease in " fiber" — that portion of the cane which is insoluble in water. As a result of several years of experi- mentation at this station this has been found to be the case.' To recapitulate : it has been found in the course of this investi- gation that the juice of the nodes of the cane is quite different from that of the internodes, containing markedly less reducing sugars, more "solids not sugars," and more coagulable bodies; that the "fiber" of the nodes contains more albuminoids, more 1 Bull, shortly to be published. 464 Roelofsen. insoluble carbohydrates not sugars which readily pass into reducing sugars; that as the cane deteriorates reducing sugars are formed more rapidly in the nodes than in the internodes; and that probably glucose is the first visible product of plant-assimila- tion by the young cane. In our opinion these facts can be best explained by the hypothesis previously stated, namely, that the physiological function of the node in the cane is similar to that of the seeds in the case of flowering plants — to store food in the region of the eye for the use of the young plant before it has taken sufficient hold of the earth to draw sustenance from the atmos- phere and soil. This hypothesis is further confirmed by the fact that the isolated nodes of the cane when planted will germinate and grow to maturity.' If this be true in regard to the cane, it may be true also in the case of all varieties of plants which propagate from the nodes or joints. Louisiana Sugar School and Experiment Station. NOTES OF WORK FROM THE CHEMICAL LABORA- TORY OF THE UNIVERSITY OF VIRGINIA. NO. XIX. Communicated by F. P. Dunnington. 160.— ON THE SOLUBILITY OF CREAM OF TARTAR IN ALCOHOL OF VARIOUS STRENGTHS AND AT VARIOUS TEMPERATURES. By J. A. Roelofsen. In 1891 some work was done by Mr. W. H. Wenger in this laboratory on the solubility of cream of tartar in alcohol of various strengths at about 25° C, and the result was published in this Journal.* With a view to enlarge the knowledge upon the subject and to ascertain the solubility at various temperatures, the follow- ing work was undertaken. » Bull. 7, ist Series, La. Sug. Expt. Sta. ^ 14, 624. Solubility of Cream of Tartar in Alcohol. 465 Alcohol of 93 per cent, by weight was mixed with various amounts of a saturated solution of cream of tartar, previously puri- fied, to make liquids of different strengths. Lots containing 90, 80, 70, 60, 50, 40, 30, 20, 10, and o per cent, of alcohol respec- tively were made, and small bottles holding about 125 cc. filled with them. These were tightly corked and exposed for a number of hours, in no case less than six, to the following temperatures: 0°, 5°, lo^ I5^ 20°, 25°, 30^ 35°, 40°, 45°, 50° C. The bottles were frequently shaken. It was found necessary in the case of the weaker alcohols at the higher temperatures, and with water at all temperatures, to add some cream of tartar so as to maintain an excess and prevent supersaturation ; for, except in the stronger alcohols, it was found that the amount thrown down from solution on addition of the alcohol was not sufficient to saturate the liquids at the higher temperatures. For each determination 50 cc. were used, and this amount was drawn up into a pipette to the lower end of which was attached by means of rubber tubing a piece of glass tube 10 cm. long and 6 mm. in diameter, tightly packed with cotton-wool. In this way undissolved particles were kept from being drawn up. In a few cases it was very difficult to obtain perfectly clear solutions. The bottles subjected to the higher temperatures were packed in cotton after being taken from the hot-air chamber, to prevent loss of heat by radiation during the taking of samples, and observation of the temperature of the remainder of the liquid after the samples were taken showed no appreciable loss of heat. The samples were put into beakers, diluted with water, and determinations of the cream of tartar present made with decinormal solution of sodium hydrate, phenolphthalein being employed as an indicator. The sodium-hydrate solution was freed from carbonic acid by barium hydrate, and preserved in a bottle provided with a rubber stopper through which passed a syphon to draw off the solution as needed, and a U-tube containing solid potassium hydrate through which air was admitted. The solution was standardized against normal sulphuric acid, and this was repeated several times during the course of the work, which extended over some weeks. Four determinations of the dissolved cream of tartar were made in each case, and wherever these results gave a considerable devia- tion from a regular curve more than these were made. The averages of these are given in the table, and the diagram indicates <»° 5" 10° I 5'" 2 0° 2 5" 30' 35° 40= 45° Ha 10 / / / / / 90 /■ / 80 / / / / 70 / / / / / / / ■ ■ 7 / / so ^ ^ ^ r — 1 _^ ^ / V / 3«^ 40 i^ ^ / / / ^ / / 30 / / ^ ^ / / /' y ^ 20 y y ^-^ y -^ ,^-- ^ -^ ^ ^ ^ 40' ^ ^ -^ r . — — ' y -- — — = ~ SO cr: m ^- ^^ J __-- =:== ii:^ 60 70 80 90 • •*=^^ ~ "^ Solubility of Cream of Tartar in Alcohol. Abscissae represent temperatures; ordinates, milligrams dissolved in 10 cc. liquid. Figures on right of drawing represent percentage strength of alcohol. lodine-absorpiion of some of the rarer Fatty Oils. 467 by a separate line for water and for each strength of alcohol em- ployed, the weight in milligrams of cream of tartar dissolved in 10 cc. as determined for each 5° of temperature from 0° to 50° C. Temp. Milligrams of cream of tartar dissolved Alcohol of percentage 90 80 70 60 50 40 30 20 10 0° 6.2 6.4 49 6.0 6.0 6.2 7.0 10.8 17-3 30.1 5° 5-5 6.0 5-t 6.0 6.8 6.8 7-1 13.2 18.8 32.0 10° 6.2 6.2 5-1 5.8 6.4 7.0 8.6 16.0 27.0 41. 1 15° 5-3 6.2 6.2 6.2 5-5 7-7 8.8 15.8 23-9 44-3 20° 6.4 6.4 6.2 6.4 7.0 9.6 "•3 17.1 29-3 49.0 25° 4-7 5-5 6.0 6.8 7.0 10.3 II. 7 21.4 36.4 54.1 30° 4-7 6.0 6.8 7-5 8.5 II.O I3-I 24.8 39-9 69.2 35° 1.9 5-1 5-9 6.8 9.0 12.4 18.8 28.7 49-3 83.8 40° 1-7 5-3 5-8 7.0 10.2 14.9 23.1 37-7 53.6 95-9 45° 1-7 5-3 6.0 7-9 10.7 16.5 25.8 44.2 72.6 II 2.8 50° 1-5 5-1 6.0 8.1 12.8 19.0 29,7 53-6 87.2 124.8 161.— THE IODINE-ABSORPTION OF SOME OF THP: RARER FATTY OILS. By J. A. ROELOFSEN. The following figures on the iodine-absorption of some fatty oils were obtained according to the directions given in Allen's Com- mercial Organic Analysis where the process of Baron Hiibl is fully described. The time allowed for the absorption was two hours in each instance. The oils came from thoroughly reliable sources ; those marked * were obtained from exhibitors at the Philadelphia Centennial Exposition, and those marked f were gotten from and personally extracted by Baboo Kanny Loll Dey, Professor of Chemistry at the Medical College, Calcutta, and honorary and corresponding member of the Pharmaceutical Society of Great Britain. The oils were kept in securely stop- pered bottles in a protected place. Upon opening, all, with the exception of one, which is especially noticed in the following list, were found to be in good condition. A few of the more common oils, the absorption-figures of which were already well known, and also some few of which but a single figure was found on record, were worked for the sake of comparison, but the absorp- tion-figures of most of the oils have, as far as I have been able to ascertain, never before been determined. 468 Roelofsen. "2 *^'H '3 a = - 1) . . (U ^ ^ ^ ° 5 .-s ^ .r. ^ !=■ o o o *j c c ft i! "oJ "MI'S 13 tifl 60 to " ft „ „ _o rt X iJ ■ vo 00 00 00* tJ- g" ■5-vd « >-< 'fi'O >^ 0\»p on oi - r-^ ro d od CO roco fO 0\^ t^ vOOn— 0>-iOnO\ 0000'^ rt*- -« CQ ►fl.aq u K X •^ — N W3.2 2 .2 .2 'n ■ n !S 2 S "^ '5 "^ "5 £ s ^ -^ « 5 s § ■^ s^s ■2 TV ! 5 1 -§ s s s « S s i 'i « « 5 "^ "^ « 1^ ij ^ c/2 '^ iy ^^ ^ . ^ ^ ^ ^ =^ "^ S J c 2 si oo o X Sl^O a3--3oCur: t^ o c« rt o ys to-'- ■ft^-- ■5 ju "5 >, o.-::5ftS^«>^'^ £ -a -a o o o Iodine- absorption of some of the rarer Fatty Oils. 469 3 .2- M 1-1 3.2" o o ^ ^ Jld £ . o 13 "5 T3 5 t* c p S rt c _, 5 1 I Tucuman. ium herbaceufti. ays. ocarpus laevis. her a trijuga. achta Indica. "dium occidentale, rpus anacardium. i 1 a ^ ^ i |:^ ^■\ 11 ^s ^ s 1; ^ 5 § •■^^ « 2 s ■S ^ ■li ^ cq:c ^N C^c^^ ^:^ ^^^5 < tl :1 i iii^ ^ 'c -bad deco a-tel n pr c « >,-i^ •" ^ J '^^•^'S S . . u S '^-^ = e (4 « 'r 2 ?„ e' It E : • 0) — c J '5 -I nut oi was fo -nut 5ura ( - — > ■5 '0 ^ ■U 13 y. o_c rt "S j: iirjun- acassa im-ka- ashew (This arking e-no-a •a ::: — lit S 3 « c C J^C ^ ^ C! S ^ U >.> t-j H CA M s h m REVIEWS AND REPORTS. Recent Progress in Physical Chemistry. Die Dichten ges'ditigter D'dmpfe und ihre Beziehung zu den Gesetzen der Erstarrung tmd Verdampfung der Losungs- mittel. F. M. Raoult: Ztsclir. phys. Chem. 13, 187 (1894). — Van't Hoff's thermodynamical theory of the vapor-pressures of solutions is essentially a calculation of the change of free energy' accompanying the separation of solvent from the solution. When n g-mols. of non-volatile substance are dissolved in iVg-mols. of solvent, a separation, by isothermal vaporization and subsequent condensation, of that amount Njn g-mols. of solvent in which one g-mol. is dissolved, requires the work (change of free energy) — I vdf-=. — RT log -7T when the density d of the vapor is the theoretical. In this formula v and Z'are the molecular volume and absolute tempera- ture of the vapor, /"and/'' are the vapor-pressures of solvent and solution, R is the " gas-constant." When the vapor-density has an abnormal value J (as is the case with the fatty acids), the change of free energy is d N ^rr, f —T' — RT log ^ ; and since the same quantity, when the separation is effected osmotically by a semi-permeable membrane, is RT^ we have as the most general result, Raoult points out that for most solvents Ajd is nearly unity, yet the variation therefrom may be appreciable. A most satisfactory confirmation of the theory is furnished by the following agreement between the Ajd values calculated by Raoult from the vapor- pressures and those directly observed by (chiefly) Battelli, and Ramsay and Young : Solvent. Temperature. A/ Ztschr. phys. Chem. 12, 244. * Published under the s«ine title in book form (Basel, 1893). 472 Reviews and Reports. The latter question is answered emphatically in the affirmative, for Kahlbaum finds the vapor-pressures of water and of mercury, by employment of the dynamical method, to be as good as identical with those obtained by Regnault and Hertz by the statical one. A repetition, further, of Landolt's determinations with the fatty acids resulted in a correction of Landolt's figures, of sufficient amount to establish a satisfactory agreement here also between the results of the two methods. It is thus at last demon- strated that the easily executed dynamical method has every advantage over the statical one. The paper is concluded by a tabulation of the data obtained in the study of thirteen fatty acids and of various mixtures. Primare oder secund'dre elektrolytische Wasserzersetzung. M. Le Blanc : Ztschr.phys. Chefn. 13, 163. — In the present article LeBlanc upholds against Arrhenius the view developed by him- self in recent papers upon the electromotive forces of polarization, that in the electrolysis of those aqueous solutions from which hydrogen and oxygen are evolved by the current the water present suffers a primary decomposition. To precipitate an ion upon an electrode requires an electromo- tive force sufficient to overbalance the counter electrolytic solu- tion-pressure of the precipitated substance, i. e. its tendency to go into solution again as an ion. This solution-pressure varies from substance to substance in the same way that the vapor-pres- sure varies, and with low impressed electromotive forces the sub- stances with smallest solution-pressures must necessarily be the first to be thrown out. In all aqueous solutions the hydrogen cation of water is present — in very slight concentrations, it is true, but yet appreciably present — and since the electrolytic solution- pressure of hydrogen is relatively low, this substance is usually the first to be precipitated. (The metals with still lower solution- pressures are copper, mercury and silver.) Although the current is carried through the solution in great part by the ions of the dissolved electrolytes, the concentrations of these ions being rela- tively great, yet the precipitation of material upon the electrodes is another thing, and involves first those substances which go out the most easily — involves for most electrolytic solutions the hydrogen and hydroxyl ions of the solvent. And as fast as these are thrown out at the respective electrodes their places are supplied by the continued (electrolytic) dissociation necessary to maintain their equilibrium-concentrations there. The decomposi- tion of the liberated hydroxyl into water and free oxygen is of course a purely secondary reaction. When the current-density is so increased that more current passes than these ions can carry, the electromotive force immediately rises to a point at which one or more of the dissolved ions present are precipitated also. The older hypothesis of a primary separation of all the cations at the cathode, and similarly for the anode, would require, to Reviews and Reports. 473 account for the behavior of a solution of potassium, zinc, copper and silver salts, the complex and somewhat ridiculous assumption that secondarily the liberated potassium replaces the zinc, hydro- gen, copper and silver, the zinc then replacing the last three, the hydrogen the remaining two, and the copper finally the less "positive" silver, which is thereupon deposited. It is much simpler to regard the current as carried by all the ions present, and the successive separation of the cations to Ibllow the order of their respective electrolytic solution-pressures — as is actually observed. From this point of view it is apparent that the quan- titative separation of ions from one another by electrolysis depends more essentially upon the electromotive forces employed than upon the amount of current — an entirely new development. LeBlanc was led to this view of the primary separation of the least positive ion, by the identical values of the electromotive forces required to electrolyze a large number of aqueous acids and bases studied by him. The idea is a distinct advance in the theory of electrolysis, and the paper in question is a clear and dignified exposition of it. Beitr'dge zur Stochiometrie der lonenbeweglichkeit. G. Bredig : Ztschr. phys. Cheni. 13, 191. — The investigation published under the above title is an unusually exhaustive and careful piece of work. The author has determined, with the employment of a high degree of experimental skill, the maximum electrical conduc- tivities of a hundred and fifty salts in aqueous solution at 25°, and from these and older data has calculated the migration-velocities of some three hundred ions. An ingenious detail of the work is the successful measurement of the conductivities of salts of weak organic bases, after repressing their hydrolytic dissociation by the addition of an excess of the free base: an especially interesting application of the law of mass-action (or reaction-isotherm). The author points out that the migration-velocity of the elemen- tary ions is a pronounced periodic function of their atomic weights. In connection with many interesting facts concerning the relative velocities of complex ions, he brings forward the unexpected result that the migration-velocity of a metameric cation is the greater the more symmetrical its substitutions. He proves further that purely isomeric cations migrate at equal rates, and that in all cases the velocity decreases asymptotically with increasing com- plexity of structure. The paper is an important contribution to our knowledge of the ion-velocities. Ueber die Affinii'dtsgrdssen der Basen. G. Bredig : Ztschr. phys. Chem. 13, 289. — With the aid of the migration-velocities of the organic cations obtained in the preceding paper, the author has determined the dissociation-constants (affinity-constants) of over forty amines and ammonium bases, finding the dissociation-equi- librium to be accurately represented in all accessible cases by the 474 Reviews and Reports. theoretical isotherm employed by Ostwald for the acids. The quaternary bases are as strong as the fixed alkalies, the secondary bases are notably weaker (of 30-40 per cent, dissociation in yi(y-normal solutions), the tertiary and primary amines are the weakest of all. Ammonium hydroxide is but slightly stronger than acetic acid; the substituted phosphonium, arsonium, stibonium, sulphinium and tellurinium bases are very strong ; the tin and mercury bases very weak. The author demonstrates that meas- urement of the electrical conductivity of the hydrochlorides of very weak bases yields data from which the affinity-constants of the free bases can be found with the aid of the law of mass-action, and he concludes the paper by pointing out various errors and misapprehensions into which Lellmann and his pupils have fallen in their recent attacks upon the ion-theory of electrolytic solutions. Ueber die Zersetz2ing des Jodwasserstoffgases in der Hitze. Max Bode7istein: Ztschr. phys. Chem. 13, 5b. — The material for this exhaustive study of the dissociation of hydriodic acid was prepared synthetically by conducting iodine-vapor and hydrogen over hot, platinized asbestos, and absorbing by water the acid pro- duced. It could then be obtained when desired by heating this solution and purifying the evolved gas (from water) by phosphorus pentoxide and (from iodine) by a pure red phosphorus. The yield was about 90 per cent, of the iodine employed. The usual statement that dissociation begins near 180° was shown to be incorrect, for an appreciable although slow decomposition takes place at 100°. At 448°, the temperature of boiling sulphur, the dissociation according to the equation 2HI^H2-f I2 ' amounted to 21.5 percent, of the gas. used, and the amount remaining uncombined in the formation of the acid from free hydrogen and iodine was 21.0 per cent., so that the dissociating system was beyond question in a true state of thermodynamical equilibrium. The degree of dissociation in vessels filled under atmospheric pressure changed from 16.4 per cent, at 290° to 23.6 per cent, at 518°, passing through a minimum near 325°. The rise from this point proves that the heat of formation at all higher temperatures is positive ; it must accordingly be negative (heat absorbed) at lower temperatures, which is in accordance with the thermochemical determinations of Thomsen. The velocities with which the dissociation- and formation- reactions occur were unexpectedly found to be entirely regular; the progress of the dissociation was accurately represented by the integrated form of the differential equation Reviews and Reports. 475 representing the velocities of bimoleciilar reactions; Cand Care constants, the dissociation is x per cent, at the time /. This is the first case in which a uniform reaction-velocity has been observed in gaseous systeois. The velocity increases ol course enormously with rising temperature. A remarkable result of the investigation is the very consider- able increase of the dissociation under rising pressure, a phenom- enon seemingly explainable only by the lack of strict proportion- ality between the partial pressures of the gases and their mole- cular concentrations. J. E. Trevor. Cornell University. Handbuch der Stereochemie. Edited by C. A. Bischoff, Professor of Chemistry at the Polytechnicum of Riga, with the assistance of Paul Walden, Decent in Physical Chemistry. First volume, pp. 448. 8vo. 111. (M. 14). Frankfurt : Bechhold, 1893. To those who have not followed with special attention the progress of stereochemistry, it will be a surprise to find Le Bel speaking of this as "a positive science which deserves the place it has gained in our colleges "; to know that for some years a "professor of stereochemistry " has been lecturing at the Univer- sity of Geneva, and finally to learn that a work so extensive as Prof. Bischoff's could be written about a branch of chemistry of which seven years ago nothing was heard. As its name implies, this book is intended chiefly for reference. The bulk of the first volume consists of data concerning the optical rotation of all active organic substances; and the great number of constants here collated by Dr. Walden will have permanent value. The " Historical Introduction" by Prof. Bischoff is full without being tedious. The principal chemists concerned are allowed to speak in their own words ; and (except in the case of his own theories) there is no criticism by the author. It must be observed that the experimental work of Burch and Marsh (called Bush and March) receives no more attention than that accorded to the spelling of their names ; and the aromatic- ring theory of Marsh is also overlooked. On the other hand it is a satisfaction to find that the theories of Van't Hoff and of Le Bel are here recognized as two and distinct. Due importance is accorded to Le Bel's proof of the existence of asymmetric nitrogen, and his attempts to prove the activity of non-saturated carbon compounds " promise a still greater revolu- tion in the prevailing principles of stereochemistry." Unless such a revolution should occur we shall find confronting us this singular situation : the elaboration of the idea of the three-dimen- sional arrangement of atoms will have shown that in most cases either the arrangements are two-dimensional or there is no per- manent arrangement at all. For if a substance CaRiR2R3R4 (for Vol. XV I. -35: 476 Reviews and Reports. example) is inactive, this must be either because the atoms, like those of CiRiR2R3R4, are in the same plane, or because they incessantly change places. (The tetrahedron hypothesis demands, of course, a three-dimensional arrangement for CsRiRsRsRj.) The second volume of Prof. Bischoff 's book will be awaited with interest: it will treat of the influence of space-relations on chemical reactions, and of the motion of the atoms within the molecule. Arnold Eiloart. A System of Instruction in Qualitative Chemical Analysis, By Arthur H. Elliott, Ph.D., Professor of Chemistry and Physics, and Director of the Chemical Laboratory in the College of Pharmacy of the City of New York. Second Edition, 1894. Published by the Author. The first edition of Prof. Elliott's manual has already been favorably criticized in this journal.' In the preface to the second edition the author says that it has been adopted as text-book by a number of institutions. The book deserves success and can be recommended for its clearness and attention to details. e. r. Centenary Commemoration of Antoine-Laurent Lavoisier, 1794— May 8 — 1894. Dr. Gustavus Hinrichs, of St. Louis, has issued a circular letter calling upon chemists to contribute the funds necessary for the erection in the city of Paris of a monument to Lavoisier. The principal part of the letter is here reproduced : " An entire century has rolled over the unmarked grave of the Copernicus of Chemistry ; with to-day, the second century takes its beginning. Will not the chemists of the world deem it a priv- ilege to erect a monument to the memory of their common master who fell a victim to the frenzy of the Reign of Terror ? And will not the chemists of France open the way and furnish a place for it to permit us to enjoy that privilege? "The life-work of Lavoisier was deeper and broader than the discovery of any new substance, and affected the very foundation of the science of chemistry. He broke through the veil of mere phenomena and discovered beyond it the reality of chemical pro- cesses. " History teaches that in every science man has first mistaken the direct testimony of his senses, the mere appearances or phe- norhena, for the essence of things real, and that only a master-mind can penetrate this veil in which our senses wrap the reality of things and hide it from our reason. Reviews and Reports. 477 " The work of Lavoisier is exactly of the same character as that of Copernicus ; his system is equally bold, and as plainly con- tradicted by the immediate testimony of the senses. That is the reason why Scheele and Priestley never surrendered, though they themselves had discovered the very substance which, in the hands of Lavoisier, became the corner-stone of the new system of chemistry. The direct testimony of the senses is undoubtedly that fire-stuff, phlogiston, is set free when carbon, sulphur and the metals do burn. The customary superficial representation of the phlogistic period is as unjust to the eminent chemists of that period as to their opponents. " Indeed, it required the mind of a Lavoisier, and its persistent labor for twenty years, to accomplish the Copernican revolution in chemistry. He had to devise those admirable experiments with air and metals which demonstrate, once for all, that also in chem- istry the testimony of the senses has been misleading, and that where phlogiston appears to be given off, an invisible substance is really taken up. Lavoisier did not tire till he had, balance in hand, demonstrated that in all chemical processes, the sum-total of matter involved remains exactly the same (Trait6, I, p. 141), precisely as in the Copernican planetary system the spheres retain their identity and maintain their well-ordered motions unchanged, how- ever much they may appear to wander astray and become retro- grade when seen from this earth. " By the use of his calorimeter, Lavoisier even made a good beginning on the investigation of the dynamic phases of chemical processes. " Laplace has given (Systeme du Monde, IIL 6d.,p. 357 ; 1808) a beautiful word-picture of the life-work of Copernicus ; by simply replacing the astronomical terms by their chemical equivalents, this picture changes to a faithful graphic representation of the work of Lavoisier. " One other fact should be noted to avoid further misconcep- tions, and especially to forever bar the recurrence of national jealousies. Also on this point the parallel holds good. "Copernicus is not extolled as a mere observer. Several of his immediate predecessors, as well as many of his followers, greatly surpassed him in this direction. Yet his mind saw clearly many things that the human eye unaided could never have perceived. Not until Galilei, long after the death of Copernicus, had armed himself with his new telescope, did mortal eye behold the moon- like phases of Venus, mentally described by Copernicus ; at the same time the telescope revealed a miniature Copernican system in the world of Jupiter. " In the same manner we will cheerfully grant that some of the contemporaries of Lavoisier may have been more skilled experi- menters in some directions, and that no doubt he left much for his followers to do. Nevertheless, his Traite ele'me7itah-e de 478 Reviews and Reports. Cliimie is unquestionably the first rational exposition of the science of chemistry, entirely resting on experimental evidence, largely his own, and admitting to the list of entities of matter nothing that was not actually produced; and since that day, chemistry is the science of the real elements. This true insight in the nature of matter permitted him to see into the future, as Copernicus had done, and to describe elements before the instru- ment required to produce them had been invented. When Davy, by means of the powerful galvanic battery of the Royal Institution of Great Britain, produced the light metals, he con- firmed what Lavoisier had seen mentally more than thirty years before. (Trait6, I, pp. 194, 195.) " The life-work of Lavoisier has been admirably summed up by Berthelot in the following words (La Revolution Chimique; Paris, 1890; p. 207): "'That which we have a right to admire, that which the uni- versal judgment of the civilized world with every day hallows more and more, is the positive work which he has accomplished ; namely, the final erection, upon a permanent basis, of one of the fundamental sciences, Chemistry. No work in the history of civilization is greater, and therefore the name of Lavoisier will live forever in the memory of humanity.' " To these clear and true words nothing can possibly be added, except it be the expression of the hope that committees will form in all countries and co-operate with the one that should spring into being at the place of the work and glory, the suffering and death, of Lavoisier. It would be a most encouraging token for the continued progress of our race if the coming twentieth cen- tury should find, at the city of Paris, a monument erected to the memory of Lavoisier by the chemists of all the nations of the earth. " I make bold to suggest the union for this noble purpose by adopting as my own the words spoken in the College de France, by Dumas, at the close of his memorial lecture on Lavoisier, on May 7th, 1836: " ' May .... this tribute, the feeble expression of the senti- ments of veneration which fill my soul .... awaken memories and excite the zeal of men in positions of greater influence .... that there may be erected, to the memory of Lavoisier, a monu- ment .... which to posterity shall express our profound admiration for his genius and our lasting sorrow for his pre- mature death.' " Vol. XVI. [November, 1894.] No. AMERICAN CHEMICAL JOURNAL, INSTRUMENTS FOR THE GRADUATION AND CALI- BRATION OF VOLUMETRIC APPARATUS. By H. N. Morse and T. L. Blalock. The apparatus represented in Fig. i is employed both for the graduation and the verification of liter and half-liter measuring- flasks. The delivering capacity of the bulb between the mark on the upper, stem and the zero mark on the lower one must not exceed 500 cubic centimeters at the highest temperature at which the instrument is to be used ; while the combined delivering capacity of the bulb and the graduated portion of the lower stem must not be less than half a liter at 0°. The lower stem is graduated in millimeter divisions which are numbered in both directions, in order that the instrument may be used in the inverted position if desired. For the purpose of explaining the method of preparing an instrument of this kind for use, an account of an actual calibration will be given. The water delivered by the bulb, and also that delivered by the graduated portion of the stem, was weighed and its temperature noted. The delivering capacity of the bulb and of the stem was then calculated by the familiar formula, in which P\s the weight of the water when weighed in the air with brass weights (497.769 grams for the bulb and 3.0708 grams for the stem) ; Vol. XVI.-36. 500 c cm '^&iX \l Graduation and Calibration of Volumetric Apparatus. 481 / is the weight in a vacuum of one gram of water weighed in the air ( 1. 00 1059 grams at ordinary temperatures) ; and d is the density of the water at the temperature of weighing (21.4"). In this way the capacity of the bulb at 21.4° was found to be 499,324 cubic centimeters, and that of the stem, at the same temperature, 3.081 cubic centimeters. The capacity of the bulb at 0°, 4°, 10°, 12.5°, 15°, 17.5°, 20**, 22.5°, 25°, 27.5° and 30°, was calculated by the formula, Fi = M{i +r (/i— /)}, in which P,p and d have the same values as in the previous equation; Y is the expansion-coefficient of glass (0.000025) ; /i is the temperature, 0°, 4° or 10°, etc., and t is the temperature of the water at the time of weighing. The results of these calculations are given in the second column of the table which follows. No corresponding calculation was made for the stem, because the change in its capacity between 0° and 30° amounts to less than 0.0025 of a cubicentimeter. The inequalities of its bore were also found to be insignificant. Each millimeter division of the stem was therefore regarded as having a delivering capacity of 0.03081 cubicentimeter at all tempera- tures between 0° and 30°. The next step was to find to what points on the graduated stem the water must be drawn in order to graduate a half-liter flask for each of the temperatures included in the first column. This was accomplished by dividing the differences between 500 and the several numbers in the second column by 0.03081. The results are given in the third column. To prepare the instrument for use in the graduation and verifi- cation of apparatus when the so-called Mohr system is to be employed — i. e., when the correction for air-displacement is to be dispensed with, the quantities in the second column were divided by the volumes at the different temperatures of one gram of water when weighed in the air with brass weights. These are : At 0° , 1.001185 ccm. at 20" , 1.0028 1 2 ccm, 4° , 1.001057 22.5^ 1.003363 10° , 1.001322 25° , 1.003974 12.5°, 1.001581 27-5°, , 1. 00465 1 15° , 1.001917 2,0^ . 1.005380 17.5°, 1.002330 * 482 Morse and Blalock. The results are given in the fourth column of the table. The final step was to find to what point on the stem the water must be drawn in order to graduate flasks on the Mohr system for each of the temperatures recorded in the first column. For this purpose the capacity of the stem (3.081 cc.) was divided by 1.00233, the true volume of the Mohr unit at 17.5°, which gave 3.0738 as the capacity of the stem in Mohr units, or 0.030738 as the capacity of a single division. The differences between 500 and the several numbers in the fourth column were then divided by 0.030738, giving the numbers which are recorded in the fifth column. Strictly speaking, each difference should be divided by a different number ; but the maximum error which could ever result from dividing all the differences by the volume of a gram of water when weighed in air at 17.5° (Mohr's standard temperature) is less than one hundredth of a cubicentimeter, and therefore inappreciable. Table. Capacity of Divisions of Stem Capacity of Divisions of Stem Temperature. Bulb in cc. to be added. Bulb in Mohr units. to be added. 0° 499.048 30.9 498.457 50.2 4° 499.098 29.3 498.571 46.5 10° 499.176 26.7 498.517 48.2 12.5° 499.208 25.7 498.420 514 15° 499,241 24.6 498.286 56.1 17.5° 499.273 23.6 498.112 61.4 20° 499.304 22.6 497-937 67.1 22.5'' 499.338 21.5 497.664 76.0 25° 499.372 20.4 497-395 84-7 27.5° 499.403 19.4 497.091 94.6 30° 499.436 18.3 496.763 105.3 Capacity of one division of the stem in cubicentimeters, 0.03081. Capacity ' of one division of the stem in Mohr units. , 0.030738. The use of the instrument is exceedingly simple. It is attached at the lower end to a Greiner and Friedrich's stop-cock, which is in turn connected by means of a rubber tube with a reservoir of distilled water placed somewhat higher than the instrument itself. The water should have about the temperature of the room. Its exact temperature, however, is a matter of indifference. Suppose now it is required to graduate a flask in such a way that it shall Graduation and Calibration of Volumetric Apparatus. 483 have a capacity of 500 cubicentimeters at 20®. The whole apparatus is filled from the reservoir and emptied, once or twice, in order that the glass may acquire the temperature of the water. It is then again filled with water to the mark on the upper stem, and the water is delivered into the flask until the point indicated in the third column of the table (22.6) is reached. The flask is then placed upon a turn-table and marked with a diamond in the usual manner. If the flask is to be graduated for the same tem- perature on the Mohr system, the point to which the water must be drawn off (67.1) will be found on the same horizontal line in the fifth column. In the same way a flask may be graduated for any other temperature included in the table. The graduation will be correct for the given temperature whatever may be the tem- perature of the water at the time of graduation; since equal changes of temperature produce equal changes in the capacity of both the standard pipette and the flask. The instrument may be used equally well for the graduation of large gas-measuring apparatus. Suppose, again, that it is required to find the capacity, at any given temperature, of a flask already graduated. The flask is filled to the mark from the pipette, and to the capacity of the bulb at the given temperature there is added the product of the reading on the stem and the volume of one millimeter division of the stem. If the capacity of the flask in cubicentimeters is to be found, the numbers in the second column will be employed for the bulb, and the value 0.03081 for the stem ; while, if it is desired to find its capacity in Mohr units, the corresponding numbers in the fourth column will be used for the bulb, and 0.030738 for the stem divisions. This method of graduation and verification has been found very satisfactory in practice. To determine what degree of accuracy may be expected from it, fourteen liter and half-liter flasks, representing the product of six different makers, were selected, and the capacity of each was determined both by means of the standard pipette and by weighing the water required to fill the flask. The maximum difference in the results obtained by the two methods was less than ^ ^ j ^ ,^ of the capacity of the flask, while the average difference was gTToT- ^^ is interesting, as show- ing to what extent the graduated apparatus of different makers may vary, to note that flasks of nominally equal capacity were in some cases found to differ by one-tenth of one per cent. 484 Morse and Blalock. The interior of the pipette must, of course, be scrupulously cleansed. The apparatus represented in Fig. 2 is used for the graduation and verification of smaller measuring-flasks. The capacities of the bulbs and of the stems are subject to the same limitations as in the preceding case ; that is, the delivering capacity of the smaller bulb must not exceed 50, and that of the larger one 200 cubicentimeters, at the highest temperature at which the in- strument is to be used, while the combined capacity of each bulb and its stem must not be less than 50 and 200 cubicentimeters respectively at 0°. The delivering capacity of each bulb and stem, at different temper- atures, is determined, and the results are tabulated in the manner previously ex- plained. 50-, 200-, or 250-cubicentimeter flasks can be graduated or their capacities determined by a single filling of the pi- pette. To graduate a loo-cubicentimeter flask the smaller bulb must be twice filled. The stop-cock may, of course, be attached to either end of the instrument. Fig. 3 represents a modification of Ost- wald's calibration pipette for burettes.' It differs from the original form in that a Greiner and Friedrich's stop-cock has been substituted for the side tube and pinch- cocks of the Ostwald apparatus. The ad- vantage of such a substitution lies in the fact that it does away with the only objec- tion to the original pipette, viz., the danger that the compression of the rubber con- necting-tubes by the pinch-cocks may not pjg 2. always have the same effect upon the ca- pacity of the ungraduated portion of the apparatus, i. e., that portion of it which lies between the gradua- tion of the burette and of the pipette. Another advantage of the 1 Ztschr. anal. Chem. 22, 549. Fig. 3. Fig. 4. 486 Morse and Blalock. modified form is that mercury, as well as water, may be employed for the calibration of burettes. If mercury is used, the walls of the rubber connecting-tubes should be supported by winding them with cord or copper wire. A still better form of calibrating pipette is shown in Fig. 4. The largest bulb, together with the graduated upper stem, is employed to determine the total delivering capacity of a burette at any temperature, while the smaller ones are used to ascertain the inequalities of its caliber. The largest bulb delivers somewhat less than 50 cubicentimeters at 30°, while the bulb and the upper graduated stem together deliver more than 50 cubicentimeters at o''. The smaller bulbs and the graduated portions of the stem which belong to them are subject to the same relative limitations in respect to delivering capacity. In other words, at 30° the bulb marked 3 ccm. delivers a little less than 3, and that marked 2 ccm. a little less than 2 cubicentimeters, while the first together with the graduated portion of the stem above it, and the second with the graduated stem below it, will deliver not less than 3 and 2 cubi- centimeters respectively at 0°. The delivering capacity of each bulb and of each graduated section of the stem is determined for all required temperatures, and the results are tabulated in the manner already described. The procedure with a burette is as follows : The burette and the pipette are both filled with water to the upper limits of their graduation. It is convenient for the purpose of filling to insert between the burette and pipette a T-tube, one branch of which is supplied with a ground-glass stop-cock, and to connect this by means of a rubber tube with a supply of water which is placed somewhat above the upper limit of the graduation on the burette. Having filled both instruments, the water in the pipette is drawn down to the beginning of the graduation under the large bulb. The water in the burette is then allowed to flow into the pipette until the graduated portion of the former is emptied. The delivering capacity of the whole burette, at any temperature, may then be ascertained by reference to the table. The method employed by us in correcting for the inequalities of caliber does not require the exact capacity of the smaller bulbs to be known. Equal quantities of water — approximately 2, 3 or 5 cubicentimeters — are drawn off. and a record is made of the suc- cessive readings upon the burette. The subsequent procedure will be better explained by an example. A burette which had Graduation and Calibration of Volumetric Apparatus. 487 been graduated on the Mohr system for 15°, but which was to be used for solutions prepared at 20°, was found to have at the latter temperature a delivering capacity of 50.258 cubicentinieters. The burette was filled, and the water drawn off in quantities which exactly filled the smallest bulb of the pipette. The readings upon the burette corresponding to the successive pipettefuls are given in the table under R. The 25th and last reading was 49.52. The capacity of the burette to this point on its graduation was found by the proportion 50 : 50.258 : 49.52 : x, to be 49.7755 cubicen- timeters. The volume of a bulbful was, therefore, ^-^^i^^* o*" 1. 99102 cubicentimeters. The correct reading corresponding to each actual reading was then found by multiplying i. 99102 by the series of numbers from i to 25. The results are given under C. R. The differences, z. e., the quantities to be added as corrections to the actual readings, are given under D. Finally, a curve of cor- rections for the burette was plotted upon cross-ruled paper, using a line representing the graduation upon the burette as the axis of abscissas, and the corrections under D as the ordinates. R. C. R. D. I 1-95 1.99 --O.O3 2 394 3-98 - - 0.04 •3 5-93 5-97 4-0.04 4 7.94 7-96 -}- 0.02 5 9.91 9-96 + 0.05 6 11.90 11.95 -f 0.05 7 13-85 13-94 -f 0.09 8 15.84 15-93 + 0.09 9 17-83 17.92 + 0.09 10 19.78 19-91 + 0.13 II 21.74 21.90 -fo.i6 12 23-75 23.89 + 0.14 13 25-74 25.88 + 0.14 14 27.70 27-87 + 0.17 15 29.70 29.87 + 0.17 16 31.68 31.86 4-0.18 17 33-65 3385 + 0.20 18 35-62 35-84 4-0-22 19 37.60 37-83 + 0.23 20 3959 39.82 + 0.23 21 41.58 41.81 + 0.23 22 43-56 43-80 4-0.24 23 45-55 45-79 + 0.24 24 47-54 47.78 + 0.24 25 49-52 49-78 + 0.26 50.00 50.26 ■H-0.26 488 Michaud. It will be readily seen that the table belonging to the pipette is so constructed that a curve of corrections for any other tempera- ture, or for the Mohr system, can be drawn with equal facility. Note. The apparatus here described was made by Mr. Emil Greiner of New York. Contribution from the Chemical Laboratory of the Costa Rican Government. NOTE ON THE INFLUENCE OF CERTAIN METALS ON THE STABILITY OF THE AMALGAM OF AMMONIUM. By Gustave Michaud. The metallic radicle ammonium, which cannot be obtained in a state of purity, seems to acquire some stability in the state of alloy with mercury. Other metals than mercury might possibly form alloys still more stable with the ammonium, and with a view to testing this I prepared a series of amalgams, each of which con- tained, besides mercury and ammonium, one of the metals given in the list below. These amalgams were made by pouring loo cc. of a solution of ammonium chloride, saturated at 25°, on a mixture of 5 cc. of sodium amalgam with 5 cc. of an amalgam of each metal. This latter amalgam, in turn, was obtained by placing sodium amalgam in contact with a salt of the metal. As a result of the method used for its preparation, the amount of the metal whose action was to be examined was proportional to the chemical equivalent of this metal. The sodium amalgam used was the same through- out all the experiments and contained 0.85 per cent, of sodium. The experiments were made simultaneously in graduated cylin- ders of 200 cc, in order to allow the measuring of the volume of the amalgam at any time. One of the graduated cylinders con- tained nothing but the sodium amalgam, to which 5 cc. of pure mercury had been added in order to have the same percentage of sodium as in the other cylinders. From the beginning of the experiment two facts were plainly perceived, viz., that the volume of the amalgam of ammonium, On the Stability of the Amalgam of Ammonium. 489 when made in the presence of a third metal, was in nearly every case less than that of the same amalgam when pure, and that, in some cases, there was no increase of volume at all, the ammonium being decomposed, in proportion as it was formed, into a mixture of ammonia and hydrogen, which escaped with a considerable effervescence. The following table shows these facts and the relative influence of the principal metals on the stability of the amalgam of ammo- nium.. Increase of volume after Increase of volume after Metal added to the addition of loocc. of a Metal added to the addition of loo cc. of a amalgam of am- solution of chloride of amalgam of am- solution of chloride of monium. ammonium. monium. ammonium. None 115 cc. Fe no cc. Ag 50 " Mg 115 " Al 115 " Mn no " Au 35 '• Pb 65 " Bi 5 " Pt " Cd 80 " Sb 30 " Co 30 " Sn 25 " Cu 105 " Zn no " It is impossible not to recognize a certain relation between the electric polarity of metals and the influence which they have on the stability of the amalgam of ammonium, the less electro- positive metals being as a rule those which prevent the formation of the radicle. On the other hand this rule shows several excep- tions. Aluminium and magnesium do not seem to have any influence whatever on the stability of ammonium amalgam, but they would probably interfere with it if it were possible to intro- duce them in large quantities in a liquid amalgam. At least an experiment in which I tried without result to prepare the amalgam with alum of aluminium and ammonium instead of ammonium chloride, led me to that conclusion, as it was evident that the presence of aluminium was the only factor that could prevent the formation of the ammonium. To sum up the results of this investigation I may deduce the following conclusions: 1. When the radicle ammonium is set free in presence of an alloy of mercury and of one of the aforesaid metals, its stability is inferior to what it would be when associated with mercury alone. 2. A very small percentage of platinum in the amalgam of ammonium absolutely prevents the formation of the ammonium. 490 Herty, 3. The power of decreasing the stability of the radicle ammo- nium is greater in the electro-negative metals, but is not propor- tional to their polarity. 4. In spite of the contrary conclusion of several chemists,' the amalgam of ammonium does contain the ammonium, for when- ever a simple mixture of hydrogen and ammonia is evolved during the experiment, it has not the power to dilate the mercury and to give it the characteristic appearance of the amalgam of ammonium. MIXED DOUBLE HALIDES OF ANTIMONY AND POTASSIUM. By Charles H. Herty. Hoping to obtain evidence bearmg upon the question of the constitution of double halides, R. W. Atkinson'' took the plausible position that if the theory of molecular combination be true, as regards the formation of these compounds, then by mixing one equivalent of antimony trichloride with three of potassium bro- mide, and one of antimony tribromide with three of potassium chloride, two distinct isomeric salts of the general formula KsSbCbBrs should be obtained, namely, SbCls.sKBr and SbBrs, 3KCI. But on mixing the antimony and potassium salts in these proportions he obtained the same product in each case. This was confirmed by analysis of the compounds and by measure- ment of the crystals. From the analytical results he deduced the formula SbChBrsKj.iJHsO. Atkinson found further evidence that no molecular combina- tion had taken place in the fact that, on heating the above com- pound, fumes of antimony salts escaped, and a residue remained consisting of equivalent parts of potassium chloride and potassium bromide. From this fact he concludes that " This . . . also answers sufficiently the objection, that by the act of solution a redistribution of elements takes place, and that therefore we must obtain the same compound from whichever pairs we start. Grant- 1 Landolt: Ztschr. Chem. [2], 5,429. 2 J. Chem. Soc. 1883, 289. Mixed Double Halides of Antimony and Potassium. 491 ing that it were so, the crystals resulting would have one or other of the two constitutions SbBrs.sKCI or SbCl3.3KBr, and on heat- ing either all the bromine or all the chlorine would be driven off, not one-half of each." By using varying proportions of antimony and potassium salts he obtained compounds to which he ascribed the formulas Sb2C]6Br3K3.2H20 and SbChBrK.HsO. In a recent article' I have shown that double halides of lead and potassium containing more than one halogen are not true salts, but isomorphous mixtures of double halides containing only one halogen. It has seemed advisable, therefore, to study these salts of Atkinson more carefully and determine whether they are true chemical compounds or isomorphous mixtures. Preparation of Compounds. If it be true that definite chemical compounds are formed by the union of an alkali halide with the halide of a heavy metal, the halogen being different in the two substances, then by varying the proportion of one or the other of these substances we should expect the same salt to be formed under at least slight variations of conditions. In the case of mixed double halides of lead and potassium it has been shown that, by using the same quantities of potassium bromide, but slightly varying quantities of lead iodide, compounds could be obtained, each of which showed a slight variation in composition from that of any of the other compounds. Again, by substituting for portions of potassium bromide equiv- alent quantities of potassium iodide, and for lead iodide equivalent quantities of lead bromide, compounds could be obtained at will, in accordance with the variations in the composition of the solu- tions from which they were formed. In the following investigation the latter method has been adopted for determining the character of the compounds. In this manner all of the solutions contained the same weights of anti- mony and potassium, but varying weights of bromine and chlorine. 60 cc. of water, heated almost to boiling, was used in the prepa- ration of each solution. Beginning with the proportions indicated by the formula SbCh.sKBr, and using in each case the same quantity of antimony chloride, a series of solutions was prepared by substituting for varying quantities of potassium bromide equiv- > This Journal 15, 8i. 492 Herty. alent quantities of potassium chloride. Another series of solu- tions was then prepared by using in each case the same quantity of potassium bromide, as represented in the above formula, but replacing varying quantities of the antimony chloride by equiva- lent quantities of antimony bromide. The actual weights in grams of salts used in the preparation of these solutions were as follows : Table I. SbClj. SbBrj. KCl. KBr. B 40.00 ... 30.0624 15-156 C 40.00 22.5468 27-156 D 40.00 7-5156 51.156 E 40.00 ... ... 63.156 F 20.00 31.80 ... 63-156 G 4.00 57-24 63-156 The weights of the s several elements in the above six solutions are as follows : Table II. Sb. Br. CI. K. Total. B 21.20 10.18 33-09 20.75 85.22 C 21.20 18.23 29-53 20.75 89.71 D 21.20 34-35 22.38 20.75 98.68 E 21.20 42.41 18.80 20.75 103.16 F 21.20 63.61 9.40 20.75 114.96 G 21.20 80.57 1.88 20.75 124.40 The solutions were allowed to crystallize slowly by standing in a drying oven. In the oven were placed also shallow vessels containing sulphuric acid. The temperature of the oven was kept constant at 35°. At first crystals of potassium chloride and potas- sium bromide alone separated from the solutions. Finally crystals were obtained which on being placed in water showed instant decomposition with liberation of white, flocculent anti- mony oxy-salts. The crystals varied in color from lemon-yellow to colorless, as the proportion of bromine decreased and chlorine increased in the solutions from which they were formed. These solutions showed the same gradation in color. The crystals prepared by using the proportions indicated by Atkinson, SbCh. 3KBr, (E), were tetragonal pyramids, and agreed perfectly with the description of Atkinson's compound given by Solly.' The 1 J. Chem. Soc. 1883, 293. Mixed Double Halides of Antimo7iy and Potassium. 493 crystals from F were greenish yellow and seeming hexagonal plates superimposed ; those from G were bright yellow and showed a well-defined orthorhombic form. On the other hand the crystals from B and C were almost colorless and showed an hexagonal form, while those from D were yellow tetragonal pyramids. Method of Analysis. In a weighed portion of the compound, water was determined by loss of weight on standing over phosphorus pentoxide in a desiccator. The dried residue was then dissolved in water acidu- lated with hydrochloric acid. Into this solution, kept warm, hydrogen sulphide was passed for one hour. During the last fifteen minutes of the passage of the gas the temperature of the solution was raised to gentle boiling. While still hot the solution was tightly corked and allowed to stand one hour. The precipi- tated antimony sulphide was then filtered on to a Gooch crucible, dried for one hour at a temperature of 110° in a current of dry carbon dioxide, and finally heated for two hours at 210°. This effected the conversion of the antimony sulphide from the red form to the black. In the filtrate, potassium was determined by conversion into the sulphate by means of sulphuric acid, the last traces of which were removed by heating in an atmosphere of ammonia. Another weighed portion of the salt was dissolved in water acidulated with tartaric acid. From this solution bromine and chlorine were precipitated as silver salts, dried, weighed, and finally heated in a current of chlorine gas. Analytical Results. I. 0.6230 gram B gave 0.2420 gram Sb2S3 = o.i727 gram Sb; 0.2963 gram K2SO4 =10.1330 gram K. II. 0.3493 gram B gave 0.6417 gram AgBr and AgCl. Trans- formed into 0.6156 gram AgCl, which gives in original substance 0.0469 gram Br and 0.1314 gram CI. III. 0.4986 gram C gave 0.1858 gram Sb2S3=: 0.1326 gram Sb; 0.2253 gram K2SO4 = 0.1011 gram K. IV. 0.3891 gram Cgave 0.6971 gram AgBr and AgCl. Trans- formed into 0.6528 gram AgCl, which gives in original substance 0.0796 gram Br and 0.1261 gram CI. V. 0.3591 gram D gave 0.0178 gram H2O, 0.1115 gram Sb2S3 = 0.0796 gram Sb ; 0.1439 gram K2S04 = 0.0646 gram K. 494 Herty. VI. 0.3044 gram D gave 0,4938 gram AgBr and AgCl. Trans- formed into 0.4333 gram AgCl, which gives in original substance 0.1087 gram Br and 0.0590 gram CI. VII. 0.2181 gram ^gave o.oiio gram H2O, 0.0667 gram SbaSs = 0.0476 gram Sb ; 0.0809 gram KsSOi^ 0.0363 gram K. VIII. 0.9125 gram E gave 1.4619 grams AgBr and AgCl. Transformed into 1.2560 grams AgCl, which gives in original substance 0.3700 gram Br and 0.1465 gram CI. IX. 0.7966 gram -Fgave 0.2277 gram Sb2S8 1=0.1625 gram Sb; 0.2818 gram K2S04=: 0.1265 gram K. X. 0.6642 gram i^gave 1.0704 grams AgBr and AgCl. Trans- formed into 0.8604 gram AgCl, which gives in original substance 0.3774 gram Br and 0.0454 gram CI. XI. 0.5879 gram G gave 0.0429 gram H2O, 0.1491 gram SbaS* = 0.1064 gram Sb; 0.1755 gram K2S04 = 0.0788 gram K. XII. 0.4943 gram G gave 0.7238 gram AgBr and AgCl. Transformed into 0.5578 gram AgCl, which gives in original sub- stance 0,2983 gram AgBr and 0.0056 gram CI. Table III. Sb. Br. Cl. K. Total. 27.72 1343 37.62 21.35 100.12 26.59 20.46 32.41 20.28 99-74 22.17 35-71 1938 17.99 100.21 B C D 4,96 E 5,04 21.82 40.55 16.06 16.65 100.12 F ... 20.40 56.82 6.84 15.88 99-94 G 7.30 18.10 60.35 1. 13 13-39 100,27 From the above table it will be seen that the compounds B, C and /^ which have no water of crystallization are the three which show six-sided, seemingly hexagonal, forms. This lack of water of crystallization in members of a series showing such uniformity in variation in other respects is very peculiar. Specimens oi B, Cand /^were prepared also by allowing solu- tions similar to those described in Table I to stand over sulphuric acid in desiccators at the ordinary temperature of the laboratory. The compounds thus formed were tetragonal and showed varying proportions of water of crystallization. The results of analysis of these compounds calculated to a dry basis corresponded closely with those of the anhydrous compounds above. I have not been Mixed Double Halides of Antimony and Potassium. 495 able to determine what is the cause of the separation of the anhy- drous compound in some cases and the hydrous in others. From the analyses, however, we can conclude that the hexago- nal form is characteristic of the anhydrous compound. In order better to compare the analytical results, all have been reduced to a dry basis. Table IV. Sb. Br. Cl. K. Total. B 27.72 1343 37.62 21.35 ICO. 1 2 C 26.59 20.46 32.41 20.28 99.74 D 23.32 37-57 20.39 18.93 100.21 E 22.98 42.70 16.91 17-53 100.12 F 20.40 56.82 6.84 15-88 99-94 G i9-5i 65.10 1.22 14.44 100.27 From the close analogy of these results to those similarly obtained in the mixed halides of lead and potassium, it seemed fair to suppose that these compounds were mixtures of the simple halo-antimonites, and since Poggiale' states that the salt SbCl3.3KCl crystallizes from a solution in water of the two salts in the proportion indicated in the formula, it seemed reasonable to expect that the compounds would be mixtures of the salts SbCl3.3KCl and SbBr3.3KBr, If, then, we assume E to be such a mixture, the amount of antimony required would be 20.20 per cent., but there was found 22.98 per cent.; further 19.78 per cent, of potassium is required, while there was found only 17.53 P^J^ cent. Evidently this assumption is wrong. To gain further knowledge of the nature of these compounds, it was deemed best to prepare the simple double halides at each end of the above series by mixing together the antimony and potassium salts indicated in the formulas SbCl3.3KCl (Poggiale) and SbBr3.3KBr, dissolving each mixture in hot water and allow- ing to crystallize by sl^nding in an oven at 35°. From the solu- tion of the chlorides there separated a few columnar, colorless crystals. After removing these, there separated in large quantities six-sided crystals, colorless, and closely resembling the salt RbssSbioClss described by Saunders.' The existence of this rubidium salt has been recently confirmed by Wheeler.' The above crystals, labeled A, gave on analysis the following results : > Compt. rend. 30, 1180. ' This Journal 14, 161. 3 Am. J. Sci. [3] 46,273. Vor.. XVI. -37. 496 Herty. I. 0.3302 gram A gave 0.1398 gram SbsSazz: 0.0998 gram Sb ; 0.1667 gram K.iS04 = 0.0748 gram K. II. 0.3150 gram A gave 0.5992 gram AgCl = 0.1482 gram CI. Calculated for Found. KasSbjoCltj. K,SbCI, (Poggiale). Sb 30.22 30.14 26.64 CI 47.05 47.24 47.28 K 22.66 22.62 26.08 99.93 100.00 100.00 The atomic ratio of antimony to potassium is found from the above results of analysis to be Sb : K as 1 : 2.297, giving thus the formula ioSbCl3.23KCl, and not SbCl3.3KCl as stated by Poggiale. From the solution of the bromides there separated at first a few small crystals of potassium bromide; after removing these from the solution, beautiful transparent yellow crystals, possessing a well-defined orthorhombic form, were depos- ited. These crystals were much larger than those formed from the solution of the chlorides. They were labeled //". On analysis the following results were obtained : I. 0,4830 gram //"gave 0.0351 gram H2O ; 0.1207 gram SbsSsn: 0.0861 gram Sb ; 0.1443 gram K2S04 = o.o648 gram K. il- 0-5433 gram ^gave 0.7858 gram AgBr = o.3344 gram Br. Calculated for Found. Calculated for Found. KjaSbioBrja.syHjO. Calculated to dry basis. K„Sb,oBr„. H.O 7.27 7.12 ... Sb 17.83 17-57 19.22 18.92 Br 61.55 62.12 66.38 66.88 K 13.42 100.07 13-19 14.47 14.20 100.00 100.07 100.00 The atomic ratio of antimony to potassium in the dried salt is, Sb: K as 1 : 2.31, giving thus the formula ioSbBr3.23KBr. Knowing now the character of the simple salts formed at each end of the series, and assuming the interniediate compounds to be mixtures of these two, we should find : Mixed Double Halides of Antimony and Potassium, 497 Table V. Antimony. Potassium. Found. Required. Found. Required, B 27.72 27.80 21.35 20.86 C 26.59 26.47 20.28 19.86 D 23.32 23.64 18.93 17.74 E 22.98 22.87 17-53 17-17 F 20.40 20.43 15.88 15.34 G 19.51 19.20 14.44 14-40 Atkinson deduced his formulas for these compounds from determinations of bromine and chlorine alone. That these data are not sufficient for calculating formulas for these compounds will readily be seen from the following : Chlorine per cent, calculated for : K3SbClG= 47.28 KsaSbioCUs =: 47.24 KC1 = 47-54 SbClsr= 47.01. Bromine per cent, calculated for : KaSbBre = 66.91 K23Sb.oBr63 =66.88 KBr = 67.i4 SbBr3 = 66.67. That the compounds are mixtures and not true salts is rendered still more probable by a comparison of the composition of the compounds and the solutions from which they were formed. The composition of the solutions, reduced to a basis of parts in one hundred for the sake of comparison with the percentage compo- sition of the compounds, is as follows: Table VI. Sb. Br. Cl. K. B 24.87 11-95 38.83 2435 C 23.63 20.32 32.92 23-13 D 21.48 34-81 22.68 21.03 E 20.55 41. II 18.23 20.11 F 18.44 55-33 8.18 18.05 G 17.04 6477 I-5I 16.68 That the compounds should show a slightly larger percentage foantimony and smaller percentage of potassium than the solu- 498 Herty. tions from which they were formed is to be expected in the light of the fact already mentioned, that, when the solutions were allowed to stand, potassium bromide and chloride first separated in small quantities, then came the mixed double halides. Again, if we tabulate the variations of the percentages of bromine and chlorine in the series of salts on the one hand, and in the series of solutions on the other, we find : Table VII. Solution. Bromine. Salt. Chlorine. Solution. Salt. B-C 8.37 7-03 5-91 5-21 C-D 14.49 I7.II 10.24 X2.02 D-E 6.30 5-13 445 348 E-F 14.22 14.12 10.05 10.07 F-G 9.44 8.28 6.67 5.62 Effect of Heat upon the Mixtures. Atkinson states that if the compound SbCl3.3KBr.ijH20 be heated to 300°, all antimony is volatilized as halogen salts, while chlorine and bromine remain in the residue combined with potas- sium in equal atomic proportions. No analytical results are given. Hoping to obtain a simple method for the determination of potas- sium in the compounds, and also to gain further evidence as to the character of the compounds experiments were made with B and F. Specimens of these were slowly heated to 300° and kept at this temperature two hours. On reheating for one hour at 300° no further loss of weight occurred. The residues were then dissolved in water, and the halogens determined by precipitation with silver nitrate. In this way results were obtained which showed rough agreement with those required by calculation, if it be assumed that the compounds are mixtures. But on removing the excess of silver from the filtrates, and passing hydrogen sulphide through, a decided precipitate of antimony sulphide was observed, show- ing that though a constant weight had been obtained, there still remained portions of antimony salts in the residue. For this reason the bromine and chlorine results in the residues cannot be used as satisfactory evidence as to the nature of the compounds. In a general way, however, they tend to confirm the view that the compounds are mixtures of the halo-antimonites containing only one halogen. Mixed Double Halides of Antimony and Potassium. 499 The evidence obtained in this investigation, taken in connection with that obtained in the case of the mixed double halides of lead and potassium, seems to show that these compounds of Atkinson are not true salts, but isomorphous mixtures of the simple halo- antimonites. Mr. H. Hillyer is now at work, in this laboratory, upon Pitkin V so-called salt, PtCU.2KBr. I expect to begin work at once upon mixed double halides in which potassium is replaced by such a base, for example, as aniline. In a recent article'' I described certain green tabular crystals formed by dissolving lead iodide in dilute solutions of potassivm bromide and allowing the solutions to crystallize. At the time of the publication of that article, only one type of potassium halo- plumbite was known, namely, KPbXs. Analyses of these green crystals led me to the conclusion that they were mixtures of KPbl3, KPbBrs and PbBre. Since that time Wells' has prepared a salt to which he ascribes the formula KPbsBrs. In a later article* he puts forward the view that the green tabular crystals which I had prepared were mixtures of KPbsBrs and KPb2l5, and not of KPbIs, KPbBra and PbBrz. I must confess that I consider his view much more plausible than mine. So far as agreement between theoretical and analytical results is con- cerned, one view is practically as good as the other. Other reasons, however, tend to show the correctness of his interpreta- tion of the analytical results. In the first place, there is great crystaliographic similarity between the salt KPb2Br6and the green tabular crystals. In the second place, the conception of " free lead bromide " in the crystals necessitates, as Wells observes, the assumption of the isomorphism of lead bromide with KPbIs and KPbBrs. While this is not impossible, yet it is much more im- probable than the assumption of the isomorphism of KPb2Br6and KPb2l5. Of greater importance still is the fact that though repeated efforts were made to obtain crystals containing very small or very large proportions of " free lead bromide," in no case did the proportion fall below 30 per cent, of the weight of the crystal, or exceed 36 per cent. This constancy is hardly to be expected if " free lead bromide " is present in the crystals. University of Georgia, July 13, 1894. 1 J. Am. Cliem. Soc. 1, 472. » This Journal 15, 97. » Am. J. Sci. 1893, 46, 134. ■• Ibid. 1893, 46, 38. 500 Noyes. Contributions from the Chemical Laboratory of the Kose Polytechnic Institute. XI.— CAMPHORIC ACID.' Second Paper. By W. a. Noyes. In the first paper''' a /5-camphoraminic acid, obtained by f'^e action of caustic soda on camphoric imide, was described. 'T have since learned that this compound had been previoi sly described by Hoogewerf and Van Dorp in the Proceedings of/ ■ Dutch Academy of Science for January 27th, 1894. These auth s do not give the melting-points of the two camphoraminic acids, but establish the isomerism of the bodies on the rotatory power and CN on the conversion into two cyan-lauronic acids, C8Hi4<^pQ xj This last name is based on the relation between the formula of these acids and that of the lauronolic acid of Fittig and Woringer.^ At that time these acids might with equal propriety be referred either to lauronolic acid or to the isomeric campholytic acid of Walker.* Since, however, as will be shown in this paper, one of these acids is closely related to campholytic acid while the other is probably closely related to lauronolic acid, it seems desirable to express these facts in the names of the bodies. I would therefore suggest that the following nomenclature be followed for the present. The two carboxyl groups of camphoric acid and the corresponding positions of other groups are distinguished as a and ??. ^ , . . A m^ /CONHs a a-Camphoraminic acid, L8rlH<^pQ tt a , /J-Camphoraminic acid, C8Hu<^pq^tt^ ^, Cyan-lauronic acid,^ C8Hi4<^pQ tt a, Dihydro-cyan-campholytic acid,^ C8Hu<:^/-.j^t" j , ' Communicated to the American Association for the Advancement of Science at the Brook- lyn meeting. "This Journal 16,307. > Ann. Chem. (Liebig) 337, 6. ••J. Chem. Soc. 68,498. * Hoogewerf and Van Dorp : Loc. cit. Camphoric Acid. 501 Amino-lauronic acid/ C8Hi4<^P^ ''t? o, Dihydro-amino-campholytic acid," C8Hi4<;j^tt" o", Campholytic acid, CsHiaCOaH a Lauronolic acid, CsHisCOaH /S. Since, as will be shown below, the camphoraminic acid of Claissen and Manasse^ is the a-acid, we have for camphor the relation CsHu-cc^^q''^ , and, since Walker obtained his campho- lytic acid from the ortho-ethyl ester of camphoric acid, this ester must have the relation CsHi4<|p^^TT^ „° ". Preparation of Camphoric Acid. Considerable experience has been gained in the preparation of camphoric acid, and some of its more common derivatives, and some of the methods used may be of interest to other workers in this field. For the preparation of camphoric acid 150 grams of camphor with 1200 cc. of nitric acid (1.42) and 800 cc. of water were placed in a three-liter flask having a wide test-tube filled with water fitting closely in its neck. The flask was then heated con- tinuously on a steam-bath for 60 to 65 hours, or until a small por- tion of the liquid, after cooling and saturating with ammonia, left but a small amount of camphor undissolved. Instead of evapo- rating the nitric acid as directed in Beilstein's Handbook, it is found better to cool the acid thoroughly and filter ofl" the camphoric acid which separates. The latter is washed with water, the washings being kept separate from the nitric-acid mother-liquors. The camphoric acid weighed 93.2 grams. The nitric-acid mother- liquors measured 1800 cc. and had a specific gravity of about 1.26. To these were added 250 cc. of strong nitric acid (1.42) and 150 grams of camphor, and the oxidation and separation of camphoric acid effected as before. This time 116.7 grams of the acid were obtained. To the nitric-acid mother-liquors, measuring 1900 cc, 400 cc. of strong acid were added, bringing the sp. gr. up to 1.28 to 1.29. With this solution 171 grams of camphor were oxidized, giving 121. 3 ^^rams of camphoric acid. In all 331 grams of camphoric acid were obtained from 471 grams of cam- phor, or 70 per cent, of the weight of camphor used. ■This paper. » This Journal 16, 310. ' Ann. Chem. (Liebig) 274, 71. 502 Noyes. Camphoric Anhydride. — The camphoric acid obtained as described contains a small amount of camphor and possibly some other impurities, but is sufficiently pure for the preparation of the anhydride. For this purpose loo grams of the anKydi-itie are mixed with 75 cc. of acetic anhydride, or perhaps better with 65 cc. of the anhydride and 10 cc. of acetyl chloride, and boiled with an inverted condenser for half an hour. After cooling, the anhy- dride was washed with water and recrystallized from hot alcohol, using the mother-liquors over repeatedly. From 80 to 85 grams of the pure crystallized anhydride were obtained. Camphoric Imide. — 50 grams of finely crystalline camphoric anhydride and 125 cc. of alcohol are placed in a 500-cc. distilling bulb, and the ammonia obtained by heating 75 cc. of strong ammonia (0.90) with a reversed condenser is passed into the mix- ture until all of the anhydride passes into solution and a consider- able amount of the ammonia passes through unabsorbed. The bulb is then connected with a condenser and its contents distilled rapidly over the free flame till the boiling-point of the imide is reached. Just at the close of the distillation a considerable amount of the imide condenses in the tube of the distilling-bulb, and it is best to loosen the cork for fear of explosion. The contents of the bulb is poured immediately into a mortar, and when cold, the solid mass is broken up and dissolved in a cold ten-per cent, solu- tion of caustic soda. After filtering, the imide is precipitated as quickly as possible by a rapid current of carbon dioxide. After filtering off and washing there were obtained 30 grams of the pure imide, or about 60 per cent, of the weight of the anhydride taken. From the alkaline filtrate from the imide, hydrochloric acid pre- cipitates a mixture of camphoric acid and camphoraminic acids. This mixture may be treated with acetic anhydride, and from the resulting product a new portion of imide can be obtained, raising the total yield to perhaps 70 or 75 per cent, of the theoretical. a- Camphoraminic Acid. — 25 grams of camphoric anhydride are placed in a flask with 50 cc. of alcohol, and ammonia is passed in exactly as m the preparation of the imide. The solution is then poured out into a beaker and allowed to stand until the ammo- nium salt of the camphoraminic acid has separated as far as pos- sible. The mother-liquors are then removed as far as possible by means of the pump and the salt dried on a porous plate. The salt is then dissolved in a small amount of water, and the cam- Camphoric Acid. 503 phoraminic acid precipitated by a little more than the theoretical amounL of hydrochloric acid. The acid should be added slowly, and care should be taken to secure the separation of the camphor- aminic acid in the crystalline form by adding a small amount of the crystalline acid. 100 grams of the anhydride gave 51 grams of the acid melting at i67°-i70°. This is sufficiently pure for the preparation of amino-lauronic acid, as the purification of the latter is more easily effected than that of the camphoraminic acid. The alcoholic mother-liquors may be used to prepare the imide as directed above. /?- Camphoraminic Acid. — 25 grams of the imide are dissolved in 80 cc. of 15 per cent, caustic soda, and the solution heated on the water-bath for an hour and a half. On cooling, and especially on the addition of a fragment of the/J-acid, the sodium salt, which is rather difficultly soluble in caustic soda, separates and is filtered off and sucked as dry as possible with the pump. The salt is then dissolved in a small amount of water and the acid precipi- tated with hydrochloric acid, using the same precautions as with the a-acid. The yield is about 15 grams of an acid melting at lyQ^-iSi"* and sufficiently pure for conversion into the dihydro- amino-campholytic acid. The alkaline mother-liquors contain some of the a-acid. Dihydro-ami7io-campholytic Acid. The preparation of this acid from camphoric imide was described in the first paper.' The method there described gives a pure acid. The mother-liquors, however, contain a small amount of the amino-lauronic acid, which can be separated and identified in the form of its chloride, which melts at 303°-305° and crystallizes in a quite different manner from the chloride of the dihydro-amino-campholytic acid. For the preparation of the latter it is better, therefore, to use the comparatively pure /5-cam- phoraminic acid prepared as described above. 19.9 grams (i molecule) of the /3-camphoraminic acid are dis- solved in 100 cc. of lo-per cent, caustic soda. 5.1 cc. or 16 grams of bromine (i molecule) are dissolved by shaking quickly with 140 cc. of lo-per cent, caustic soda, and the solution added to that first mentioned. Considerable heat is evolved by the reaction and the temperature rises spontaneously about 32°. The solution > This Journal 16, 310. 504 Noyes. is then heated on the water-bath to about 75** for 15 to 20 minutes. After coolings, a little sodium sulphite is added, and the solution is neutralized with hydrochloric acid till a few drops develop no color on boiling in a test-tube with phenol-phthalein. The solution is then evaporated and the amino acid separated by crys- tallization, and purified by recrystallization from water. The chloride of dihydro-amino-campholytic acid crystallizes from concentrated solutions, on cooling, in long, thick needles which melt with decomposition at 26i°-262°. The crystals show no tendency to a concentric grouping, and in this respect differ very markedly from those of the chloride of the isomeric amino-lauronic acid. 0.1909 gram of the salt gave 0.1322 gram AgCl, corresponding to 17.13 per cent, of chlorine. Theory requires 17. 11 per cent. The nitrate crystallizes in long needles which meU with decom- position at 2I2°-2I3''. The chloroplatinate is very easily soluble in water, and sepa- rates, on slow evaporation of its solution, in dark orange-yellow plates, sometimes of considerable size. 0.2239 gram of the salt gave 0.0581 gram Pt, or 25.95 per cent. Theory requires 25.93 P^'^ cent. Anhydride of Dihydro-avii7io-campholytic Acid. An attempt was made to obtain an amine by distilling dihydro- amino-campholytic acid with quicklime. The lime acts, instead, as a dehydrating agent, and an inner anhydride of the acid is formed. If a little care is taken to avoid too high a temperature the reaction is almost quantitative, and the nearly pure anhydride condenses in the front part of the tube as a white, waxlike solid. The substance is best purified by crystallization from a very small amount of warm ligroin. The crystals which separated from the concentrated solution on cooling were separated from the mother- liquors by spreading them on a porous plate. The crystallized substance retains it waxlike character, but melts sharply at i88°- 189°. It boils without decomposition at 285°-287°. The substance is indifferent and is very easily soluble in water, alcohol, benzene, chloroform, ether and carbon bisulphide. 0-1596 gram of the substance gave 0.1426 gram H2O and.0.4120 gram CO2. 0.1153 gram gave 0.01065 gram N. Camphoric Acid. 505 Calculated for CeHi,<^°>. j Found. ^^ C ".70-59 7040 H 9.80 9.92 N 9.15 ... 9.24 Attempts to obtain an amine by the distillation of dihydro- amino-campholytic acid with barium hydroxide, with soda-lime, and with sodium methylate, were unsuccessful. In some cases very considerable amounts of permanent gases were formed, and it was evident that the reaction consisted chiefly in a decomposi- tion of the substance in a manner which would not yield any simple derivatives of the acid. Canipholytic Acid.^ When dihydro-amino-campholytic acid is dissolved in the calcu- lated amount of sulphuric acid, and sodium nitrite is added to the solution, decomposition begins at once, accompanied by a rise of temperature. The acid is best separated from the solution by distillation with a current of steam. It partly separates from the distillate in oily drops which are very slightly heavier than water. The acid may be separated from the distillate by extraction with ether, or it may be converted into the sodium salt by means of sodium carbonate and the solution evaporated, apparently without any decomposition. The free acid boils at 240°-243°. When treated with bromine in carbon-disulphide solution the acid yields a dibromide which melts, when heated quickly, at 113°-! 14°. Walker^ gives the melting-point of the dibromide of campholytic acid as iio°, but states that it depends somewhat on the manner of heating, and in a private communication states that in one case a melting-point of 115° was observed. My dibromide shows the same tendency to blacken when* exposed to the light which was observed by Walker. It also decomposes with ammonia or sodium carbonate, forming an oil which is insoluble both in acids and bases. The barium salt remains on evaporation as a sticky mass. The only respect in which my acid appears to differ essen- tially from that of Walker seems to be in that mine gives precipi- tates when zinc chloride or silver nitrate is added to a solution of its sodium salt, while Walker states that the ammonium salt of his 1 Dr. Walker writes me that he is at present interes ted n other matters and has very kindly left the further study of this substance to me, 2 J. Chem. Soc. 63, 499 {1893). 5o6 Noyes. acid gives no precipitate with salts of the heavy metals. The dif- lerence is probably due to his use of the acpmonium salt instead of the sodium. Zinc Campholytate. — This salt was prepared by adding a solu- tion of zinc chloride to a solution of the sodium salt of campho- lytic acid. It separates in a semi-crystalline form, and is very difficultly soluble in water. It is easily soluble in ether. 0.2015 gram of the salt gave 0.0445 gram of zinc oxide. Calculated for (CgH, 3062)220. Found. Zn 17.52 17.72 An attempt to remove four hydrogen atoms from campholytic acid by oxidation with potassium ferricyanide, a method which has been successful with dihydro- and tett^hydro-terephthalic acids,' gave a negative result. The acid appears to remain unchanged. The study of the acid in other directions will be continued. Amino- lauronic Acid."^ ' When a-camphoraminic acid is treated with sodium hypobro- mite, it is converted into amino-lauronic acid in the same manner in which the /3-acid is converted into dihydro-amino-campholytic acid. Instead of exactly neutralizing the solution and attempting to separate the free acid, it is better, however, to render the solu- tion strongly acid with hydrochloric acid and evaporate till the chloride of the acid begins to separate. On cooling most of the chloride will separate from the solution together with some sodium chloride. It may be purified by recrystallization from water. It is very noticeable that in the treatment of a-camphoraminic acid with sodium hypobromite the rise in temperature is very 'Ann. Chem. (Liebig) 358, 2 and 49. 2 la a private communication from Professor W. A. Van Dorp he tells me that he and Professor Hoogewerf have prepared amino-lauronic acid, the work having been done before my first paper appeared. As they have not yet published anything on the subject, they have very kindly left the further study of the acid and its decomposition-products to me. On account of their work, the study of the amino-lauronic acid itself has been carried no further than was necessary in connection with the other work described in this paper. Professor Van Dorp gives the melting-point of the free amino-lauronic acid as 260°, and that of its anhydride as 204°-2o6°, a little higher than found by myself. They have also prepared the following compounds: (C9Hi,N02).^H2PtCl„ (CeH„N02)HCl.AuCls, CsHieAgNOs, C„H,,N02 C6H2(N02)30H, Cs,H,eN02.CeH5C0-melts at 206°. Camphoric Acid. 507 much less than when /?-camphoraminic acid is treated in the same manner. In the case of the /5-acid the rise in temperature is about 32°, while with the a-acid it is only about 15°. Dr. Claissen-very kindly sent me a specimen of the camphoraminic acid obtained by the action of hydrochloric acid on iso-nitroso camphor.' When one gram of this acid was treated with caustic soda and sodium hypobromite in the proportions given, the rise in temper- ature was 13°. From the solution, the chloride of amino-lauronic acid, crystallizing in its characteristic forms and melting at 303°- 305°, was readily obtained. Since in Claissen's acid, unless we suppose the reactions to be accompanied by a deep-seated rearrangement of the molecule, which seems highly improbable, the amido group must be combined with the same carbon atom which is in the methylene group of camphor, these facts establish the correspondence of that methylene group with the amidic group of a-camphoraminic acid, and the correspondence of the carbonyl of camphor with the amidic group of the /3-acid. The chloride of amino-lauronic acid crystallizes in needles which have a strong tendency to a concentric grouping. It appears to be less easily soluble than the isomeric chloride of dihydro-amino-campholytic acid. It melts at 303°-305°, The chloroplatinaie crystallizes in light orange-yellow granular crystals which under a glass are seen to consist of small plates, some of them in well-shaped six-sided forms. The salt is very much less easily soluble in water than that of the isomer. 0.1416 gram of the salt gave 0.0368 gram Pt. Calculated for (C8H„- Found. N 9.15 9.20 >Ann. Chem. (Liebig) 874, 79. 5o8 Noyes. The study of the products of the decomposition of amino- lauronic acid when its chloride is heated with a solution of sodium nitrite is not yet satisfactorily completed, but, as the further study of the subject is likely to take a considerable amount of time, it seems desirable to give here a brief statement of the results already obtained. The products of the decomposition appear to be much more complex than in the case of dihydro-amino-cam- pholytic acid. An unsaturated acid is formed which resembles very closely the lauronolic acid of Fittig and Woringer.' The acid gives with fuming hydrobromic acid a crystalline but extremely unstable hydrobromide. The calcium salt of the acid separates on the surface of its solution during evaporation in dendritic forms, as observed by Woringer. It contains, however, two, instead of three, molecules of water. On account of this dif- ference the two acids cannot be said to be identical, but, until a further comparison has been made, it does not seem besitogive the acid a new name. Attempts to obtain a salt with a different amount of water, by evaporating the solution at a low temperature, were not successful. 0.0638 gram of salt lost 0.0058 gram H2O at 135° and gave 0.0226 gram CaSO*. 0.1270 gram of the salt lost 0.0113 gram H2O at 135°. 0.1074 gram of the salt lost 0.0102 gram H2O at 135° and gave 0.0381 gram CaSO*. Calculated for ound. (C8H,3C0j)3Ca + 2H20. I. II. III. 2H2O 9.42 9.09 8.90 9.50 Ba 10.47 10.42 ... 10.43 Solutions of the sodium salt of the acid give precipitates with zinc chloride and with silver nitrate. The zinc salt is easily soluble in ether. Besides the unsaturated acid there is, in the decomposition of the amino-lauronic acid, an evolution of carbon dioxide and the formation of a hydrocarbon which boils at about 120°, but which has not yet been obtained in sufficient amount for identification. It is probably tetra-hydroxylene. There is also formed a small amount of a body which conducts itself like a lactone and which is probably the campholactone of Woringer. It is insoluble in a solution of sodium carbonate, but dissolves in a solution of barium hydroxide, and is regenerated by acidifying the solution with • Ann. Chem. (Liebig) 327, 7. Camphoric Acid. 509 hydrochloric acid and distilling. It solidifies at a low tempera- ture, but it has not been obtained in sufficient amount to purify the body and obtain a satisfactory melting-point. The experiments described in this paper cannot be said to lead to any positive conclusion with regard to the structure of camphoric acid. They do furnish, however, an interesting confirmation of the conclusions arrived at by Walker.' All of the observations made by myself, as well as those made by Walker, agree best with the supposition that camphoric acid contains the complex — *^CO"H I poIh' According to this formula the carboxyl group — C J. Chem. Soc. 63, 506 (1893). » Wreden : Ber. d. chem. Ges. 6, 1379 \ Wallach : Ibid. 85, 923. 5IO Noyes. The latter of these formulae was proposed by Collie,' and has been referred to as satisfactorily explaining the resnjts of his work, by Walker." So far as his results and my own are concerned, the two formulae are equally satisfactory. In relation to camphor, however, the first formula seems to be more satisfactory. It is closely related to a formula for camphor proposed by Armstrong,^ and more recently referred to by Wallach,* who favors this formula for camphoric acid : H=C ^ =CH2 ^ ch/ Everything points to such a close relationship as this formula supposes to exist between camphor and camphoric acid. Especi- ally the conversion of camphor into iso-nitroso camphor and a-camphoraminic acid,^ and the formation of camphor chinon and camphoric acid from camphor-carboxylic acid,^ are scarcely con- sistent with any considerable rearrangement of the molecule in the conversion of camphor into camphoric acid. Aschan' favors this conclusion. This formula also gives a more satisfactory explanation of the formation of cymene from camphor than most of the formulae which have been proposed. If we suppose the ring to separate between the carbon atoms numbered i and 2, it might easily be reformed between the carbonyl of the camphor and the carbon atom numbered i. Such a formation of rings under the influence of a carbonyl or aldehyde group has been frequently observed of late. If at the same time there occurs a separation between the carbon of the carbonyl group and the carbon atom numbered 3, there would be formed the skeleton of cymene. Aschan" has objected to Bredl's formula for camphoric acid that the bromine of brom-camphoric anhydride cannot be removed, except by reagents which cause the opening of the anhydride ring, while in many other cyclic compounds bromine is easily > Ber. d. chem. Ges. 25, 1116. 2 J. Chem. Soc. 68, 510 (1893). » Ber. d. chem. Ges. 16, 2260. *^ Ibid. 36, 921, 'Claissen and Manasse : Ann. Chem. 374, 71. 'Aschan : Ber. d. chem. Ges. 27, 1448. ' Loc. cit. 8 Ber. d. chem. Ges. 37, 1441. Reduction of Paranitrobenzoic Acid. 511 removed along with the hydrogen combined with an adjacent carbon atom. The same objection, if valid, would apply to the formula here proposed. It seems possible, however, that the presence of the anhydride ring gives such a configuration to the molecule that the removal of bromine and hydrogen with the establishment of a double union is impossible. So far as I am aware, no observations with compounds of known structure which would have a bearing on this question are known. It is hoped that' a further study of some of the compounds described in this paper may lead to more positive evidence of the truth or falsity of these formulae. In conclusion it may be worth while to remark that the rela- tive positions of the carbonyl and methylene groups in Armstrong's formula for camphor are in accordance with conclusions reached in the first part of this paper. A small portion of the experimental work described in this paper was done by Dr. H. H. Ballard and Mr. Elmer Brown, to whom I wish to express my sincere thanks. Tbrre Hautk, June 29, 1894. THE ELECTROLYTIC REDUCTION OF PARANITRO- BENZOIC ACID IN SULPHURIC-ACID SOLUTION. By a. a. Noyes and A. A. Clement. In an article published some time ago' we showed that, when nitrobenzene in sulphuric-acid solution is submitted to the action of electrolytic hydrogen, the resulting product is not aniline, but the sulphonic acid of paramidophenol. That this reaction is characteristic of nitro compounds in general has since been proved by Gattermann,' who investigated a great variety of them — nitro-hydrocarbons, nitro-anilines, and nitro-acids — and found them all to undergo change in a similar way, the reduction of the nitro group being accompanied by a migration of one of its oxygen atoms to the para position. Only those compounds in which the para position is already occupied by another element > Ber. d. chem. Ges. 26, 990. ^Ibid. 86, 1844. Vol. XVI.-38. 512 Noyes and Clement. or group than hydrogen form exceptions to this rule. Of such compounds only two were investigated — paranitroorthotoluidine and paranitrotoluene. The reduction-product of the former con- sisted of diamido-cresol [CH3(i)NH<2)NH<4)OH(5)], a migra- tion of an oxygen atom to the ortho position having taken place. In the second case there was obtained a complex body which was shown' to result from the condensation of the primary reduction- product, amidobenzyl alcohol, with a second molecule of nitro- toluene ; thus, in this case, substitution of hydroxyl in the (methyl) group occupying the para position occurs. What reac- tion takes place seems therefore to depend on the nature of the para group, and the investigation of other para compounds in this direction is of considerable interest. We wish to describe here the reduction-product of one of these, i. e., of paranitrobenzoic acid. We expected to obtain from it amidooxybenzoic acid, or possibly amidobenzoic acid itself; but, remarkably enough, neither of these substances is formed. The paranitrobenzoic acid used was prepared^ by boiling for one hour with return-cooler portions of 20 grams of paranitro- toluene with 60 grams of potassium permanganate dissolved in two liters of water, filtering out the precipitate, concentrating the filtrate, and acidifying it. In this way 18 grams, 73 per cent, of the theoretical yield, were obtained. The apparatus used for the electrolysis consisted of a small beaker, to the sides of which a large platinum electrode was closely fitted, and in which was placed a porous cup containing sulphuric acid and a small platinum electrode to serve as the positive pole. The solution, made by dissolving 12 grams of the nitrobenzoic acid in 100 grams of warm concentrated sulphuric acid, was poured into the beaker, which was surrounded with asbestos to prevent loss of heat, and a current of one ampere passed through for twenty-four hours. To isolate the product, the solution was diluted with water, cooled, and filtered. The precipitate was found to be very little soluble even in boiling water, showing the absence of sulphates of bases. It was treated with sodium-carbonate solution, the residue consisting of sulphur was filtered out, and the filtrate was acidi- fied. A voluminous precipitate separated; it was collected on a filter, and washed, first with water, then with alcohol and ether, to » Ber. d. chem. Ges. 26, aSio. a Ibid. 10, 580. Investigations on the Sulphon- Phihalehis. 513 remove any unchanged nitrobenzoic acid. Four grams of pro- duct were thus obtained. Its properties and analysis showed it to be paramidophenolsulphonic acid. It was only very slightly soluble in boiling water and is insoluble in alcohol and ether. It reduced silver nitrate solution in the cold, with production of a purple color. The solution of its salts turned brown rapidly in the air. It gave the following results on analysis : Calculated for Found.! CeH3.NH3.OH.SO3H. C 37.70 38.09 H 3-75 3.70 SOs 42.40 42.33 To give further proof of the identity of the acid, some of it was heated to 175° in a closed tube with concentrated hydrochloric acid. At this temperature the sulphonic group was split off, as is the case with the paramidophenolsulphonic acid.'' The hydro- chlorate, which crystallized out on cooling, was filtered off, dried, and boiled with a large excess of acetic anhydride. The resulting product after crystallization from water showed a melting-point of 152°, identical with that of diacetylparamidophenol. The electrolytic reduction-product of paranitrobenzoic acid is therefore paramidophenolsulphonic acid, the tendency of the oxygen atom to assume the para position being great enough to expel the carboxyl group originally occupying it. Massachusetts Institute of Tkchnologv, Boston, yune, 1894. INVESTIGATIONS ON THE SULPHON-PHTHALElNS. III.— PHTHALEINS OF ORTHO-SULPHO-PARA-TOLUIC ACID.^ By James A. Lyman. Ortho-sulpho-para-toluic acid was made according to the method described by Randall,' from methyl-saccharin that had 1 0.3964 gram substance gave 0.5479 gram CO^ and 0.1336 gram HjO ; 0.4356 gram substance gave 0.5380 gram BaS04. 2 Ber. d. chem. Ges. 26, 992. ' From the author's thesis, submitted to the Board of University Studies of the Johns Hopkins University for the degree of Doctor of Philosophy, June, 1892. * This Journal 18,256. ^14 Lyman. been placed at the disposal of Prof. Remsen by the " Badische Anilin- und Soda-fabrik." The methyl-saccharin in portions of 25 grams was boiled in a flask fitted with a reflux condenser with a mixture of 100 cc. of concentrated hydrochloric acid and 500 cc. of water. After boiling for two hours, all the sulphinide had been transformed, and the solution was then decolorized and evaporated to crystallization. The acid ammonium salt may be converted by treatment with phosphorus pentachloride directly into the chloride of the acid ; but as some of the chlor-phosphuret of nitrogen, PsNsClc, is always formed,' it was found best first to make the neutral potas- sium salt and to make the chloride from this. By boiling the chloride with water, ortho-sulpho-para-toluic acid is formed. In the preparation of the phthaleins care must be taken that the ortho-sulpho-para-toluic acid used is anhydrous, as the presence of water seems to promote the formation of tar-like substances. The acid is best dehydrated by long heating below 135°; it should not be allowed to melt. The best method of analysis of the products was found to be the burning in a platinum boat in a stream of oxygen, the front half of the tube being filled with chromate mixture. Action of Phenol. Phenol-para-ineihyl-s2ilphon-phthale'in. — Carefully dehydrated ortho-sulpho-para-toluic acid and dry, crystallized phenol in the proper molecular proportion (i to 2) are heated together in a glass tube closed at one end. The glass tubing is used in prefer- ence to a test-tube on account of its greater strength. The tube is held in a sulphuric-acid bath by a clamp. It should be short enough to allow a ready escape of the water formed, without danger of the contents frothing over the top. As heat is applied to the bath, the contents of the tube melt to a colorless liquid, which grows pink as the temperature rises and finally becomes dark red at 150°, at which point it becomes opaque, and a rapid giving off of water begins. More heat is applied, until the ther- mometer used in stirring the contents of the tube indicates 170°, at which point the reaction goes on most satisfactorily. After several hours the mass takes on a green luster, and the ther- mometer leaves marks on the sides of the tube, as though from ' White: Dissertation, 1891. Investigations on the Stdphon-Phthalelns. 515 the crushing of some crystalline body. Careful examination, however, failed to reveal crystals, and the marks presently dis- appear. The evolution of water slackens as the mass grows viscous, and after heating for fifteen hours the contents of the tube are nearly solid. The thermometer is then taken out and the tube allowed to cool, when the contents usually break loose and can be readily removed. When the reaction has proceeded properly the product is a black, resinous solid showing a green luster. It was powdered and washed with hot water to remove any phenol and ortho-sul- pho-para-toluic acid remaining. Next it was dissolved in caustic potash, precipitated from its hot solution with hydrochloric acid, and washed with hot water until no test for chlorine was obtained with silver nitrate. The precipitate becomes more compact on boiling with water. It was dried by long standing in a desiccator. Analogy with phenol-phthalein would lead to the supposition that a similar compound had been formed, having this structure : C0H4OH CeH^OH LO -i The results of the analyses are in accordance with this suppo- sition. I. 0.5552 gram substance gave 0.3207 gram BaS04. II. 0.2080 gram substance gave 0.1197 gram BaS04. III. 0.1790 gram substance gave 0.0809 gram H2O and 0.3724 gram CO 2. IV. 0.1992 gram substance gave 0.0881 gram H2O and 0.4154 gram CO 2. Calculated for Found. CjoHisOeS + sHjO. I. II. IH- IV. c 56.87 56.75 56.88 H 5.21 5.02 4.91 S 7.59 7-94 7.91 An attempt was made to determine the water given off on heat- ing, but when a loss had taken place closely approximating that required by the above formula, traces of hydrogen sulphide were given off and the attempt was abandoned. This phenomenon has been noted by White in his investigation of sulphon-fluores- cein.' 1 Dissertation, 1891. Calculated for c 60.75 H 4.81 S 8.10 516 Lyman. Pure di-phenol-para-methyl-sulphon-phthalein is not always obtained by the process just described, and an attempt to purify a mixed product was made, based on the solubility in water. After thorough washing the substances were treated with large quantities of boiling water ; the hot filtered solution was evapo- rated to dryness on the water-bath, and a product was obtained having the same properties as the pure di-phenol derivative. Analysis showed their identity, except that the product last obtained contained less water, as might be expected from its manner of preparation. I. 0.2641 gram substance gave 0.1126 gram H2O and 0.5854 gram COs. II. 0.4167 gram substance gave 0.2540 gram BaS04. Found. I. II. 60.45 4.74 8.37 Phenol-para-methyl-sulphon-phthalein is a dark-red, amorphous powder, slightly soluble in alcohol, ether and chloroform, and in warm glacial acetic acid. When these solvents are heated, a larger quantity dissolves, which on cooling comes down as a flaky precipitate. In alkaline carbonates, more readily in caustic alka- lies, it dissolves to a red liquid much like an alkaline solution of phenol-phthalein. It shows a striking purple-red in dilute solu- tion, which changes to yellow on the addition of acids. The purple color is not aifected by exposure to the air, even for a long time. All attempts to obtain a crystalline product by the use of the ordinary solvents were failures, but the use of a solution in water containing a little phenol was attended with better success. A clear, filtered solution was obtained, and, as nothing was deposited, the vessel containing it was covered with a watch-glass and allowed to stand in a sheltered place for several weeks. After some time, examination showed the presence of groups of short, yellow prisms floating from the surface of the liquid. Further evapora- tion caused these groups to grow, though the crystals always remained very small. When exposed to the air they took on a green luster. Their crystal character was not examined, as it was impossible to keep them without change after their removal from Investigations on the Sulphon- Phthale'ins. 517 the solution. The amount was too small for complete analysis. So far as is known, this is the only case of a crystalline sulphon- phthalein yet noted. Action of bromine on phenol-para-methyl-sulphon-phthalein. — The substance was treated with the theoretical amount of bromine required to form a di-brom substitution-product. The bromine was used as a twenty-per cent, solution in glacial acetic acid, and was poured upon the sulphon-phthalein suspended in glacial acetic acid in a small Erlenmeyer flask. Reaction took place at once, accompanied by the production of heat and the evolution of hydrobromic acid. After standing a short time the whole was heated for half an hour on the water-bath and filtered through glass-wool. A small residue of unchanged phthalein was left on the filter. As the addition of water failed to produce any con- siderable precipitate, the solution was evaporated to dryness on the water-bath, and the residue was powdered and placed in a desiccator over caustic potash. The product thus obtained was a light brown powder, soluble to a considerable extent in water. In alkalies it dissolves with a deep-blue color, appearing greenish when dilute. The solution stains the skin a persistent blue which , acids change to yellow. A determination of bromine gave the following result : Calculated for Calculated for CaoHuBrjOgS. CjoHisBraOsS. Found. Br 32.32 39.59 35.99 A second preparation was analyzed for bromine and sulphur. I. 0.2152 gram substance gave 0.1954 gram AgBr. II. 0.2400 gram substance gave 0.0801 gram BaSO*. Br Calculated for CjoHigBrsOgS. 39-59 Found. I. II. 38.64 S 5-29 4-5< It seems, then, that a tri-brom substitution-product is formed ; but, owing to the difficulty of separating it from the unchanged phthalein and from other substitution-products, it was not obtained in a perfectly pure form. Treatment with other propor- tions of bromine gave similar impure tri-brom derivatives. Action of phosphorus pentachloride on phenol-para-methyl- sulphon- phthalein. — When this substance is treated in a small Erlenmeyer flask with a small excess of phosphorus pentachloride, 5i8 Lyman. slight reaction takes place at once, and the mass becomes some- what coherent, but after this first action nothing further takes place. Accordingly the flask was heated to 130° in a sulphuric- acid bath, when the mass melted and a lively effervescence followed. The heating was continued until all action had ceased. The product was a dark-red, viscous mass that when treated with water fell to an insoluble pink powder. The most natural expectation would be that a product had been obtained with this structure, f C0H4CI OH4CI [O -I The substance obtained, however, does not exhibit the properties of such a compound. It does not go into solution even when boiled with strong caustic potash, but is apparently subjected to a deeper decomposition, forming a turbid liquid with the potash, and giving off a phenol-like odor. Analysis of a specimen gave results that could not be readily interpreted, and, considering the prop- erties of the substance, it is doubtful whether it was pure. When ortho-sulpho-para-toluic acid is heated with four and with six molecules of phenol, products are obtained that appear to contain four and six residues of phenol respectively. The evidence as to their composition is, however, not conclusive. Action of Pyrocatechin. To make a pyrocatechin-sulphon-phthalein, one equivalent of ortho-sulpho-para-toluic acid was heated with two equivalents of pyrocatechin in the usual way. Vigorous reaction took place, accompanied by a copious evolution of water, and the whole mass soon became viscous and clung to the sides of the tube. After two hours of heating at a temperature of 165° the reaction was com- pleted, the substance being a black, friable mass. With water it became gelatinous and entirely prevented filtration. It was accord- ingly evaporated to dryness and powdered, when it still presented some of the same difficulties; but the washing was finally accom- plished with a considerable loss of substance. Analysis showed surprising results. I. 0.1000 gram substance gave 0.0203 gram BaS04. Investigations on the Sulphon-Phthale'ins. 519 II. 0.1280 gram substance gave 0.0246 gram BaS04. III. 0.1 172 gram substance gave 0.0468 gram H2O and 0.2897 gram CO2. IV. 0.1477 gram substance gave 0.0632 gram H2O and 0.3569 gram CO2. Calculated for C,„H,eO,S. Q, (2 mol. pyrocatechin.) C 60.00 Calculated for ,H3„0,„S+2H,0. (6 mol.) 67.17 Calculated for C56H380„S+3>iH (8 mol.) 67.40 .0.^ Found. 11. in. IV. . . 67.42 67.20 H 4.00 4.32 4.51 .. 4-44 4-75 S 8.00 4.07 3.21 2.79 2.73 .. The results of analysis seem to agree best with the require- ments of a compound containg eight residues of pyrocatechin. Though the percentages of carbon and hydrogen agree nearly as well for a hexa-pyrocatechin derivative, the percentages of sulphur are much nearer the requirements of the former compound. This formation of an octo-derivative would not be an isolated case; a similar compound of resorcinol will presently be described, and its character can be better discussed then. But it is unex- pected that such a compound should be formed with the propor- tions used. It is to be regretted that the loss involved in its preparation left so little of the substance that its more detailed study was impossible. The compound is a black powder, dis- solving in alkalies with an evanescent, grass-green color. Action of Resorcinol. According to Fahlberg and Barge,' when ortho-sulpho-benzoic •acid and resorcinol are heated in the molecular proportions of one to two or one to four, a condensation-product is obtained containing four resorcinol residues. White found that when the proportion of one to two was used a tetra-product resulted ; but that when the proportions of one to four or one to six were used, a product was obtained containing six resorcinol residues, the conclusion being supported by analyses of a number of specimens. A series of three resorcinol-sulphon-phthaleins was thus obtained. The resorcinol-phthaleins of ortho-sulpho-para-toluic acid have been treated by Jones.'* Para-methyl-sulphon-fluorescein was made by heating para-methyl-dioxy-benzoyl-benzene-sulphonic acid, and chlorine and bromine compounds were obtained similar to those obtained with sulphon-fluorescein. It was also found 1 Ber. d. chem. Ges. 23, 765. 2 Dissertation, 1891. 520 Lyman. that heating together resorcinol and ortho-sulpho-para-toluic acid in the proportion of two to one gave a tetra-resorcinol product, while the*use of the proportions four to one and six to one gave a compound containing six residues of resorcinol. A series of three melhyl-sulphon-phthaleins was thus obtained. As it seemed desirable to investigate these latter compounds further, the follow- ing work was done. Treatment of ortho-sulpho-para-toluic acid with various pro- portions of resorcinol. — One molecule of ortho-sulpho-para-toluic acid was heated with two molecules of resorcinol. The resorcinol melted and dissolved the acid, forming a red liquid. Reaction began at 150°, and the mass was allowed to remain at 170° for about two hours. It gradually grew stiff, and on cooling was a. brittle, black solid, giving a dark-red powder. It was purified in the usual way. In water it dissolved slightly with a yellow color, and with alkalies it gave the characteristic red-green fluorescence. A determination of sulphur was made. Calculated for Calculated for C2oH,409S4-H20. CssH-ijOsS-fsHjO. Found. I2 mol.) u mol.) I. n. S 8.39 5.17 6.16 5.86 The results indicate that a tetra-resorcinol derivative had been obtained. When ortho-sulpho-para-toluic acid and resorcinol in the pro- portion of one to four were heated together, reaction was com- pleted in about two hours, the product being an exceedingly britde, black substance having properties similar to those of the substance just described. Determinations of sulphur in the washed product were made. I. 0.2028 gram substance gave 0.0634 gram BaS04. II. 0.3122 gram substance gave 0.0973 gram BaSO*. Calculated for Calculated for Found. C44H3oO,oS.4H20. C44H3oO,oS.3H30. I, 11. S 3.90 3.98 4.30 4.28 A fairly pure hexa-resorcinol derivative was obtained ; but the substance was obtained in a better form by the use of other pro- portions. Hexa-resorcinol-para-methyl-sulphon-phthalem. — Ortho-sul- pho-para-toluic acid and resorcinol in the molecular proportions of one to six were heated in a short tube. At 150° vigorous reac- Investigations on the Sulphon-Phthale'ins. 521 tion began, and the temperature was allowed to rise very slowly with constant stirring by the thermometer until 170° was indi- cated. At the end of five hours the almost solid mass was allowed to cool. It was, like the other resorcinol compounds, a brittle, black mass showing electrical properties when rubbed. When powdered and placed in water, it softens and then grows hard again, forming small porous grains. It was twice dissolved in caustic potash, precipitated with hydrochloric acid and washed on the filter. The product is a dark-red powder rather difficult to handle when dry, as it becomes electrified when stirred. It dissolves in water to a slight extent, giving a light-yellow solution with a tendency to fluorescence. When boiled with water it becomes more compact. It dissolves at once in alkalies, less readily in alkaline carbonates, giving a red solution showing a strong green fluorescence. When diluted it shows a pink-green fluorescence. On standing in alkaline solution for several days, a slight reddish precipitate settles and the color of the solution becomes less intense. Zinc dust and caustic potash reduce it to a yellow solu- tion ; this when filtered from the zinc dust quickly oxidizes again. It is slightly soluble in alcohol and ether, and in warm glacial acetic acid. Hot water containing resorcinol dissolves a considerable quantity, which on cooling separates as an impalpable powder, rendering the liquid turbid. Attempts to crystallize the substance were unsuccessful. The results of analysis are as follows : I. 0.2928 gram substance gave 0.0830 gram BaS04. II. 0.2197 gram substance gave 0.0622 gram BaS04. III. 0.1507 gram substance gave 0.0640 gram H2O and 0.3623 gram CO2. IV. 0.1668 gram substance gave 0.0703 gram H2O and 0.401 1 gram CO2. C44H3oO,oS.4H20. C44H30O.0S.3H2O. orC44H3aO,3S. Fou nd. (Jones). I. II. III. IV. c 64.21 65.67 ... ... 65-56 65-59 H 4-63 4.48 ... 4.71 4.67 S 3-90 3-98 3-90 3-89 ... ... Whether the three molecules of so-called water exist as such or as hydroxyl groups in the resorcinol residues, could not be deter- 522 Lyman. mined. Attempts to drive off the water by heating were unsuc- cessful, the substance melting at 200° before a loss corresponding to three molecules of water had occurred. No trace of hydrogen sulphide was detected. Aciion of eight molecules of resorcinol on ortho-sulpho-para- ioluic acid. — Ortho-sulpho-para-toluic acid and resorcinol, one equivalent of the former to eight of the latter, were heated in the usual manner in a sulphuric-acid bath. Reaction took place readily, about ten hours being required for its completion. At the end of that time the mass had become too stiff to be stirred with the thermometer, and was accordingly allowed to cool, when it broke loose from the sides of the tube and could be obtained in the form of a stick. When rubbed it exhibited electrical proper- ties, and when powdered and thrown upon water it at first soft- ened and then hardened to a friable mass. When the stick was broken it showed a radiating structure, with curved surfaces. It was powdered and washed with warm water and then twice dissolved in alkali and reprecipitated with hydrochloric acid, being each time filtered and washed. It was finally dried for a long time in a desiccator. Two determinations of sulphur were made. I. 0.2485 gram substance gave 0.0594 gram BaS04. II. 0.2836 gram substance gave 0.0658 gram BaS04. Calculated for Calculated for C44H3eO,3S. C6.H4eO,8S. Found. (6 mol. resorcinol.) (8 mol. resorcinol.) I. II. S 3.98 3.18 3.29 3.19 As only a small quantity of the compound had been made, a complete analysis was not attempted ; accordingly a second prep- aration was made, using larger quantities. This product was in every way similar to the one first obtained. I. 0.1727 gram substance gave 0.0696 gram H2O and 0.4238 gram CO2. II. 0.1660 gram substance gave 0.0677 gram H2O and 0.4065 gram CO2. III. 0.1547 gram substance gave 0.0628 gram H20and 0.3805 gram CO2. IV. 0.1298 gram substance gave 0.0524 gram H2O and 0.3174 gram CO2. V. 0.2197 gram substance gave 0.0495 gram BaS04. Investigations on the Suiphon-Phthale'ins. 523 Calculated for Found. C56H490,6S. I. II. III. IV. c 66.80 66.86 66.75 67.07 66.69 H 4-57 4.48 4-53 4-52 4.49 S 3.18 ... ... ... ... 3.10 A third product, similar to the other two, was made and analyzed. I. 0.1543 gram substance gave 0.0625 gram H2O and 0.3774 gram CO2. II. 0.2426 gram substance gave 0.0570 gram BaSO*. III. 0.2629 gram substance gave 0.0626 gram BaS04. Calculated for Found. C68H4eO,8S. I. II. III. C 66.80 66.71 H 4.57 4.50 S 3.18 ... 3.23 3.27 The following is the average of all the determinations made, compared with the composition of various phthaleins. Average CB6H460,e|6. C44HsoO,oS.3H50. C44HgoO,oS.2HjO. Found. (8 mol.) (6 mol.) (6 mol.) C 66.81 66.80 65.67 67.17 H 4.50 4.57 4.48 4.32 S 3.21 3.18 3.98 4.07 While the evidence in favor of the formula CseH^OieS is, there- fore, strong, it seems nevertheless premature to speculate in regard to the structure of this curious product. Action of bromine on octo-resorchiol-para-methyl-suiphon- phthaleln. — Two grams of the phthale'in were treated in a small Erlenmeyer flask with a large excess of bromine in the form of a twenty-per cent, solution in glacial acetic acid. Heat was at once developed. After about one hour's gentle heating on the water- bath, with the occasional addition of bromine, solution had taken place, giving a dark-red liquid, rom which, on cooling, amorphous crusts were deposited. The addition of water threw down a pink, finely-divided precipitate, very hard to wash. The heated liquid was accordingly made alkaline with caustic potash, and the sub- stance in solution was precipitated by hydrochloric acid as a red precipitate resembling ferric hydroxide. A determination of bro- mine gave Br : 47.60. A substituted octo-resorcinol-para-methyl- sulphon-phthalein of the formula CeeHsBBmOieS would give 46.8 per cent, of bromine. 224 Lyman. A further endeavor to obtain a pure product was made. The same process was followed, the heating on the water-bath being succeeded by heating in a sulphuric-acid bath to the boiling- point of the acetic acid. A compound similar to the first was formed, which gave the following results on analysis : I. 0.1540 gram substance gave 0.0295 gram HsO and 0.1943 gram CO 2. II. 0.217 1 gram substance gave 0.2507 gram AgBr. Found. Calculated for C56H34Br„0,eS. I. c 3444 3442 H 1.74 2.13 Br 49.00 ... 49.14 It seems that bromine readily enters the compound at first, but that this action takes place less readily as more and more enters, until prolonged action is needed to force the eleventh and twelfth atoms into combination. These substances are very similar to eosin in their properties. They dissolve readily in alkalies, giving a deep-red color when concentrated. When diluted the solution shows the delicate fluorescence of tetra-brom-fluorescein. Action of Orcinol. The action of orcinol upon ortho-sulpho-benzoic acid has been investigated by Mr. J. E. Gilpin, the results being published in his dissertation (1892). He obtained a di-orcinol and a tetra-orcinol derivative of the acid, orcinol showing less readiness than resorcinol to enter into combination. One equivalent of ortho-sulpho-para-toluic acid was heated with two equivalents of orcinol. The orcinol melted and dissolved the acid, but reaction was somewhat slow in taking place, being com- pleted only after ten hours' heating at 170°. The resulting solid mass was powdered and washed with hot water, with which it does not become gummy. It was then treated in the usual manner. I. 0.1774 gram substance gave 0.0734 gram H2O and 0.4197 gram CO a. II. 0.1747 gram substance gave 0.0744 gram H2O and 0.4118 gram CO 3. III. 0.2068 gram substance gave o.iioo gram BaSOi. Investigations on the Sulphon-Piithale'ins. 525 Calculated for C^jHigOeS. (2 mol. orcinol.) 64-39 I. 64.52 Found. II. 64.28 III. 4-39 4.60 4-73 ... 7.80 7.31 ance seems to be a di i-orcinol derivative containing, perhaps, a little of a higher phthalein from which it could not well be separated. It is a bright-red powder, dissolving in alkalies with a red color, and showing a green fluorescence considerably weaker than that of resorcinol phthaleins. Action of Hydroqtdnone. Reaction between hydroquinone and ortho-sulpho-para-toluic acid does not readily take place, a compound being formed only after long-continued heating. When the two substances in the molecular proportion of two to one were heated in the usual manner a clear liquid was formed at 170°. At a temperature of 185° slight coloration began, accompanied by a gradual evolution of water. After about fifty hours an opaque viscous mass resulted, which on cooling formed a brittle, black solid. It was purified in the usual manner, and analyzed with the following result : I. 0.1643 gram substance gave 0.0586 gram BaS04. II. 0.1676 gram substance gave 0.0601 gram BaS04. III. 0.1750 gram substance gave 0.0670 gram BaS04, IV. 0.1617 gram substance gave 0.0582 gram H2O and 0.3928 gram COs. V. 0.1624 gram substance gave 0.0577 gram H2O and 0.3935 gram CO 2. Calculated for (2 mol. hydroquinone.) Calculated for Ca^Hs^O^S.HjO. (4 mol. hydroquinone.) c 63.09 65-75 H 3.66 4.12 S 8.38 5-47 rouna. III. IV. V. ... 66.25 66.17 ... 4.00 3-95 5.26 ... ... 4.90 4.93 It thus appears that a di-hydroquinone product is not formed, but, instead, one containing four residues of the phenol. Tetra-hydroquinone-para-methyl-sulphon-phthalein is a com- pact brown powder soluble in a large quantity of water with a yellow color. In alkalies it readily dissolves with a red color. Zinc dust and caustic potash reduce it to a nearly colorless substance. 526 Lyman, Action of Pyrogallol. An experiment is described by White,' in which, on heating together pyrogallol and ortho-sulpho-benzoic acid, he obtained a substance having properties hke those of gallein ; the compound was not analyzed. Attempts, described below, to obtain a methyl- sulphon-analogue of gallein were successful, resulting in the for- mation of at least two phthaleins. Par a-77iethy I- sulpJj on- gallein, — This compound was formed by heating together ortho-sulpho-para-toluic acid and pyrogallol in the molecular proportion of one to two. The clear liquid formed when these melted grew darker, and at i6o° vigorous reaction took place ; to prevent the fused mass from rising to the mouth of the tube, continuous stirring was necessary. After two hours' heating, the product was a stiff substance that clung to the sides of the tube. On cooling, it formed a brittle mass that when exposed to the air for any length of time became sticky. Accordingly it was quickly powdered and mixed with water, when it dissolved to a large extent, giving a brown solution. Attempts to filter off the undissolved substance being unsuccessful, as the paper became packed with the gelatinous mass, the liquid was treated with caustic potash, and then acidified with hydrochloric acid, but attempts to filter met with no better success. The acidified liquid was then evaporated to dryness on the water-bath and the residue powdered. It now presented more agreeable characteristics : it did not become sticky in the air, and, though still rather easily soluble in water, the part not dissolved sank to the bottom of the beaker and was easily filtered off and washed. The dried product gave the following analytical results: I. 0.1298 gram substance gave 0.0435 gram H2O and 0,2649 gram CO2. II. 0.1274 gram substance gave 0.0426 gram H2O and 0.2605 gram CO2. III. 0.1588 gram substance gave 0,0820 gram BaS04 (Liebig). IV. 0.1477 gram substance gave 0.0749 gram BaS04 (Liebig). V. 0.1361 gram substance gave 0.0701 gram BaS04 (Liebig). 7.03 7,08 Calculated for CsoHj^OsS. (/-methyl-sulphon-gallein; 1. I. Found, II. III. C 55-55 55-67 55-76 ... H 3.70 3-72 3.72 ... S 7.40 ... 7.IC •Dissertation, 1891. Investigations on the Suiphoji- Phthale'ins. 527 The fact that the percentages of carbon are high, while sulphur comes persistently low, would indicate that the substance analyzed may have contained a small quantity of a higher phthalein; this would cause the variation noted from the theoretical percentages. Para-methyl-sulphon-gallein is a blue-black powder, rather easily soluble in water, its dilute solution having a yellow color. If a drop of alkali is allowed to fall upon such a solution, a beau- tiful sky-blue color is produced. On stirring, and especially in more concentrated solutions, the blue is soon succeeded by a rich purple, often so deep as to appear black. An excess of acid brings back a yellowish-brown color, but the precipitate formed will not settle. When allowed to stand in the air, the alkaline solution gradually fades and becomes almost colorless, and a brown sub- stance is deposited. The addition of a little stannous chloride makes the blue color more persistent. Hexa-pyrogallol-para-methyl-sulphon-galle'in. — One equivalent of ortho-sulpho-para-toluic acid was heated with four equivalents of pyrogallol. Reaction was even more vigorous than in the pre- ceding case, and after an hour's heating at 160° a solid mass resulted. On cooling it was brittle, black and quickly became sticky. It presented in an exaggerated form the difficulties encount- ered in the purification of methyl-sulphon-gallein, but after being evaporated to dryness it was filtered and washed, a considerable loss resulting from its solubility. The most probable product from the reaction of the propor- tions used was a compound containing four residues of pyro- gallol. If, however, the reaction should take place as with resor- cinol, a substance containing six such residuesmight be expected. Analysis showed that the latter was the case. I. 0.1812 gram substance gave 0.0644 gram H2O and 0.4001 gram CO2. II. 0.1740 gram substance gave 0.0615 gram H2O and 0.3850 gram CO2. III. 0.2340 gram substance gave 0.0639 gram BaSO*. IV. 0.3566 gram substance gave 0.0995 gram BaS04. 3.83 Calculated for C35H,50„S C, (4 mol. pyrogallol.) Calculated for ,4H3„0,eS.i%H,0. (6 mol.) I. Found. II. III. c 60.90 60.47 60.22 60.34 ... H 3-49 3-78 3-94 3.92 ... S 5.08 Vol. XVI. -39. 3.66 ... ••• 3.7, 528 Gilpin. Hexa-pyrogallol-para-methyl-sulphon-phthalein is a black pow- der rather easily soluble in water, and giving a rose color when such a solution is diluted. A drop of alkali causes a beautiful, evanescent sky-blue color. If some stannous chloride is added to a water solution of the phthalein and the alkali is then added, the blue comes out beautifully and remains for some time. After the disappearance of the blue, a purple takes its place, slowly fading to a brown and finally to a yellow. After standing for a week, even in sealed tubes, all color disappears. Zinc and caustic potash give a yellow solution. With most of the phthaleins described, attempts were made by treatment with bromine, with acetic anhydride, and with phos- phorus pentachloride to get definite products, in the hope of throwing light upon the nature of the phthaleins, but these attempts did not lead to satisfactory results. IV.-ORCIN-SULPHON-PHTHALEIN. By J. E. Gilpin. Orcin and ortho-sulpho-benzoic acid (2 molecules of the orcin to I of the acid) were well mixed and put in a short tube which was suspended in a sulphuric-acid bath. A thermometer was used as a stirring-rod in order that the temperature might be easily regulated. The temperature was slowly raised to 80°, at which point the orcin melted to a straw-yellow colored liquid. The acid dissolved in this as the temperature rose, and at i6o°-i70° water began to come off. Constant stirring, especially in the early stages of the action, is necessary, as in many cases it would not be complete without it. The color gradually turned to dark red, and after the mass had become slightly stiff it was heated to 180° until it was too stiff to stir. If the tube is plunged into cold water while still warm, the material on cooling suddenly breaks away from the glass and can be extracted without difficulty. After being powdered it was washed well with the aid of the filter-pump, to remove any excess of the orcin or acid, and then dissolved in alkali. The sub- stance is obtained from this solution by adding hydrochloric acid, when it is precipitated as a flocculent mass. It was found much easier to wash this precipitate if the liquid was first evaporated Orcin-Sidphon- Phthale'in. 529 off and the mixture of the compound and sodium chloride well powdered. This was washed until the sodium chloride was removed and the wash-water showed no marked color, then dried on porous plates and dryinor-paper. The powder prepared in this way is of a dark-red color, easily soluble in alkali, with a marked red and green fluorescence; nearly insoluble in water, and slightly soluble in alcohol and ether. On evaporating the solu- tions, however, it always comes down as a powder and shows no tendency to crystallize. Four specimens were made in this way, using the same proportions; but only one of these was approxi- mately pure. The general tendency seem to be towards the formation of the di- and tetra-products mixed, whether the orcin is present in the proportions of two or four molecules. Fischer,' in his work on the action of orcin on phthalic-acid anhydride, found that one of the oxygen atoms was replaced by two orcin residues, water being eliminated in the reaction. It was expected that with the ortho-sulpho-benzoic acid a similar action would take place, leading to the formation of a diorcin derivative of ortho-sulpho-benzoic acid. From the work on this product it can hardly be doubted that diorcin-sulphon-phthalein is formed ; but generally there is some tetraorcin-sulphon-phtha- . lein farmed at the same time. As Baeyer showed that in the action of resorcin on phthalic anhydride it was not the anhydride oxygen which was replaced but one of the others, we should expect the same to hold good here. Preparatio7i I. 0.131 gram of the substance gave 0.0698 gram BaSO* (Liebig). 0.1752 gram of the substance gave 0.0707 gram H2O and 0-3875 gram CO2. Preparation II. 0.1376 gram of the substance gave 0.0692 gram BaS04. 0.1305 gram of the substance gave 0.0613 gram BaS04. 0.14 1 1 gram of the substance gave 0.0686 gram BaS04. 0.1009 gram of the substance gave 0.0475 gram BaS04, 0.1682 gram of the substance gave 0.0693 gram H2O and 0.3730 gram CO2. 0.2992 gram of the substance gave o. 1184 gram H2O and 0.6725 gram CO2. ' Ber. d. chem. Ges. 7, 1214. 530 Ullmann. Preparation III. 0.1826 gram of the substance gave 0.1048 gram BaS04. 0.145 1 gram of the substance gave 0.0634 gram H2O and 0.3340 gram CO 2. 0.1703 gram of the substance gave 0.0715 gram HaO and 0.3921 gram CO2. Preparation I V. "■ 0.1 1 28 gram of the substance gave 0.0571 gram BaSOi. Found. Calculated for I. II. III. IV. C,,H,80,S. C 60.86 60.31 60.47 61.29 62.77 62.78 ... H 4.34 4.48 4.57 4.43 4.82 4.66 ... S 7.72 7.32 6.91 6.45 6.68 6.46 7.88 ... 6.95 The first of these was comparatively pure diorcin-sulphon- phthalein. The others, however, contained amounts of carbon and sulphur which would correspond to mixtures of the di- and tetraorcin-sulphon-phthalein, varying according to the proportions of each present, but in general showing a greater tendency towards the tetraorcin-sulphon-phthalein. From this it was thought that a pure tetra-product could be obtained by the use of more orcin in proportion to the acid, but it was found that this was not the case. Experiment showed that when the acid is heated with orcin in the proper proportions an approximately pure tetra-product was formed, but it could not be obtained in pure condition. ON PARA-CHLOR-META-SULPHO-BENZOIC ACID AND SOME OF ITS DERIVATIVES.' By H. M. Ullmann. Para-chlor-benzoic acid from commercial saccharin. — My attention was directed to /-chlor-benzoic acid by observing its • From the author's thesis, submitted to the Board of University Studies of the Johns Hop- kins University for the degree of Doctor of Philosophy, June, 1892. Para-chlor-meia-sulpho-benzoic Acid. 531 occurrence in commercial saccharin. In the preparation of (3-sulpho-benzoic acid from the benzoic sulphinide contained in sac- charin, the latter substance is boiled for several hours with dilute hydrochloric acid in a flask connected with a reflux condenser ; the benzoic sulphinide is thus converted into the acid ammo- nium salt of i7-sulpho-benzoic acid, which is soluble in water; the ^-sulphamine-benzoic acid originally contained in the saccharin remains unchanged and is insoluble in cold water.' During this operation there is frequently formed a volatile product which con- denses in white flocculent masses and is returned with condensed water-vapor, and is insoluble in cold water. After the operation has been completed the contents of the flask are allowed to cool ; the soluble acid ammonium salt of -chlor-benzoic acid could be obtained. In the methods of analysis hitherto devised for the determina- tion of the benzoic sulphinide and the impurities contained in saccharin, the presence of/»-chlor-benzoic acid as an impurity has not been regarded. According to the method of Remsen and Burton,^ which is based on the quantitative conversion of the sulphinide into the acid ammonium salt of chlor-benzoic acid off from the ^-sulphamine- benzoic acid and weighing each separately. In the method of Salkowski no acid is used ; the /-chlor-ben- zoic acid remains combined with its base as a soluble salt and is 1 Remsen and Dohme : This Journal 11, 332. 2 This Journal 11, 403. 532 Ulhnann. estimated, along with all substances which are more soluble than /•-sulphamine-benzoic acid, as benzoic sulphinide. The presence of a salt of /-chlor- benzoic acid will explain the occurrence of chlorine in the ash of saccharin pointed out by S. Hirschfeld.' The volatility with water-vapor of /-chlor-benzoic acid has not hitherto been recorded. The w-chlor-benzoic acid and the ^-chlor-benzoic acid are not volatile with water-vapor. Beilstein and Kuhlberg^ found that the di-chlor-benzoic acid in which one of the chlorine atoms is in the para position relative to the car- boxyl group, is volatile with water-vapor. As far as these obser- vations go, it would seem that the volatility with water-vapor of chlor-substitution-productsof benzoic acid is in someway depend- ent on the para position of the chlorine. The /"-chlor-benzoic acid obtained from commercial saccharin agreed in all respects with the descriptions given by Beilstein and Geitner,^ and by EmmerlingV and with ^-chlor-benzoic acid which I prepared later on from /-chlor-toluene. It is precipitated by acid^ from its soluble sodium salt in white flocculent masses, and sublimes in needles which melt at 234°. The calcium salt crystallizes in arborescent forms with three molecules of water of crystallization. Analyses of the calcium salt resulted as follows : I. 0.2584 gram salt gave 0.0S74 gram CaS04^ 0.0252 gram Ca. II. 0.2189 gram salt gave 0.0730 gram CaS04=ro.o2i5 gjam Ca. Calculated for Found. (C,H4ClCOO)2Ca4-3H20. I. II. Ca 9.88 9.92 9.81 The identity of the/)-chlor-benzoic acid obtained from commer- cial saccharin was thus fully confirmed. ' Thesis : Ein Beitrag zur Saccharinfiage. Erlangen, i Hirschfeld has again used the empirical method devised by Salkowski for the analysis of sac- charin after so exact a method as that of Remsen and Burton has been devised. I have had occa- sion to prepare pure benzoic sulphinide from commercialsaccharin, and from personal experience and the experience of many others I know that no reliance whatever is to be placed on the method of Salkowski, which depends on the suppositions that benzoic sulphinide and/-sulpliamine-ben- zoic acid can be separated by three fractional crystallizations in water, and that the resulting mother-liquors upon evaporation give approximately pure benzoic sulphinide. Hirschfeld con- firms the erroneous statement of Salkowski that benzoic sulphinide after long-continued boil- ing with dilute hydrochloric acid is left unchanged. Hirschfeld again confirms Salkowski's error in the melting-point of pure benzoic sulphinide, and gives it at 213°. It is 222°-223°. The various connections in which the name of Professor Ira Remsen is used in this thesis by Hirschfeld can be ascribed only to ignorance of the literature. "Ann. Chem. (Liebig) 152. 3 /^rf. 139, 336. < Ber. d. chem. Ges. 8, 8S0. Para-chlor-meia-sulpho-benzoic Acid. 533 Para-chlor-benzoic acidfrojn p-suJphamine-be7izoic acid. — For a supply of/-chlor-benzoic acid to be utilized in a reinvestigation of its sulphonic-acid derivative, it was found necessary to use some other method of preparation than the distillation from commercial saccharin. It was thought that by treating /!>-sulphamine-benzoic acid with an excess of phosphorus pentachloride at a high tem- perature, the sulphonamide group would be driven out and replaced by chlorine. Then upon saponifying the resulting /•-chlor-benzoyl chloride it would be transformed into/>-chlor-ben- zoic acid, and could then be distilled with steam from the accom- panying impurities. This reaction has been successful with many sulphonic acids ; a case bearing directly on the experiment in question is the conversion of di-chlor-para-sulpho-benzoic acid into tri-chlor-benzoic acid accomplished by R. Otto.' In the case of a sulphamine-benzoic acid the action of phosphorus pentachlor- ide might result in the formation of /-chlor-benzonitrile, analogous to the formation of f7-chlor-benzonitrile from benzoic sulphinide and phosphorus pentachloride.^ Both of these suppositions were realized, but the yield of /-chlor-benzoic acid was too small to recommend the use of this method as a practical one for the prep- aration of the acid in quantity. Para-chlor-benzoic acid from p-chlor-tohiene. — While the experiments described above were in progress, other experiments were being carried on with Emmerling's method for the prepara- tion of/-chlor-benzoic acid.^ By this method small quantities of /-chlor-toluene are treated with dilute solutions of potassium per- manganate containing one part of permanganate to three parts of /-chlor-toluene used. Taking into account that ordinary /-chlor- toluene contains from twenty-five to fifty per cent, of ^-chlor- toluene, and that for a complete decoloration of the dilute perman- ganate solution about nine hours are necessary, this method is not more rapid than the one above described for getting ^-chlor-ben- zoic acid out of /-sulphamine-benzoic acid. But it was found that concentrated solutions of potassium permanganate with an excess of/»-chlor-toluene yield satisfactory results. The following method is the best one available for the preparation of /-chlor-benzoic acid ; Into a four-liter globe-flask containing three liters of dis- tilled water are put one hundred and fifty grams of potassium 'Ann. Chem. (Liebig) 123, 226. ^ Remsen and Dohme: This Journal 11, 332. » Ber. d. chem. Ges. », 880. 534 Ullmann. permanganate. To this is added fifty to seventy cubic centime- ters of ^-chlor-toluene. The flask is placed in a bath of boiling water and is connected with a reflux condenser. After having been kept at the temperature of boiling water for twelve to fifteen hours the color of the permanganate disappears. The excess of ^-chlor-toluene is readily recovered by distillation with steam. While still hot the soluble salts of para- and ortho-chlor-benzoic acid are filtered from the insoluble hydroxides of manganese. The solution of potassium para- and ortho-chlor-benzoate is allowed to cool and is then treated with sulphuric acid, whereby the ^-chlor-benzoic acid is completely precipitated together with a little ^-chlor-benzoic acid. The greater part of the -chlor-benzoic acid when boiled with water. The method adopted as the most practicable is the following one : Into a receiver is placed dry powdered /!>-chlor-benzoic acid, and into it is distilled sulphuric anhydride from a retort contain- ing fuming sulphuric acid. A brownish viscous mass is formed. By occasionally heating this mass to the boiling-point of the sul- phuric anhydride, the reaction is hastened and is fully completed in the course of a day. The contents of the receiver are now transferred to a retort and the excess of sulphuric anhydride is distilled off. The ^-chlor-sulpho-benzoic acid thus formed is 536 Ullmann. poured into water and dissolves completely. The solution is treated with barium carbonate, and the filtrate from the barium sul- phate gives after concentration excellent crystals of barium /-chlor-;«-sulpho-benzoate, difficultly soluble in cold water. The final mother-liquors from the crystallizations of the barium salt of /-chlor-w-sulpho-benzoic acid contained a much more soluble barium salt. This more soluble salt is not the salt of /-chlor-(? sulpho-benzoic acid, as was found by comparing the free acid and some of its derivatives with the substances obtained by J. J. de Roode from /-chlor-toluene-<7-sulphonic acid.' It is probably the barium salt of a/-chlor-di-sulpho-benzoic acid. The acid obtained in free condition by exactly precipitating the barium in barium />-chlor-»2-sulpho-benzoate with sulphuric acid, crystallized from water in long needles. It is very soluble in water. An analysis of the acid gave the following results: I. 0.2671 gram acid heated to constant weight at 150° lost 0.0492 gram H2O. II. 0.1225 gram, acid gave by Carius method 0.1004 gram BaS04 ^0.0138 gram S, and 0.0609 gram AgCli=o.oi5i gram CI. Cl Calculated for CeHc \ SO3H + sH^O. CO^H Found. H.O 18.59 18.42 S 11.04 11.24 CI 12.21 12.29 The copper, sodium, potassium, lead and zinc salts were made and were found to agree with the descriptions given by Collenand Bottinger. Upon slow concentration of the solution of the acid lead salt, this separates in long fine needles with five molecules of water of crystallization. The copper, lead and acid lead salts show a strong tendency toward the formation of insoluble basic salts. When the solutions were concentrated on a water-bath, even at a comparatively low temperature, the basic salts separated out. This was most marked in the case of the copper salt. To obtain a good crystallization of the neutral salts it was necessary to let the water evaporate slowly in a desiccator. A determination of the water of crystallization in the acid lead salt gave the following result : 'This Journal 13, 220. Para-chlor-meta-sulpho' benzoic Acid. 537 I. 0.2582 gram acid Pb salt heated to constant weight at 160°, lost 0.0333 gram. Calculated for (CTH.ClSOeljPb+sHaO. Found. HiO 11.72 11.67 Barium SalL — The barium salt is remarkably dimorphous. It crystallizes in two forms, a compact tabular form and a yielding, spongy, arborescent form, both containing three molecules of water of crystallization, and under the.same conditions losing it equally. The appearance of the spongy form when dried between folds of filter-paper is best described by comparing it with the ordinary sea-fern. The compact tabular form is the one which was most commonly obtained. Both forms sometimes occurred together in the same solution, and by redissolving one the other could be made to crystallize out under the proper conditions. The condi- tions for the arborescent form were found to be slow crystalliza- tion from a saturated solution at a low temperature. It was usually obtained by allowing a solution saturated at about 18° to stand over night during the winter months when the temperature of the labor- atory would fall to about 5°. It was sometimes formed by expos- ing a beaker containing a saturated solution to the influence of the lower temperature out of doors, care being taken to prevent a too sudden chilling of the solution. The arborescent form dissolves readily in water, the tabular form even when finely powdered is much more slowly dissolved. Both forms contain three mole- cules of water of crystallization which are lost at iqo°. Analyses of these salts gave the following results : I. 0.4365 gram tabular Ba salt dried at 180° to constant weight lost 0.0554 gram. II. 0.1918 gram arborescent Ba salt dried at 190" to constant weight lost 0.0244 gram. III. 0.4365 gram tabular Ba salt gave 0.2385 gram BaS04. IV. 0,3128 gram tabular Ba salt gave 0.17 11 gram BaSO^, Calculated Or ^.H,-jcOO>Ba + 3 „0. Found. Arborescent H=0 12.69 12.68 12.72 Ba 32.23 32.15 32.18 It was at first supposed that although these dimorphous forms might have the same composition, the molecules of water of 538 Ullmann. crystallization would be combined in different ways, which would become manifest by the rate at which the water is given off. To test this supposition a portion of the tabular barium salt and of the arborescent barium salt were heated in an air-bath under the same conditions. Both portions had been powdered to the same degree of fineness. It was found, however, that the loss of water proceeded at the same rate in both forms. Constitution of p-chlor-m-suipho-benzoic acid. — As has been stated, Collen and Bottinger in their work on the/>-chlor-sulpho- benzoic acid in question were unable to arrive at any conclusion concerning its constitution. There are but two possibilities in the case of a mono-sulphonic- acid derivative of /-chlor-benzoic acid — the sulphonic-acid group may enter into the ortho or the meta relation, relative to the car- boxyl group. Now, it has been shown in a great number of cases that, when an acid having the carboxyl and sulphonic-acid groups in the ortho relation is treated successively with phosphorus pentachloride and ammonia, the result is always an imide and not a di-amide. In the di-carboxyl acids this tendency toward an imide formation in the ortho relation is most simply illustrated in the case of phthalic acid, which by the treatment above described is converted quantitatively into phthalimide. Cases bearing more directly on a /-chlor-sulpho-benzoic acid are the conversion of /-brom-t?-sulpho-benzoic acid into/-brom-sulphinide by Remsen and Bayley,' and the conversion of sulpho-terephthalic acid into the amide of terephthalic sulphinide by Remsen and Hall.^ Therefore, if the carboxyl group and the sulphonic-acid group are in* the ortho relation in this/>-chlor-sulphonic acid, the result of treating it with phosphorus pentachloride and ammonia should be ;:>-chlor-benzoic sulphinide, CeH=Uo>^"' ^^^"^ ^^' ^""" [ci made and studied by de Roode.' The substance which I obtained by treating this />-chlor-sulpho- benzoic acid successively with phosphorus pentachloride and ammonia differs from /-chlor- sulphinide in all of its properties and will be shown to be a di-amide. Since />-chlor-benzoic sulphinide was not formed, and since the di-amide is formed, the carboxyl and sulphonic-acid groups cannot be in the ortho relation ; they must be in the meta rela- > This Journal 8, 23r. 2/^/^.3,409. ^ I6id. \Z, 22<). Para-chlor-meta-sulpho-benzoic Acid. 539 tion, and the constitution of the acid is expressed by the formula, rCOOH(i) CeHa -< S020H(3) It is/-chlor-m-sulpho-benzoic acid. (CIG). Chlorides of p-chlor-m-sulpho-benzoic acid. — The chlorides, and the amides derived from them, were also investigated by Collen and Bottinger." The results recorded by them seemed anomalous when compared with the reactions which sulpho-ben- zoic acids in general undergo when treated with phosphorus penta- chloride, and a revision of the work was undertaken. The chlorides were prepared as follows: The neutral sodium salt, dried at 160° and finely powdered, is ground together with an amount of phosphorus pentachloride equivalent to two mole- cules of phosphorus pentachloride to one of sodium salt. The reaction takes place spontaneously, leaving a whitish oily mass. When this is treated with water the liquid chloride separates out as a light yellow oil, which on repeated shaking with portions of water, becomes perfectly white and waxen, and on being allowed to stand under water, solidifies after a short time. The solid chloride is then freed from adhering water as well as possible by pressing between folds of filter-paper. It is then dissolved in pure anhydrous ether, in which it is very easily soluble, and is freed from a small portion of insoluble inorganic salts which are some- times still present. If the ethereal solution is now placed in a large desiccator and the ether is allowed to evaporate slowly, needle-shaped crystals are deposited. They have the general appearance of a pure substance. On determining the melting- point, however, it is found that they begin to fuse at ss^.are-semi- liquid from 70°-i30°, and melt completely between i30°-i55°, the range of the fusing and melting-points varying with each preparation. Recrystallization from anhydrous ether does not lead to a substance with a more satisfactory melting-point. During the process of crystallization and recrystallization in ethereal solu- tion, fumes of hydrochloric acid were given off, which points to a decomposition of the chloride in ethereal solution. In fact, a pure chloride with a sharp melting-point, which was obtained later on by the use of another solvent, underwent change in an ethereal solu- tion. The same chloride could be kept under water for several days without undergoing any very marked change. The use of more 1 Ber. d. chem. Ges. 9, 1250. 540 Ullmann. than the equivalent of two molecules of phosphorus pentachloride, and heating to 150°, leads to a compound with a tendency to become semi-liquid at a low temperature. When phosphorus pentachloride is used in the proportion of one molecule and the operation carried on at as low a temperature as possible, a com- pound with a tendency to become semi-liquid at a higher tem- perature is formed ; but these variations in conditions and amount of phosphorus pentachloride did not yield a pure product. The crystallized compound melting from 35°-i55° is evidently a mixture of chlorides. Its properties agree with those of the mixture obtained by CoUen and Bottinger, and which they sup- posed to be a mixture of two monochlorides : r CI (CI CeHa-^ SO.Cl and CeHs-^ SO=OH. IcOOH (COCl It is in the first place highly improbable that under the same conditions the carboxyl group would be acted on in one molecule and the sulphonic-acid group in another. Furthermore, the chief product of the reaction of ammonia on this mixture is a di-amide rci CsHs^ SO2NH2, which shows the preponderance of the di-chloride (.CONH= fCl CeHs \ SO2CI, in the mixture. And it was found possible to sepa- Icoci rate from the mixture of chlorides two distinct chlorides, the one fCl being the di-chloride CcHrJ SOoCl, and the other the mono-chlo- (COCl rc! ride CeH-iX SO^OH. This was accomplished as follows: (COCl After the chlorides have been dissolved in ether to rid them of inorganic impurities, the ethereal solution is rapidly evaporated to dryness by blowing over it a stream of dry air. In this way small portions of water which were enclosed in the chlorides before dissolving them in ether are gotten rid of. The dry chlorides are then shaken with successive portions of cold anhy- drous ligroin (7o°-8o°), in which the di-chloride is rather soluble and the mono-chloride insoluble. The solution of the di-chloride is concentrated by evaporation in the water-bath, care being taken to avoid the contact of water-vapor by placing over the beaker a Para-chlor-nieia-sulpho-benzoic Acid. 541 watch-crystal which just covers the mouth. From the concen- trated solution, when cooled slowly, the di-chloride separates in flakes which are nearly pure, and by a recrystallization from ligroin are rendered perfectly pure. Di-chloride of p-chlor-msidpho-bejizoic acid. — This chloride crystallizes from ligroin in flakes melting sharply at 42°-43°. From ether it crystallized in needles. In ethereal solution it slowly undergoes change, as is shown by a lowering and indefi- niteness of its melting-point. It can be kept under cold water for a considerable length of time without decomposition. In water at a temperature above its melting point it undergoes rapid decom- position with the formation of />-chlor-»2-sulpho-benzoic acid. Treated with ammonia it is converted into the corresponding di-amide. A determination of the chlorine was made by boiling the chlor- ide with water containing a little nitric acid in a balloon-flask with upright condenser. After boiling for a short time the insol- uble chloride is completely decomposed into />-chlor-7«-sulpho- benzoic acid and hydrochloric acid. The hydrochloric acid is then precipitated as silver chloride. An analysis gave the result : I. 0.3492 gram chloride gave 0.3625 gram AgCl = 0.0897 gram CI. " (Cl Calculated for C.Ho^ CO Cl (SO2CI. Found. Cl, replaceable by OH 25.93 25.68 Mono-chloride of p-chlor-m-sulpho-benzoic acid. — This chloride is formed in small quantity together with the di-chloride when the sodium saltof/»-chlor-;/z-sulpho-benzoic acid is treated with phos- phorus pentachloride. It remains undissolved when a mixture ot the chloride is treated with cold ligroin. It crystallizes from ether in needles. It has not yet been obtained in perfectly pure condi- tion, as is evident from analyses and from the melting-point, which varies between 163° and 167°. Treated with ammonia, it yields a very little of the di-amide and a soluble salt. The presence of the di-amide is accounted for by the impurity of the mono-chlor- ide hitherto obtained, which contains a little of the di-chloride. The soluble salt formed has not yet been isolated in pure condi- tion for study. It is in all probability the corresponding amide ammonium salt. 542 Ullmann. Determinations of the replaceable chlorine in the mono-chloride were made, as in the case of the di-chloride,by boiling with water. Results were obtained as follows : I. 0.3142 gram chloride gave 0.1799 gram AgCl = 0.0445 gram CI. II. 0.1820 gram chloride gave 0.1003 gram AgCl. Calculated for CjHa I SO3OH Found. I cub. I. II. CI, replaceable by OH 13.91 14.20 14-27 A determination of the chlorine replaced by the action of ammonia was also made. For this purpose a known weight of chloride is boiled with ammonia in a balloon-flask. After acidi- fying with nitric acid the chlorine is precipitated as chloride. This determination of the replaceable chlorine resulted as follows : 0.3086 gram chloride gave 0.1792 gram AgCl =0.0443 gram CI. ici Calcu ated for CeHj.^ SOjOH ( COCl. Found. CI, replaceable by NH2 13.91 14-36 In the above analyses the amount of replaceable chlorine found is too high for the mono-chloride. It is evident that the mono- chloride melting at i63°-t67° still contains a litde of the di-chloride. Di-amide of p-chlor-m-sulpho-benzoic acid. — By treating the mixture of the chlorides, or the di-chloride of /-chlor-w-sulpho- benzoic acid with ammonium hydroxide the di-amide is formed. This is soluble in the excess of ammonia used. For the recovery of the di-amide, the solution was evaporated until most of the free ammonia had been given off, and upon acidifying with hydro- chloric acid the di-amide is precipitated. By washing with cold water the soluble salts were dissolved. The di-amide is readily soluble in hot water, and after two recrystallizations separates in excellent needle-shaped crystals which melt at 233°. This di-amide has been described by CoUen and Bottinger as an amide- ammonium salt derived from what they supposed to be a mixture of mono-chlorides. It has been shown that this mixture is one consisting mainly of the di-chloride which, as would be expected, yields the di-amide. The di-amide resulting from the mixture of chlorides is far more insoluble than would be expected for an Para-chlor-meia-sulpho-benzoic Acid. 543 ammonium salt, and does not give off ammonia by treatment with cold caustic soda. The analyses point most clearly to the di-amide. I. 0.2791 gram di-amide by absolute method gave 28.0 cc. N at 18° =10.03294 gram N. II. 0.3126 gram di-amide by Liebig's method gave 0.3094 gram BaSO^. III. 0.1923 gram di-amide by Liebig's method gave 0.1914 gram BaSO*. Cl Calculated for CeHa SO^NHo CONH2. Found. N 11.97 II.9I I. II. S 13.66 13.60 13.68 Amide-ammonium salt and di-anilide. — By treating the mono- chloride which melts at i63°-i67° with ammonia, an amide- ammonium salt is formed which probably has the constitution ( CONHs CsHa^ SO2ONH* (ci. It is much more soluble than the di-amide. It has not yet been obtained in pure condition. The di-anilide formed by treating the di-chloride with aniline was obtained in pure needle-shaped crystals which melt at 219°- 220°. It is the intention of the author to pursue this investigation farther, and to reinvestigate the work done on />-brom-sulpho- benzoic acid, in regard to which statements analogous to those on /-chlor-sulpho-benzoic acid have been recorded. This investigation was undertaken at the suggestion of Profes- sor Remsen and carried out under his supervision, and I desire here to acknowledge my indebtedness to him. Vol. XVI.— 40. 544 Mabery. Contributions from the Chemical Laboratory of the Case School of Applied Science.' IV.— ON THE DETERMINATION OF SULPHUR IN VOLATILE ORGANIC COMPOUNDS.' By Charles F. Mabery. The great quantity of products introduced into the petroleum industry froin the fields in Ohio and Canada yielding the sulphur oils has involved many sulphur determinations, and the necessity of a rapid methodcapableof affording results of extreme accuracy, especially in oils containing a small fraction of one per cent, of sulphur. Several of the older methods leave nothing to be desired in point of accuracy, but they are not sufficiently expedi- tious for service in manufacturing operations, or in investigations which depend upon immediate information concerning the per- centage of sulphur. The first attempt to determine sulphur in organic compounds by combustion in oxygen was made by C. M. Warren,^ the sulphuric acid formed being absorbed within the combustion-tube in plumbic peroxide. The oxides and acids formed by combustion were first distilled and collected in bromine-water as an oxidizing agent, by Sauer,^ and this method was still further improved by Mixter,"" who avoided the use of a rubber cork in the forward end of the combustion-tube, carried forward the volatilized substance by a current of carbonic dioxide, and suggested more efficient means for oxidation by bromine and absorption. All these methods depend upon the formation of sulphuric acid and pre- cipitation as baric sulphate, which involves considerable labor when a large number of determinations are necessary in a limited time. To overcome this difficulty Burton' suggested a modification of the method of Sauer, which consists in absorbing the oxidized sulphur in a standard solution of potassic hydrate and titrating the excess of alkali with standard sulphuric acid. Besides these methods the only other suitable means for the determination of sulphur in oils with large percentages of sulphur ' Proceedings of the American Academy of Arts and Sciences. Aid in the work described in this paper was given by the Academy from the C. M. Warren Fund for Chemical Research. * This paper is one of tlie series on the composition of the sulphur-petroleums. ' Proc. Am. Acad. 6, 472. * Ztsch. anal. Chem. 13, 32. s This Journal 2, 396. * Ibid. 11, 73. Determination of Sulphur in Volatile Organic Compounds. 545 is the well known method of Carius, in which the substance is oxidized in a closed tube by means of fuming nitric acid. In its applicability to all classes of compounds, and in the accuracy of results of which it is capable, this method leaves little to be desired except perhaps in the analysis of oils containing less than o.oi per cent, of sulphur. On account of the limited weight of substance that can be oxidized in a Carius tube another method must be selected for substances containing less sulphur. Our experience has shown that the Carius method may be relied upon in sulphur determinations to yield concordant results wiihin a few hundredths of one per cent. Oxidation of the less volatile oils containing a small percentage of sulphur, without doubt, may be accurately accomplished in an open vessel, but with larger amounts of sulphur the action of nitric acid is so violent that it must entail loss by volatilization, unless indeed the sulphur oil is considerably diluted by a sulphur-free oil, in which case the solvent must be completely oxidized. The great number of sulphur determinations in crude oils and products obtained from them, connected with the extended exam- inations which have occupied my attention during several years past, has demanded a careful comparison of the various methods as to their efficiency and economy of time. Particular attention has been given to details of the Carius method, with the precau- tions necessary in its successful application to the analysis of sulphur oils. The first requisite is a furnace of suitable construc- tion to maintain an equal temperature, easily controlled in all the tubes within the furnace, without a great loss of heat by radiation. For this purpose and for Carius analyses in general I have recently had a furnace constructed which differs in certain features from any other I have seen, and it shows such a high degree of efficiency that a brief description may not be entirely devoid of interest. The body is of the ordinary cylindrical form, 75 cm. long and 25 cm. in diameter, of heavy sheet iron, and it is sur- rounded by two outer jackets of sheet iron each enclosing a half- inch space, and extending beneath on either side to within 6 cm. of the heating tube ; it is supported upon legs of strap-iron ^-inch thick and 2 inches wide, each entirely encircling the body at either end, for rigidity. These two air-spaces retain the heat so effectually that the hand may be borne on the outside of the furnace when the thermometer within indicates a temperature of 546 Mabery. 200°. The iron tubes are as usual of gas-pipe, with threads at either end, with caps easily movable by the fingers. With a small hole in each cap for the escape of gas, these tubes retain all glass in the most violent explosions. When several tubes are in the furnace at the same time a record of them may conveniently be kept by suspending metal tags numbered consecutively, from the holes in the caps by means of bent wire. Fig. I shows the arrangement of the outer air-spaces with the position of the heating-tube. The furnace is heated by means of a gas-stove heater 45 cm. in length, with thirty- two gas-jets that will burn continuously with a flame 2 mm. high, giving a temperature within the furnace of less than 60°; by inter- posing an asbestos or an iron plate a consider- ably lower temperature may be maintained. The heating-tube is supported on two iron straps bolted to the legs, one at either end of the furnace ; by means of it the heat is very equally distributed with little waste, and the Fig. i. glass tubes being thus evenly heated there is less danger of loss by explosion. A temperature of 200° may be obtained within twenty-five minutes after lighting the jets, and it may be maintained with jets fifteen millimeters in height, requiring a small consumption of gas; the hand maybe held without dis- comfort for some time directly beneath the heater. The variation in temperature at different heights within the furnace is small ; with the thermometer at 275" at the level of the upper tubes, the temperature at the level of the lower tubes is about 9° higher. For temperatures higher than 275° a second heating-tube is necessary. It is frequently convenient to be able to regulate within close limits the flow of gas for the required temperature without further attention after lighting the jets. The device shown in Fig. 2, which suggested itself for this purpose, consists in attaching to the end of the gas-valve by means of a screw-thread a brass cap with an index of stout copper wire moving in front of a graduated circle with a radius of about six inches.' 1 Evidently this attachment has no reference to variations in temperature caused by changes in pressure on the gas within the mains. For regulating the temperature within closer limits, a gas-regulator can be inserted. Deierminaiion of Sulphur in Volatile Organic Compounds. 547 With glass tubes of large size — those we use are 15 mm. inside diameter — well sealed, and with strict adherence to certain conditions which have elsewhere been described by A. W. Smith and me,' there is little danger of an explosion. The quantity of nitric acid should not be in excess of twenty times the weight of the substance taken, and after heating to 175° for fifteen hours the tubes are opened — best without remov- FiG. 2. ing from the furnace — resealed and heated again to 250° during five to ten hours. The serious objection to the Carius method for sulphur is the slow process of oxidation, and it seems hardly possible to hasten the operation by raising the temperature, since glass tubes will not stand the great pressure. In studying various methods depending upon the oxidation of sulphur by combustion I have found that nothing less than com- plete oxidation gives reliable results. Many experiments on fractional combustion have shown clearly that compounds with high percentages of sulphur do not yield concordant results, even when the sulphur compound is diluted with a sulphur-free oil. I have found Burton's adaptation of the Sauer method reliable and expeditious, and with certain modifications presently to be described it is perfectly satisfactory for the analysis of oils of high as well as low percentages of sulphur. In Fig. 3 the inlet tube for oxygen or air is shown as entering through the rear stopper, as proposed by Mixter, and extending just to the centre of the constriction. In the combustion of some of the oils which we have under examination, the temperature must be maintained as high as the most infusible Bohemian glass will stand, and at such tem- peratures the smaller tube within is distorted if it is placed in the forward portion of the combustion-tube in the zone of greatest heat ; if it terminates at the narrowest point of the constriction, continuous combustion is insured by thorough admixture of the volatilized substance with oxygen. Complete oxidation is still more certain in rapid combustion if that portion of the tube in front of the narrower part is left somewhat longer than is preferred by Sauer, Mixter, or Burton. The tubing we have in use is some- 1 This Journal 16, 83. 548 Mabery. what thicker in the wall than that in ordinary use, and larp^er, with an inside diameter of i8 mm. It is important that the oxida- tion proceed as rapidly as is consistent with complete absorption, and we find that this is best accomplished in a large U-tube partly filled with broken glass. Our U-tube is 34 cm. in height, 25 mm. inside diameter, and with 50 cc. of the absorbent solution a rapid gaseous stream may be passed through without danger of loss. For low sulphur oils we use a solution of sodic hydrate of such a strength that i cc. equals 0.00 10 gram of sulphur, and for Titf Fig. 3. higher percentages a solution in which i cc. equals 0.0050 gram Methyl-orange has been used as an indicator in all our determina- tions ; the change in color in titrating an alkaline solution with this indicator is well defined and exceedingly delicate. The titra- tions may be made in the U-tube without transferring the solution after washing in the acid from the combustion-tube. To carry forward the volatilized substance it is advantageous to introduce a slow current of carbonic dioxide, as proposed by Mixter, and we have sometimes used a combustion-tube closed with a rubber cork in front and sometimes a bent tube. With substances con- taining a high percentage of sulphur it is doubtless safer, as Mixter suggests, to avoid the use of a cork in front. On account of the large consumption of oxygen in burning rap- idly a considerable weight of oil — at least three times the quantity theoretically required for oxidation — and finding that the com- bustion proceeds with equal facility in air, nearly all our determi- nations have been made in a current of air supplied under pressure, with the same means for exhaustion that Burton found advanta- geous. The operation requires close attention and 0.5 to i gram Determination of Sulphur in Volatile Organic Compounds. 549 of oil may easily be burned in forty-five minutes to one hour, depending upon the nature of the substance, the heavier oils especially if containing much sulphur being the most difficult to burn. The higher sulphides will not support a continuous flame, and dependence must be placed upon a very hot tube ; with the more volatile oils it is sometimes difficult to maintain a continuous flame even in oxygen, the combustion proceeding in long inter- mittent non-luminous flashes. If the flame becomes luminous the rapidity of volatilization must be instantly checked, and the flow of air increased. The appearance of white fumes in the forward part of the combustion-tube or the absorption-tube, indicating improper adjustment as to the temperature, flow of gas, or rate of volatilization, is invariably attended with low results. The com- pleteness of the absorption in the U-tube was tested by placing a second tube beyond it containing a similar solution, but no trace of acid was found in the second tube when the excess of alkaline hydrate in the first at the end of the analysis was not less than 15-20 cc. With a smaller excess in rapid combustions there is danger of loss. The oil for analysis is weighed in a bulb or tube of hard glass, and it is sometimes convenient to transfer most of it to a platinum or a porcelain boat, which may easily be accom- plished without loss within the combustion-tube provided there .is a gentle current of air inward and the combustion-tube in front has previously been heated to the required temperature. In the examination of Ohio and Canadian sulphur petroleums for identification of the paraffin, aromatic, and unsaturated hydro- carbons, sulphur compounds, and other constituents, with which I am at present engaged, numerous determinations of sulphur have been necessary, and the extreme convenience of combustion in air has greatly facilitated the separation of the various products. As an evidence of the reliability of this method, the following results are selected, with parallel determinations by the Carius method : Distillate from crude Canada oil collected at Percentage of Sulphur. 89°-9i° after one distillation under 50 mm. Combustion in Air. Carius. and seven under atmospheric pressure. .. . 0.044 0-043 Distillate from crude Ohio oil collected at I27''-I29° afterone distillation under 50mm. and seven under atmospheric pressure .... 0.0343 0.036 Distillate from crude Canada oil collected at 1 1 5°-i 17° after one distillation under 50 mm. and seven under atmospheric pressure ... . 0.173 0.0108 550 Mabery. Distillate from crude Canada oil collected at I20°-I30° after one distillation under 50 mm. and five under atmospheric pres- sure o. 505 The same after shaking five times with alco- holic mercuric chloride 0.07 The same after shaking once with alcoholic mercuric chloride with the addition of solid mercuric chloride 0.086 I. II. Sulphur oil from Canada sludge acid 6.3 6.3 6.47 Sulphur oil from Canada sludge acid 17-36 I7'3i I. II. Sulphur oil from Canada sludge acid 16.67 16.76 Canada sulphur oil 6.15 6.01 Canada sulphur oil 13-67 13-70 Crude sulphide separated by mercuric chlo- ride from fraction iio°-ii5° of sulphur oil after the fifth distillation under 50 mm 18.85 A fraction of the same corresponding to pen- i. n. m. tyl sulphide, percentage of sulphur, 18.39. 1^.53 18.55 18.67 These results were obtained by six persons working independ- ently of one another. The oxidation of nitrogen to any considerable extent by the use of air in the combustion of sulphur compounds is evidently excluded by the close agreement of the results it yields with corresponding determinations by the Carius method. In accord- ance with the suggestion of a friend, from the fact that nitrous and nitric acids are formed to a greater or less extent depending upon conditions in the ordinary forms of combustion, it seemed of interest to ascertain whether these acids were present at all in the alkaline absorbent. In testing for the formation of nitrous acid, the exceedingly delicate color reaction was applied which is produced in an acid solution of a nitrite by the addition of sul- phanilic acid and naphthylamine chloride. An examination of our reagents showed that the purest commercial sodic or potassic hydrate gives an intense color, and even hydrates prepared from the metals are not free from color. Pure sulphuric acid gave no reaction, and pure sodic carbonate only a faint color. We finally obtained a solution that gave not a trace of color by dissolving metallic sodium and boiling the solution for some time with metallic aluminum. With this solution as the absorbent in a Reviews and Reports. 551 combustion of a sulphur oil, after the analysis the solution was as free from color as before when it had stood half an hour after the addition of the reagents. Since a pink color is distinctly visible in this reaction with one part of nitrogen in the form of nitrous acid in one thousand million parts of solution, it is safe to conclude that nitrous acid is not one of the products in this form of com- bustion. To determine whether nitric acid is formed, after the combus- tion a portion of the sodic hydrate solution was neutralized, mixed with ferrous sulphate, and concentrated sulphuric acid poured beneath the solution. No color was visible at the junction of the two liquids. In a second test another portion of the alka- line solution was nearly neutralized with sulphuric acid, evaporated to dryness, and a few drops of phenolsulphuric acid added. Upon diluting to a definite volume, no difference could be perceived between the color of this solution and that given by phenol- sulphuric acid alone in a blank experiment. In the combustion of sulphur oils in air, therefore, the atmospheric nitrogen is not affected. For efficient aid in studying the details of these methods of analysis, I should acknowledge my obligations to Mr. W. O. Quayle, and to my assistants, Messrs. D. B. Cleveland and G. M. Little. REVIEWS AND REPORTS. Recent Progress in Physical Chemistry. Die Geschwi7idigkeit des Ueberganges von Aldoximen in S'dure- niirile. A. Hantsch: Ztschr. phys. Chem. 13, 509 (1894). — The R— C— H antialdoximes || differ from the stereoisomeric synal- HO— N R— C— H doximes || in that only the latter, in the form of their N— OH acetates, are decomposed by the alkaline carbonates into nitriles 552 Reviews and Reports. and acetic acid. The readiness of this nitrile formation depends to a high degree upon the nature of the substituent R, and in the effort to obtain a numerical expression for this dependence the characteristic constant of the reaction-velocity was determined for a dozen different acetates. When the process is considered to be RH:C:N.OCOCH3 = R.C; N + CH3COOH it is mononiolecular and its velocity must remain proportional to the concentration of the undecomposed acetate, i. e., -^=C.iA-x)ov^\og [A/{A-x)-]-C, when X parts of the initial quantity A are decomposed at the time /. The velocity constant C furnishes a sensitive test for the purity of the aldoximes used. The compounds investigated fall into the following order with respect to readiness of decomposition : 1. Metanitro-derivative C=: 0.000128 2. Parachlor " 371 3. Thiophene " 408 4. Anis " 410 5. Piperonal " 474 6. ^-Methyl " 475 7. Benzaldehyde " 552 8. /)-Ethoxyl " 564 9. /)-Brom " 619 10. p-lodo " 696 II. /-Cyan " 8(?) 12. />-Nitro " 8 These results seem to indicate no simple relation between con- stitution and stability. Ueber die Besiivtmung kleiner Dissociaiionsspannungen, spe- ziell krystalhvasserhaJtiger Salze. C. E. Linebarger : Ztschr. phys. Chem. 13,500(1894). — When two phases of a heterogeneous system are in equilibrium with a third they are also in equilibrium with each other. From this principle the conclusion has been drawn by Nernst that the dissociation-pressures of a substance whose volatile dissociation-products are soluble in some liquid are identical with the partial pressures of these products above the saturated solution of the substance (itself insoluble), and can hence be found from the solubilities of these products and their solubility-coefficients. This is tested in the present paper for various hydrated salts in equilibrium with ether at the boiling- point, the solubility of the dissociated water being calculated in each case from the boiling-point of the solution. For crystallized salts of copper, strontium, magnesium and zinc values were obtained in close agreement with the direct determinations of Frowein. It was shown also that such determinations can be Reviews and Reports. 553 consecutively made in a given mass of ether if the salts employed be added in the order of their dissociation-pressures. The paper is an interesting contribution to the literature of Chemical Statics. Beitrd^e zur Molekulargewichtsbestimmung aji ^'festen L'osun- gen." F. W. Kiisier : Ztschr. phys. Chevi. 13, 445 (1894). — The chemical equilibrium established between caoutchouc, ether and water is an especially interesting- instance of an equilibrium in heterogeneous systems involving a solid solution, for pure caout- chouc dissolves ether readily, and in the system mentioned the ether is subjected to a. distribution between the solid and liquid phases. To given amounts of water and ether varying quantities of caoutchouc were added and the concentrations of the aqueous solution determined from its freezing-point. The concentration C in the caoutchouc is strictly proportional to the square of that c in the water ; therefore, according to the law of the reaction-isotherm, the ether is bimolecular in the caoutchouc and the distribution between the two phases (at 0° C.) takes place according to the ^''^^"'^ (C4H,oO> % 2C.H10O in caoutchouc in water and the equilibrium constant is given by the equation K. Cz=zc\ The equilibrium at higher temperatures, near 17° R., is of a yet more interesting character, for here the ratio of distribution K, although at first constant, increases with falling concentration. This indicates a progressive dissociation of the bimolecular ether in the caoutchouc. The ratio of the total concentrations approaches with falling concentration the limit 2.8, which is therefore the ratio of distribution of the monomolecular substance. The product of this ratio into the concentration of the (always monomolecular) ether in water yields then the concentration c' of the monomole- cular ether in the caoutchouc for each case, and this taken from the total concentration in the solid phase leaves that C of the undi.ssociated bimolecular ether there. The dissociation-isotherm requires now the relation K.C'-=.c''^ or — j=^-=z const. This is the first instance of a quantitative application of the theory of dissociation to a substance dissolved in solid solution, and it is at the same time a beautiful application of Nernst's Principle of Distribution. A continuation of the work at 0° C, whereby the concentra- tions were varied within very wide limits, yielded a quantitative result of the same character as that just cited; and it was found that the dissociation of the bimolecular ether in the solid solution is, at ordinary temperatures, for a given total concentration, about three times as great as at zero. 554 Revie7vs and Reports. Ueber die Dissociatioyi vo7i Kalhimtrijodid in wdssri^er Lbsung: A. A. Jakovkin : Zischr. phys. Chevi. 13, 539 (1894). — When iodine is dissolved in aqueous potassium iodide it is supposed that a compound of the two substances is formed in the solution. If this be true, one must expect an equilibrium of the form to be established between this compound and its dissociation- products. In the experimental work cited, aqueous solutions containing the two substances in varying relative concentrations were shaken up with carbon disulphide at 20°, whereby any free iodine present must be distributed between the liquids in the ratio I : 410 (Berthelot and Jungfleisch). The degree of dissoci- ation is then given by the ratio of the iodine-concentration in the disulphide to 410 times that of the total iodine in the iodide solution. When a g-mols of potassium iodide are present per one of iodine, x of which are free and i — x combined, in the volume v of this solution, the dissociation isotherm becomes T^ I — X a — (i — .y) / X V \ V (a — J -{-x') X K, (I — x)v on the assumption that m-=.i, i. e. that the compound KI3 is formed. A' was found in three series of experiments to remain very constant, the tri-iodide is therefore formed and the dissocia- tion equilibrium in the solution is Kh ^ KI -+- \i. Dielektrizit'dtskonstante tmd chemisches Gleich^ewicht. Ztschr, phys. Chem. 13, 531 (1894). Methode zur Bestimmung von Dielektrizitdtskonstanten. W. Nernst: Gottinger Nachrichten 1893. 762. — The electrical energy of a system of charged con- ductors, \-e V, in which e and v represent quantity and potential respectively, falls to— ^-^P^ when the system is immersed in a medium whose specific inductive capacity is D. The conductors are therefore more readily separable in a medium with a high than in one with a low specific inductive capacity, and transfer from one medium to another is accompanied by a force tending to draw the system into the one with the greater D. [This is perhaps more readily seen when we say that two charged and separated conductors exert upon each other an attraction which is dependent upon their charges and their distance apart, the energy of attraction being proportional directly to the two electrical quantities £1 and So and inversely to the distance r (Coulomb). Writing/ for the attractive force, Reviews and Reports, 555 the negative sign indicating positive attraction when the electrical quantities are of opposite sign. When the charged bodies are separated in air the proportionality-factor /?= i, in other media (dielectrics) the attractive energy falls to -j- of its value in air, D thus defined being the Dielectric Constant or Specific Inductive Capacity. The attractive energy between the bodies is then inversely proportional to the dielectric constant of the surrounding medium.] Now, smce the charged ions of an electrolyte must be regarded as existing independently, z. e., in a certain sense separated, they must accordingly be expected to dissociate (electrolytically) more readily in media with high dielectric constants than in such with low D. And this is exactly what is observed, for, e. g., the elec- trical conductivity (which, being proportional to the ion-concen- tration, roughly measures the dissociation) of -i-normal hydro- chloric acid in various solvents varies as follows : Medium. Approximate Conductivity. D. Gases o.ooooooooo 1. 00 Benzene 0.0000005 2.26 Xylene O.OOOOCO8 2.34 Ether 0.000004 4.06 Isoamyl alcohol 1.6 16.7 Isobutyl alcohol 3-1 18.7 Ethyl alcohol 16.0 25-7 Methyl alcohol 94.0 35-3 Water 310.0 79.3 This accounts for the depression of the conductivity of an aqueous solution on the addition of alcohol, etc. The discovery is a remarkable one, and must have an important bearing upon the further study of chemical equilibrium. Many bimolecular substances, e. g., fatty acids, dissociate pri- marily into monomolecular substance, and then into ions. That these two processes are related is to be anticipated, and this appears from Nernst's discovery that acetic acid is subjected to a considerably greater dissociation, (CH3COOH)222CH3COOH, in its solution in benzene (/?= 2.3) than when in vapor (z. e. in vacuum, D= i.o) of the same concentration and temperature. [This is even better illustrated by data taken by the writer from Beckmann's work on the boiling-points of the solutions of benzoic acid. This substance is entirely bimolecular in carbon disulphide (D=Z2.6), and its increasing dissociation, (C6H5COOH).^ 2C6H5COOH, with increasing dielectric constant of the solvent is strikingly exhibited by the rapid decrease of its "mean molecular weight," as tabulated : e^6 Reviews and Reports. M=122, (^>=244. Medium. Me,m M. D. Carbon disulphi Ide 244 2.6 Chloroform 214 4.8 Ethyl acetate 133 6.7 Acetic acid 123 97 Ethyl alcohol 120 257 Water <;i20 79-3 In water, and apparently in alcohol, the electrolytic dissociation comes appreciably into play. This effect should find its most rational expression in the increase of the dissociation-constant with increasing D, but for this the numerical material and the mathe- matical theory are as yet both lackino^.] That the dielectric constant is characteristic for the chemical nature of distinctive classes of compounds is evident from the following table of approximate values: Gases Z>:= i.o Esters Z>r= 6.0-9.0 Hydrocarbons 2.0-2.5 Acetic acid 9.7 Carbon disulphide 2.6 Alcohol 26.0 Ether 4.0 Water 79.3 The presumed identity of the dielectric constant with the square of the index of refraction for indefinitely long waves of radiant energv promises to be of great interest in this connection. In the last paper cited, Nernst describes a convenient method for the experimental determination of dielectric constants. Ueber gesattigte Losungeii von Magnesiumchlorid und Kaliumsidfat oder von Magnesiumsulfai und Kaliumchlorid. R. Loewenherz : Ztschr. phys. Chem. 13, 459 (1894). — As an extension of the recent work of van der Heide,' the author inves- tigates the solubility-relations of all realizable combinations of potissium chloride and magnesium sulphate in water. In addi- tion to these two salts the system is capable of producing the double compounds schoenite and carnallite, the hexa- and hepta- hydrates of magnesium sulphate, as well as potassium sulphate and magnesium chloride, under the proper conditions. The paper is prefaced by a general discussion of the different systems con- cerned, in the light of the Phase Rule of Willard Gibbs, according to which n — i of the salts must be present in solid form in order to ensure a complete saturation in solutions made up of n inde- pendent components. The various solubility-curves which were experimentally determined are successfully represented as surfaces in coordin ite-systems of four axes, whereby a general view of the possibilities offered by the different systems is obtained. It becomes possible, for example, to foresee from the diagrams what iZcx:. cit. 13, 4:6. Revieivs and Reports. 557 salts will be crystallized out upon the isothermal evaporation of any given unsaturated solution, and in what order these salts will appear. This possibility is tested in a very considerable number of cases. The paper is concluded by a tensimeter experiment employed to confirm the simple theoretical conclusion that the vapor-pressure curves of two (complex) systems must intersect at the temperature at which a transformation of either into the other may occur. Ueber die dispersio7isfreie Molekularrefraktion einiger organ- ischer Verbindungen. Hans Jahn und Guido Moller : Ztschr. phys. Chem. 13, 385 (1894). — This work is essentially a continua- tion of that of Landolt and Jahn' upon the molecular refraction of hydrocarbons and alcohols, and it consists in the determination of the dielectric constants and molecular refraction of various fatty acids, substituted hydrocarbons and of a few mixtures. The molecular refraction of the mixtures was shown not to be strictly additive ; the change in the dispersion-free refraction accompany- ing such strictly comparable chemical processes as the replace- ment of a hydrogen in a benzene ring by chlorine was found to remain approximately constant ; and the authors emphasize the fact that in general the refractive power is a pronounced " con- stitutive property." Th^oremes ghiiraux sur Vetat des corps en dissohdion. P. Duhem': Jour, de Phys. (3) 3, 49 (1894). — Any purely thermo- dynamical principle relating to a solution or other mixture is independent of all hypotheses concerning the chemical state of the components of the mixture; the thermodynamical laws bear in this respect only upon the gross composition of the system as this is determined by analysis. One cannot expect thermo- dynamics therefore, when taken entirely alone, to determine what the chemical state of any given dissolved substance must be. But the thermodynamical method, when employed in connection with certain assumptions drawn from experience, makes possible a determination of this chemical state, and the results reached will possess the same degree of reliability which attaches to the expe- rience from which the deduction is drawn. This is in reality the manner in which the theory of dissociation and the theory of solu- tions are treated by thermodynamical investigators at the present time. Duhem presents a rigorous proof for the above statements. Beiirdge zur Kenntniss der photochemischen Wirkung in L'dsungen. M. Roloff : Ztschr. phys. Chem. 13, 327 (1894). — A systematic study of two photochemical reactions, that of ammo- nium oxalate upon mercuric chloride, and the decomposition of oxalic acid by bromine. Both processes are found to be very 1 Loc. cii. 1 0, 289. 558 Reviews and Reports. complicated, yet apparently to consist essentially in a transfer of electrical charo;es from one set of ions to another under the influ- ence of light. An effort is made to bring this result into connec- tion with the electromagnetic theory of light, according to which light is a manifestation of electrical wave-motion. Ueber die Hydrolyse von Salzen schwacher S'duren und schwacher Basen. S. Arrhenius : Ztschr. phys. Cheni. 13, 407 (1894). — By means of a spectrophotometric method Lellmannand Schliemann' have recently studied the hydrolytic action of water upon the helianthine salts of several fatty acids, whereby these salts are partially decomposed into helianthine and free acid. In the present paper Arrhenius calculates their results, employing the Guldberg-Waage equation of equilibrium, Z>(water) X Z>(salt) = Z>(acid) X Z>(helianthine), the symbol D indicating the electrolytically dissociated portions of the several electrolytes. This equation is shown to require an amount of decomposition agreeing very fairly with the observed results. Lellmann and Schliemann were accordingly not justified by these experiments in attacking the hypothesis of electrolytic dissociation. Loslichkeit des sauren Kaliumtartrats bei Gegenwart anderer Salze. A. A. Noyes und A. A. Clement : Ztschr. phys. Chem. 13, 412 (1894). — Experiments to determine the solubility of the acid potassium tartrate at 25'', both alone and in the presence of varying amounts of other potassium salts. The solubility is depressed equally by the chloride, bromide and iodide, and this depression becomes successively less when the nitrate, chlorate and sulphate are employed. The acetate increases the solubility because of the formation of much undissociated acetic acid ; hydro- chloric acid and various sodium salts effect an increase likewise, and for a similar reason. The results are in satisfactory accord with the ion theory. Die Wasserstoffabspaltung bei dem sauren Kaliumtartrat. A. A. Noyes : Ztschr. phys. Chem. 13, 417 (1894). — A measurement of the velocity of the inversion of cane sugar brought about by the acid potassium tartrate at 100°, using the method employed by Trevor, shows this salt to be more strongly acid than any other which has been yet investigated, with the single exception of the acid sodium oxalate. A calculation of the ion-equilibrium in the solution indicates the presence, at this temperature, of some eighteen per cent, free undissociated tartaric acid and fifty-three per cent, of the univalent anion produced by the primary dissocia- tion. 1 Lieb. Ann. 374, i6o. Reviezvs and Reports. 559 Ueber die W'drmeausdehnung einiger Losungen in Alkohol, Aether, Benzol tend Schzvefelkohle^istoff. G. Tammann und W. Hirschberg : Ztschr. phys. Chem. 13, 543 (1894). — From the theorem maintained by Tammann that solutions behave in many respects as though consisting of pure solvent under high pressures, and the fact that water differs from all other liquids in that its coefficient of expansion -Ty, is greater under high pressures than under low ones, it is concluded that aqueous solutions must expand with rising temperature more rapidly than pure water does, the rate increasing with rising concentration, and that solu- tions in other solvents must exhibit the opposite behavior. The latter part of this proposition is tested in the present paper. Ex- periments with solutions of various substances in alcohol and ether seem to show in fact that the curves of the specific volumes as temperature-functions under constant pressure do tend to coincide with the curves for the solvents if the pressure in each case be considered to be increased a certain amount. Ueber die Affinit'dtsgrossen einiger schwefelhaliiger Substiiu- tionsderivate von der Essigsdure und der Propionsdure. y. M. Lovhi: Ztschr. phys. Chem. 13,550 (1894). — The dissociation- constants of a number of organic acids containing sulphur were determined from the electrical conductivities of their aqueous solu- tions. The substitution of SO2 was observed in every case to pro- duce a stronger acid than does the introduction of sulphur alone. Ueber die Verseifungsgeschwindigkeit einiger Ester. A. de Hemptinne : Ztschr. pliys. Chem. 13, 561 (1894). — In order to supplement the work of Reicher upon the relative velocities with which various esters are saponified by bases, the author has undertaken a determination of the corresponding velocities when the action is brought about by acids. The reactions are mono- molecular ; it was found that the velocities for the methyl and ethyl esters or the ethyl and propyl esters of given acids are always in a constant ratio, and that the same is true for the acetates and propionates or the propionates and butyrates. That is to say, from the velocities of saponification of a series of acetates the velocities for the corresponding esters of any other acid are calculable when that for one of them is known. This holds also for the saponification by bases. The nature of the component alcohol is found to have a greater influence in deter- mining the velocity than has that of the acid. The catalytic effect of the hydroxyl ion in effecting saponifica- tion greatly exceeds that of the hydrogen ion, yet there is no proportionality between the two. Two gaseous systems were investigated : hydrochloric and hydriodic acid gases were observed to exert a very considerable accelerating influence upon the saponi- fication of the vapor of ethyl acetate. Vol. XVI. -41. c6o Reviezvs and Reports Ueber den Einfluss der chemischen KonsiUution organischer Stoffe auf ihre F'dhigkeit fesie Ldsunge7i zu bilden. G. Ciami- cian : Ztschr. phys. Chem. 13, i (1894). — The tendency of pairs of organic compounds to form mixed crystals (solid solutions) may be investigated by observing whether either one when in dilute solution in the other causes abnormally small freezing-point depressions. The study of a considerable number of pairs of aromatic compounds tends to show that this tendency depends in some definite manner upon the "constitution" of the substances and also upon their melting-points and their crystal form. Note on the Law of Transformation of Energy and its Appli- cations. W. Peddie : Proc. Roy. Soc. Edinb. 19, 253 (1893). — The investigation of the relations subsisting among various phys- ical quantities, and which are exemphfied in the different trans- formations of energy, are usually treated by mathematical methods of such nature as to be unavailable to any who are unacquainted with the higher mathematics. Yet these relations may be devel- oped in an elementary manner. When a, b, c, . . . represent the quantities which fix the condition of a system, the change of energy in a process in the system is dE=z^da + ^db + ^-dc-\-... d^ d^ oc or dE — Ada + Bdb + G/^ + . . . where A, B, C, . . . may be termed the forces producing the changes da, db, dc, . . . They are the quantities upon whose magnitude the tendency to change depends. If the system return to its original state, Ada-\-Bdb-{-Cdc-{-. . . = 0, an equation expressing the law of conservation. When only two forms of energy come into question we have, reckoning energy added as positive, d.Ada — d. Bdb z=. o. This relation is employed to derive the four " thermodynamical relations," the " vaporization-formula " and the familiar equations representing the equilibrium -relations between Heat and Electrical Energy, Electrical Energy and Work, Electrification and Vapor- Pressure, Surface-Tension and Vapor- Pressure, Induced Electromotive Forces. This paper was read before the Royal Society of Edinburgh within four weeks of the date of the presentation of Ostwald's entirely similar paper to the Royal Saxon Academy at Leipzig.' 'See This Journal 15,671. Reviews and Reports. 561 This remarkable coincidence emphasizes the undoubted fact that the time is ripe for the systematic development of the Energy Theory. Elektromotorische Krdfte zwischeji verschiedeii gekriimmten Quecksilberelektroden in einer Que cksilbersalzlo sung. T. Des Co2idres : Wied. A^m. 46, 292 (1892). — Two mercury surfaces at the same level in a solution of mercurous nitrate were maintained with different surface-tensions by pressing one against parchment paper by means of a mercury column of the height h ; the other surface was left free. The electrical energy producible by the system must be equivalent to the work to be gained in a transfer of the mercury from the height h, that is to say, for an equivalent weight of 200 grams, to 200 X 980 . h in absolute units. The corresponding electrical quantity is the familiar 96540 coulombs, and the desired difference of potential between the surfaces is V\ the electrical energy is therefore 96540 F, and we have 200 X 980./^ = 96540 V F^gg^ Xg^^X^ ^Xio-^volts, 96540 since one volt-coulomb is 10' ergs. The agreement of the obser- vations with the requirements of the theory is excellent, as may be seen from the following table : h. f^ theory. Z^ observed. 36 cm. 7.2 X IO-' volts 7.4X10"' 113 23. 21. Unpolarisierbare elekirolytische Zellen unter dem Einflusse der Centrifugalkraft. T. Des Cotidres : Wied. Ann. ^g, 2S4. (i8g;^). — When a glass tube containing an electrolytic solution is sup- ported horizontally upon a disk which is rapidly rotated in a horizontal plane, the centrifugal force thereby produced tends to separate the faster ion from the one with a less velocity of migration, and to thus give rise to an electromotive force tending to restore the original uniform distribution. When platinum electrodes are inserted in the tube at given distances, r^ and r., from the axis of rotation, the difference of potential V between them, after a uniform velocity of rotation has been established, may be directly measured by suitable connections with a sensitive galvanometer. When the centrifugal energy E, is in balance with the kinetic energy E^ of the rotating system, the energy-equilibrium is expressed by dE\ = d£k, or, since Ej^zzz i mv' and V= ^-- (setting rand /for radius, and time of one revolution), dEc = m l~-J--] rdr . 562 Reviews and Reports. The centrifugal energy between the electrodes at rj and r^ is, therefore — replacing i// by R, the number of revolutions per second, AEc=.ni {27: itywdr =:2m7:"-{rl—rl')R\ This difference of centrifugal energy must be equivalent to the electrical energy which would be expended in restoring the system to its initial state, i. e., equivalent to V.i for each coulomb — and for each coulomb the actually redistributed mass m of the electrolyte is - — -, — C , — A, when C and A grams of u-^v u-\-v cation and anion respectively migrate per coulomb, and u and v stand for the relative velocities of these ions. The above relation assumes then the form \ u-\-v (rl — rf) R-.iQ-' volts. u-{-v the factor lo"' converting the absolute units into volts because of the definition of a volt-coulomb as lo' ergs. This theory was tested with a solution of cadmium iodide, the following very satis- factory results being obtained. K observed. R corresponding thereto. R employed 155 microvolts 5.2 per second. 5.8 75-3 3-7 3.87 37-4 2.6 2.9 Der zeiiliche Verlauf der Selbsipolarisation in geschlossenen Amalgam- Concenirationselementen. T. Des Coudres : Wied. An7i. 52, 191 (1894). — The velocity of the isothermal diffusion of a dissolved substance depends upon the osmotic pressure of the substance and the frictional resistance of the solvent. The same is true of a metal, as zinc, dissolved in mercury to form an amal- gam. Suppose a zinc electrode to be suspended in a zinc sulphate solution which is in contact with a mass of mercury, and a con- stant current to be passed up to the time /= 7"from the zinc to the mercury. A zinc amalgam is thereby formed. Suppose, further, the mercury to be connected through an electrometer with a second zinc rod in the sulphate solution ; the electrometer indi- cates at first the difference of potential between the poles of a cell Zn I ZnS04 | Hg, and then the rapid fall due to the electrolytic solution-pressure of the zinc in the amalgam. Since for ordinary concentrations the dissolved zinc follows the gas-laws, the observed difference of potential E is given by the relation (Nernst) E'=z ' ^^.loge — . io~^ volts, 2 ^ u in which d signifies the absolute temperature, v the concentration Reviews and Reports, 563 corresponding to the electrolytic solution-pressure of the zinc, and u the concentration of the amalgam. Des Coudres eliminates v by considering the difference ^12 of E.M.F. between two surface- concentrations 7^1 and ^2' J, _ 1.981 .^log^^J^ .10-* volts. 2 7^2 After the time T the deposited zinc diffuses from the surface into the mercury, thus causing the observed E.M.F. to gradually rise. The rate of diffusion, measuring distances x normal to the surface, is (Pick) dt- dx' (^=: diffusion-constant), and when one sets ;r=:o in the integral of this equation, the surface-concentration of the amalgam as a function of the time is obtained. The above expression for E^ is thereby converted into the form ^,2 = 0.029 logi„ '^yjll^^i=====. volts, V^2 — v4 — ^ which permits the calculation of this peculiar polarization-phe- nomenon in its change with the time. That the theory is in excellent accord with the observations appears from the following measurements made at 20° C. 'l- ii. E,„ observed. ^,2, theory. 15' 30" 20' 0.00232 volts 0.00246 volts 20' 30' 344 334 30' 40' 199 215 40' 50' 166 160 50 no' 503 535 0.01444 0.01490 A discussion of the annoying sources of experimental error leads to a theoretical treatment of the theory of the self-polariza- tion of a Zn I ZnSOi | Hg cell as time-function. After a formidable integration of the partial differential equation concerned, which is accomplished by the aid of the Fourier-Cauchy method, a satis- factory relation between the surface-concentration and the time, and in consequence between the E. M. F. and the time, is arrived at. J. E. Trevor. Die Lehre von der Elektrizitat. G. Wiedemann. 2d ed. Vol. II. Large 8vo, 1126 pp. F. Vieweg& Sohn, Braunschweig, 1894. Price : in paper, 28 marks ; bound, 30 marks. The second volume of Wiedemann's great work on electricity has followed the first with a gratifying promptness, and this 564 Reviezc'S and Reports. volume receives a special welcome from the chemical world for the reason that it treats almost entirely of electrochemistry. The style of presentation is the historical-critical, and it is probably true that no contribution of any importance to electrochemical science, appearing before July 1893, has been omitted from the discussion. This fact makes the work extremely valuable as a book of reference, and no less so as a text-book for the electro- chemical specialist. Its usefulness in the latter direction is increased by the unusually independent attitude which the author maintains towards the more recent developments of the hypothesis of free ions in electrolytic solutions. The text is carefully and clearly written and has been well printed. The book is one which no student of electrical or electrochemical matters can afford to be without. PHYSIKALISCH-CHEMISCHE TaBELLEN. H. LANDOLT & R. BORNSTEIN. 2d ed. 4to, 563 pp. J. Springer, Berlin, 1894. Bound in moleskin : 24 marks. One of the most important publications of the year in the literature of physical chemistry is the new edition of Landolt and Bornstein's tables of physical and chemical data. The amount of material included has been increased enormously over that con- tained in the earlier edition, and it has been rendered more valuable by the addition of citations of all the sources from which the data given are drawn. Among the features of more immediate interest to chemists and physical chemists are the tables of atomic weights, of the freezing- and boiling-points of liquids and solutions, of densities, solubilities, vapor-pressures, critical con- stants, specific heats, heats of fusion, of vaporization and of com- bustion, electrical conductivities and dielectric constants. As a book of reference for the laboratory the work is indispensable. The typographical work has been executed with great care and elegance. Die wissenschaftlichen Grundlagen der analytischen Chkmie. W. Ostwald. 8vo, 187 pp. W. Engelmann, Leipsic, 1894. Price : in paper, 4 marks ; bound, 4.50 marks. The little book before us contains an orderly application of the new theory of solutions to the processes of ordinary chemical analysis. By throwing light upon the nature of the operations concerned it is calculated to make them mean to the student very much more than just so many mechanical manipulations. The charmingly simple manner of presenting the theory of the chem- ical equilibria between solutions and the precipitates formed from them, the theory of titration and of indicators, and the familiar theory of the "abnormal reactions" of. certain elements, calls for especial mention. The latter half of the book is a systematic exposition of the more prominent characteristics of the different Reviews and Reports. 565 analytical groups, treated from the point of view of the physical chemistry. The simple character of the little work is its chief merit ; it should be in the hands of every student of chemistry. The Theory of Heat. Thomas Preston. Large 8vo, 719 pp. Mac- millan & Co., New York, 1894. Price $5.50. For the reason that the thermodynamical theory of chemical equilibrium has become one of the most important chapters of theoretical chemistry, the appearance at this time of a clearly written text-book in English upon the subject of heat is to be greeted with pleasure. Preston's Theory of Heat is, in a general and descriptive way, quite a comprehensive work ; it readily sup- plies therefore a good idea of the general concepts and laws of the subject, without the introduction of formidable difficulties of mathematical analysis. The long chapter on pure thermodyna- mics is of exceptional interest and value ; in it a lucid account of fundamental principles is made to lead logically and easily to the theory of the thermodynamical potential, the great energy theory of chemical equilibrium and electrochemistry. As a suitable introduction to the principles upon which the thermodynamical theory of chemical processes stands, the book is to be warmly recommended. Elektrochemie, ihre Geschichte und Lehre. W. Ostwald. Large Svo. Veit & Co., Leipsic, 1894. Price : 2 marks per number of 80 pages. In this new work Ostwald has undertaken to supply an exhaus- tive account of the development of the science of Electrochemistry from the time of the earliest observations of electrochemical action. In the three numbers already issued the experimental researches of Galvani and Volta, the more distinctly chemical work of Ritter, and the discoveries of Davy in electrolysis, are in particular very fully treated. The historical development of the subject is ren- dered especially vivid by liberal quotations from the sources which are cited, and the comment upon these sources is made in a masterly way. The appearance of such a monumental work at the present time is most opportune, for the first act of the electro- chemical drama may be said to have just come to a close, an entirely new epoch in the development of the subject having been inaugurated by the brilliant discoveries of Nernst. The book is to comprise some eight or ten numbers, all of which are promised for this year (1894); their issue will be impatiently awaited. j. e. trevor. Cornell Universitv. JOSIAH PARSONS COOKE. Josiah Parsons Cooke was born in Boston, Massachusetts, on October 12, 1827, and died in Newport, Rhode Island, September 3, 1894. In spite of rather a frail childhood, he entered Harvard College in 1845, was graduated with the class of 1848, and after a year in Europe was appointed tutor in mathematics at the college. Before that time there had been for many years no systematic instruction in chemistry given to undergraduates, and since Mr. Cooke's interest in the subject was well known, he was invited to give a series of chemical lectures in addition to his other work. His own knowledge of chemistry was chiefly self-taught, for natur- ally little was to have been gained from the disjointed lectures of his undergraduate days. During his stay in Europe he studied for some months witn Regnault in Paris ; but with this exception, his broad chemical knowledge was acquired by his unaided efforts. In 1850 he was appointed instructor of chemistry and mineral- ogy at Harvard, and just at the close of the year was made Erving Professor, succeeding Professor Webster. After another six months in Europe he returned with apparatus and chemicals, and made an immediate attempt to establish an experimental course of instruction at Cambridge. At first his work was greatly hampered by the necessity of his giving lectures at the Harvard Medical School in Boston, but for seven years he taught qualita- tive and quantitative analysis to those undergraduates who cared to do the extra work. In 1858, when Boylston Hall was built, qualitative analysis was allowed to count for the bachelor's degree, and Professor Cooke was relieved of the irksome lectures at the Medical School. Ten years later he succeeded in introducing a new elective course on mineralogy ; and since that time the growth of the Chemical Laboratory of Harvard College under his direc- tion and President Eliot's administration has been continuous and rapid. Professor Cooke's untiring zeal collected the subscriptions for the building of the new laboratory and mineral cabinet; his influ- ence was the chief power which raised chemistry to the rank of a liberal study in Cambridge; and his energetic championship of the advantages of a scientific education has had a very far-reach- ing effect upon the whole educational system of America. His elementary lectures to the Freshmen were always extremely pop- ular, and the unusual applause which greeted the Professor each day upon his appearance and his conclusion was quite sincere. Besides these lectures he gave instruction for years in physical' Obituary, 567 chemistry and chemical philosophy, and until recently in quanti- tative analysis as well. In 1882 the degree of LL. D. was conferred upon Professor Cooke by the University of Cambridge, England. He was a very energetic member of the American Academy of Arts and Sciences, having been its corresponding secretary for years before his election as president, in 1892. He belonged also to the National Academy. As a popular lecturer he was unusually successful, having delivered several courses of lectures at the Lowell Insti- tute in Boston, besides many elsewhere. About thirty-five years ago Professor Cooke married Miss Mary Huntington, of Lowell, Massachusetts, who survives him. His summers were usually spent quietly at Newport, in his house overlooking the beach. He was, however, an appreciative trav- eller, and brought home many entertaining reminiscences from his occasional journeys to Europe, Egypt, and the West. Professor Cooke is very well known as a writer. Following are the titles of his books, most of which have gone through several editions. In every case the date of the first edition is given below- Chemical Problems and Reactions (1857). Elements of Chemical Physics (i860). Religion and Chemistry (1864). Principles of Chemical Philosophy (1868). The New Chemistry (1874). The Credentials of Science the Warrant of Faith (1888). Laboratory Practice (1891). Besides these books and a number of addresses upon liberal culture and of memoirs of celebrated men. Professor Cooke wrote many monographs upon scientific subjects. The titles of these papers are given below : The Relation between the Atomic Weights. 1854.— Mem. Am. Acad. [New Ser.] 5 ; Am. J. Sci. [2] 17, 387. On two new Crystalline Compounds of Zinc and Antimony. 1854. — Ibid, [2] 18, 229. On a new Filtering Apparatus. 1854. — Ibid. [2] 18, 127. On the Law of Definite Proportions in the Compounds of Zinc and Antimony. 1855.— //'/(/. [2] 20, 222. Crystalline Form not necessarily an indication of definite Chemical Composition, i860. — Ibid. [2] 30, 194; Phil. Mag. 19, 405. 0.1 the Dimorphism of Arsenic, Antimony and Zinc. 1861. — Am. J. Sci. [2] 31, 191. On the Spectroscope. \%^z.—Ibid. [2] 34, 209. On the Cleavage of Galena. 1863.—/^/^. [2] 35, 126. An improved Spectroscope, \Z()T,.—Ibid. [2] 36, 266. Crystallographic examination of Childrenite. 1863. — Ibid. [2] 36, 257. Crystallographic examination of the acid tartrates of Caesia and Rubidia. l?,64.— Ibid. [2] 37, 70. On a Spectroscope with many prisms. 1865.— /<^/t/. [2] 40, 305. On the Projection of the Spectra of the Metals, 1865,— Z/'/aT. [2] 40, 243. On the Aqueous Lines of the Solar Spectrum. 1866. — Ibid. [2] 41, 17. Vol. XVI.— 41. 568 Obituary. Separation of Iron and Alumina, etc. 1866. — Ibid. [2] 42, 78. Analysis of Uanalite of Rockport. 1866. — Ibid. [2] 42, 73. On Cryophyllite. 1867 — Ibid. [2] 43, 217. On certain Lecture Experiments, etc. 1867. — Ibid. [2] 44, 189. Crystallographic examination of some American Chlorites. 1867. — Ibid. [2] 44, 201. A method of determining the Protoxyd of Iron in Silicates not soluble in the ordinary mineral acids. 1867. — Ibid. [2] 44, 347. Atomic Ratio. iZ6^.—Ibid. [2] 47, 386. Absolute system of Electrical Measurements. 1870. — Franklin Institute. The Vermiculites. 1874. — Proc. Am. Acad. 9, 35. Melanosiderite. 1875. — Ibid. 10, 451. On two new varieties of Vermiculites (with F. Gooch). 1875. — ^^^'^^ ^°' On a new mode of manipulating Hydric Sulphide. 1876. — Ibid. 12, 113. On the process of Reverse Filtering. 1876. — Ibid. 12, 124. Revision of the Atomic Weights of Antimony. 1S77. — Ibid. 13, i. Re-examination of some of the haloid compounds of Antimony. 1877. —Ibid. 13, 72. The Radiometer. 1878.— Pop. Sci. Monthly, May. The Atomic Weight of Antimony. 1879. — Proc. Am. Acad. 15, 251. On Argento-antimonious Tartrate. 1880. — Ibid. [3] 19, 393. On the Oxidation of Hydrochloric Acid Solutions of Antimony in the atmosphere. 1880. — Am. J. Sci. [3] 19, 464. On the Solubility of Chloride of Silver in water. 1881. — Ibid. [3] 21, 221. Additional Experiments on the Atomic Weight of Antimony. 188 1. — Proc. Am. Acad, 17, 1-22 (contains reprints also of a few of the preceding papers). The Boiling-point of Iodide of Antimony, and a new form of Air Thermo- meter. 1881. — Ibid. 17, 22. A simple method for Correcting the Weight of a Body for the Buoyancy of the Atmosphere when the Volume is unknown. 1883. — Am. J. Sci. [3] 26, 38. Possible Variability of the Law of Definite proportions. 1883. — Ibtd. [3] 26, 310. The relative values of the Atomic Weights of Oxygen and Hydrogen (with T. W. Richards). 1887.— Proc. Am. Acad. 23, 149. Additional note on the relative values of the Atomic Weights of Oxygen and Hydrogen (with T. W. Richards). 1888.— //'/Vf. 23, 182. The Chemical Elements. 1S89. — Pop. Sci. Monthly, 34, 733. On a new method of determining Gas-densities. 1889. — Proc. Am. Acad. 24, 202. Theodore William Richards. Vol. XVI. [December, 1894.] No. 8. AMERICAN CHEMICAL JOURNAL, ON ELECTROSYNTHESES BY THE DIRECT UNION OF ANIONS OF WEAK ORGANIC ACIDS. By J. B. Weems. In 1849 Kolbe' made the discovery that certain hydrocarbons are among the products of the electrolysis of fatty acid salts. The interest taken in this observation was enhanced by the belief that these hydrocarbons were the long-sought alcohol radicals, obtained at last in the free state by the decomposing action of the electric current. With the development of the theory of carbon compounds a different interpretation of these facts became neces- sary, and it was soon found that these supposed alcohol radicals were compounds, in each molecule of which two radicals of the same kind were combined. The equation representing such electrolyses, 2RCOsM' = R2 -h CO2 -f 2M', was therefore the expression of a synthesis as well as a decom- position. Since these early experiments by Kolbe, the work of many chemists has confirmed the rule that the electrolysis of salts of monobasic organic acids may in general be expected to yield a certain quantity of a compound R2, formed by the union of residues R from two anions RCO2 — , which simultaneously lose carbon dioxide. But the anions of organic acids, like those from inorganic acids, have at the same time a strong tendency to react 'Ann. Chem. (Liebig) 64, 339. Vol. XVI.-43. 570 " Weems. with the water of the solution in which they find themselves. Consequently a large part of the original acid is always re gen- erated in such experiments, and the quantity of hydrocarbons synthesized in a given time cannot be directly calculated from a knowledge of the strength of the current by an application of " Faraday's Law." Recent investigations by Crum-Brown and Walker' have shown that the conditions are most favorable for the reaction, 2RCO2 — = R«-f2COi, when the solutions employed are concentrated and the electrical density at the positive electrode is high. These chemists also made a new and interesting application of this general reaction in the synthesis of the ethers of certain aliphatic dibasic acids, in which sodium alkyl salts of dibasic acids, lower in the homolo- gous series, were used as the starting-point, 2C2H5CO + 2C0= -f 2Na. Thus sodic ethylic malonate gave a 6o-per cent, yield of suc- cinic ether, and sodic ethylic succinate in turn gave a 35-per cent, yield of adipinic ether, the yields decreasing as the masses of the molecules involved became greater. In volume 15, p. 523, of this journal, in an article on "A New Class of Organic Electro-syntheses," S. P. Mulliken has shown that in the electrolyses of certain weak organic acids which do not contain the carboxyl group, a fraction of the anions unite directly in pairs without any previous splitting, according to the equation, 2RM' = R2-f 2M'. Three cases of this kind are described in detail by Mulliken; viz. a synthesis of ethane-tetracarbonic ether from the sodium salt of malonic ether; of ethane-hexacarbonic ether from methin- tricarbonic ether ; and of tetracetylethane from a concentrated solution of acetylacetone in weak alcohol. The experimental work of the author is a continuation of this investigation of Dr. Mulliken, and was undertaken at his sugges- tion, with a twofold purpose. One object was to ascertain how general a phenomenon the pairing of anions from weak organic acids is. With this end in view the following bodies were electro- lyzed : methyl-malonic ether, ethyl-malonic ether, benzyl-malonic 'Ann. Chem. (Liebig) 361, 107. On Electrosyniheses. 571 ether, acetyl-malonic ether, acetyl-benzoyl-methane, cyan-acetic ether, acelo-acetic ether, acetone-dicarbonic ether, phenyl-methyl- pyrazolon, phthalimide, succinimide, acetamide, and benzamide. Several well known acidic types, like that of the nitro-paraffins, are not represented in this list, it being possible in the time at my disposal to study only a few of the most important and accessible classes of compounds. The other question which I have tried to settle has reference to the nature of the reaction which leads to the formation of the compounds Rs. In nearly all electrosyntheses some oxygen is set free at the positive electrode, and when alkali salts in alco- holic solution are being electrolyzed, the oxidation of organic matter is sometimes sufficient to produce a very noticeable pre- cipitate of mono-sodium carbonate. Since it is known that the conductivity of the weak acids here investigated is extremely poor, it would appear, ^ priori, that some of the syntheses described might be due to oxidation of the replaceable hydrogen in two molecules of acid according to the equation, 2RH -f O = R2 -f- H.O, and not to the reaction, 2RH = R2+H2, as previously assumed. In fact, in a discussion before the Chemical Society, in February, 1893, the action of electrolytic oxygen was suggested by Armstrong' as the most probable explanation of the electrosyntheses of Crum-Brown and Walker. Dr. Walker in reply stated that he did not consider it probable that the forma- tion of synthetic ethereal salts during the electrolysis was due to oxidation at the anode. Murrey^ had shown in case of the elec- trolysis of potassium acetate that there was no sort of proportion- ality between complete oxidation to carbonic anhydride and " partial oxidation " to ethane, as might be expected if the forma- tion of ethane was due to oxidation. The conditions of the elec- trolysis found most favorable for the production of synthetic products were such as would almost insure complete oxidation of the dissolved substance, if the primary action was the decomposi- tion of water into oxygen and hydrogen. While there doubtless are cases in which electrolytic oxygen does produce the same compounds that would result from the ' Chemical News 67, 129. ' J. Chem. Soc. 1892, 10. 572 Wee7ns. direct union of organic anions, one such case being described in the present article, my work gives new proof in other cases that electrosynthesis by the direct union of anions is noi an oxidation phenomenon. Oxidation Experiments with Malonic Ether ^ Methin-tricarbonic Ether, and Acetyl-acetone. The oxidizing agents employed in these experiments were hydrogen peroxide, potassium permanganate in sodium hydrogen carbonate solution, and an acetic acid solution of chromic acid. The fi^st two of these oxidizing agents were selected because it seemed probable that their action would most closely resemble that of the nascent oxygen present in the electrolytic cell. The use of an acid solution of chromic acid was an attempt to vary the conditions of oxidation, after it had been proved that neutral and alkaline oxidizing agents did not give the same products as electrolysis. lo grams of malonic ether and 50 grams of hydrogen peroxide (2-per cent, solution) were heated on the water-bath for 12 hours. The malonic ether had completely disappeared at the end of this time. The solution was then evaporated to dryness on the water- bath. The residue proved to be pure malonic acid melting at ISS", and there was no indication that any ethane-tetracarbonic ether had been formed. To determine what action hydrogen peroxide would have on ethane-tetracarbonic ether, if this should be formed from the malonic ether, the following experiment was carried out : 1.2 grams of ethane-tetracarbonic ether were heated with 30 grams of hydrogen peroxide on a water-bath for 15 hours. The ether was completely dissolved, but, on cooling, separated out in fine needles which proved to be unchanged ethane-tetracarbonic ether. The surprising ease with which malonic ether was saponified by a weak solution of hydrogen peroxide at this comparatively low temperature led me to study the action of distilled water upon it under the same conditions. 5 grams of malonic ether and 30 grams of distilled water were heated on the water-bath for 12 hours and frequently shaken. At the end of this time the ether had completely disappeared. The solution was evaporated on the water-bath to dryness. The On Electrosyntheses. 573 residue proved to be pure malonic acid melting at 133°. During the evaporation a slight odor of acetic acid was present. Hjelt' in some experiments on the action of water on malonic ether, heated malonic ether with water in a sealed tube at 150° and obtained as a result malonic acid, acetic acid, and acetic ether. From this he concluded that malonic ether is first changed to acetic ether and the latter to acetic acid. But from my experiments on malonic ether at temperatures below 100° it appears that the ether is first completely saponified to malonic acid, which at a higher temperature loses carbon dioxide, forming acetic acid. The presence of acetic ether in experiments with sealed tubes can be readily explained as the result of the action of acetic acid on the alcohol from the malonic ether. The employment of acid sodium carbonate with potassium permanganate in oxidation experiments, which has been much used by von Baeyer in his work on the hydrophthalic acids, was found very convenient, as such a solution contains no mineral acid or caustic alkali which would saponify the ethers experi- mented with. Experiments on formic acid showed that in such a permanganate solution, two molecules of potassium permanganate contain four available atoms of oxygen. 100 grams of a saturated solution of sodium hydrogen carbonate were placed in a flask with ten grams malonic ether. A two-per cent, solution of potassium permanganate was then run into the mixture with constant agita- tion, in quantity sufficient to oxidize one atom of hydrogen in each molecule of malonic ether. After standing for several hours the oxidized mixture was fihered and extracted with ether. The ethereal extract contained 8.3 grams of malonic ether, but no ethane-tetracarbonic ether. An examination of the aqueous solution which had been extracted with ether showed that oxalic acid was the chief product of the oxidation of malonic ether. The addition of potassium permanganate to a mixture of malonic ether and sodium hydrogen carbonate until a permanent color is pro- duced, causes the complete destrucrion of the malonic ether with- out formation of any ethane-tetracarbonic ether. In the oxidation by chromic acid, 2.07 grams of chromic acid dissolved in 20 grams of acetic acid were added to 10 grams of malonic ether in 20 grams of acetic acid. The solution was allowed to stand for three hours, filtered, acetic acid removed by evapora- 1 Ber. d. chem. Ges. 13, 1949. 574 JVeems. tion, and the residue distiHed. No ethane-tetracarbonic ether had been formed. The oxidation experiments on methin-tricarbonic ether and acetyl-acetone, which, on electrolysis, yield ethane-hexacarbonic ether and tetra-acetyl-ethane respectively, were made in the same manner as those on malonic ether, and do not require detailed description. The partial saponification of methin-tricarbonic ether gives malonic ether, and its oxidation-products were found to be the the same as given by this body. They contained no trace of ethane-hexacarbonic ether. The results of the oxidation experiments on acetyl-acetone and aceto-acetic ether were also of a negative character. Consiiiution of Ethane-hexacarbonic Ether. In the account of his synthesis of this new hexacarbonic ether, Mulliken' employs the following equations to account for its for- mation: 2 II ^^^ =Na.-f2C=(C020H6> C (CO.OH5)2 C = (CO»CsH5)3 -^ I C = (CO^GHs)^ While from analogy with the reaction that takes place in the syn- thesis of ethane-tetracarbonic ether from malonic ether, in which the carbon atoms of two anions unite, the correctness of this equation appears almost certain, the constitution of ethane-hexa- carbonic ether nevertheless requires additional proof; for several cases have been lately described by Michael," Nef,^ and Claissen,* in which the sodium in compounds of this class may be directly replaced by an acid radical. If it were possible for the reaction to take this second course, a compound having the following symbol would result, (C2H502C)2'^^ p. p. P^(CO=GH6). Saponification of the ether under consideration appeared to offer the easiest means of determining its constitution. If it is a true ' This Journal 15, 527. ^ Ibid. 14, 489-491. ' Ann. Chem. (Liebig) 266, 79. < Ibid. 877, 166. On Elecirosyntheses. 575 hexacarboaic ether, saponification should give a salt of the corre- sponding- acid, or bodies closely related to it; while, if its consti- tution were represented by the less probable formula, the action of potash would be likely to split the molecule into two molecules of the original substance, which as Conrad and Guthzeit' have shown, on saponification gives potassium malonate, CHCCO.OHOs -f 4KOH = CHCH.CH.(KC02> + 2K2CO3 + 6GH60H. The identification of this salt was completed by converting it into the corresponding silver salt, which on treatment with ethyl iodide gave the well-known ethane-tetracarbonic ether. The same salt was also prepared by saponification from a pure speci- men of ethane-tetracarbonic ether, and analyzed with the follow- ing results: I. 0.5233 gram of crystallized salt contained 0.0478 gram H2O. II. 0.3297 gram of crystallized salt contained 0.0303 gram H2O. III. 0.5233 gram of crystallized salt contained 0.41 17 gram K.SO4. IV. 0.3297 gram of crystallized salt gave 0.2912 gram K2SO4. IV. 39-71 Ethane-tetracarbonic Acid. Conrad and Bischoff' have studied the action of potash upon ethane-tetracarbonic ether and found that the resulting potassium salt, when decomposed with acetic acid, gave an oily substance, from which they separated ethenyl-tricarbonic acid melting at 155°. Saponification of the ethane-tetracarbonic ether by boiling hydrochloric acid gave the same product; (C^H.COO^CH — CH(C02C2H6)2 + H2O = (HC02).CH — CH2CO2H -\- C02-f 4GH5OH. 1 Ann. Chem. (Liebig) 314, 71. Calculated for C2H,,(C02K)4.2H50. I. Found. II. III. H20 9-13 9.21 9.21 K 39.67 ... 39.6 On Eledrosyntheses. 577 Recently Buchner/ by saponifying ethane-tetracarbonic ether, obtained a substance melting at i67°-i69° which, on heating, was transformed into succinic acid. Buchner's work was not known to me when I was engaged in these experiments, and as his description of the properties of the acid is very incomplete, my observations in regard to its formation and its properties may be of some interest. Desiring to ascertain whether, as the experi- ments of Conrad and Bischoff seemed to show, ethane-tetracar- bonic acid is really so unstable that it cannot be liberated without losing carbon dioxide and being converted into a tri- or dibasic acid, 0.7892 gram of the potassium salt of ethane-tetracarbonic acid was dried at 130° and dissolved in 20 cc. of distilled water. Just enough sulphuric acid was then added to combine with the potassium of the salt, the flask being connected with potash-bulbs by means of a long tube half filled with calcic chloride. After the mixture had been allowed to stand several hours, air was drawn through the apparatus. No carbon dioxide had been given off. The mixture was then warmed on the water-bath for one hour. In this time the acid had lost 17.16 per cent, of carbon dioxide. Conversion of the ethane-tetracarbonic acid into the ethenyl-tricarbonic acid would involve a loss of 12.29 P^r cent, of carbon dioxide. Ethane-tetracarbonic acid is therefore quite stable at ordinary temperatures, although readily splitting off carbon dioxide in the manner indicated by Conrad and Guthzeit when warmed. By heating the aqueous solution as long as the potash-bulbs increased in weight, it was found that 0.1924 gram of carbon dioxide had been lost ; a loss which closely corresponds to that (0.1939 gram) required by the equation, C=H+ 2CO2. The contents of the flask at the close of the experiment were extracted with ether. The ethereal solution, on evaporation, gave a residue of crystals of succinic acid, melting at 185°. The preparation of free ethane-tetracarbonic acid is easily accomplished by adding to a 15-per cent, aqfieous solution of its potassium salt the calculated quantity of sulphuric acid and then allowing the solution to evaporate spontaneously. All heating must be carefully avoided. The residue is extracted with ether to separate the acid from the potassium sulphate. ' Ber. d. chem. Ges. 35, 1157. 578 Weems. The ethereal solution on evaporation deposits the acid in beau- tiful tablets, apparently of the monoclinic system, which melt with decomposition at i69°-i7i°. The fused mass after solidification remelts above i8o° and consists of succinic acid. Ethane-tetra- carbonic acid is readily soluble in water, alcohol and ether, but almost insoluble in benzene and acetic acid. Its clear, colorless crystals gradually break down to a white powder on exposure to the air. Combustions of the acid gave the following results : I. 0.2504 gram of theacid dried invacuo gaveo.3213 gram COi and 0.0732 gram H2O. II. 0.2488 gram of the acid dried at 100° gave 0.3163 gram CO2 and 0.0726 gram H2O. Calculated for Found. C2H3(COaH)4. I. II. Carbon 34.95 34.99 34.67 Hydrogen 2.92 3.24 3.24 Reagents gave the following precipitates with a solution of its potassium salt : — Argentic nitrate : White precipitate soluble in ammonia, not reprecipitated by nitric acid. Barium chloride : White precipitate soluble in nitric acid. Calcium chloride : White precipitate. Plumbic acetate: White precipitate insoluble in boiling water. Mercurous nitrate: White precipitate insoluble in boiling water. Mercuric chloride: White precipitate on warming the solution. Ferrous sulphate : Light-brown precipitate, does not appear at first. Zinc sulphate: White precipitate which increases on heating. Magnesium sulphate: No precipitate. Electrolysis of Methyl-malonic Ether. — {Dimetliyl-ethane-ietra- carbonic Ether'). CHsCCOsCsHs cw " GH.ol]oNa =^^' + ^-'^C(C0=OH.> CHs,^ -, ^^^CHs ■^ (CO-.C^Hs),^^ ~'-v^(C02C2H6)2. In this electrolysis, as well as in those which will be spoken of further on, the electrolyte was placed in a small porous cylindrical cell. The anode, a spiral of thick platinum wire, was suspended On Elecirosyniheses. 579 in this cylinder, which was itself placed within a beaker of about the same height, but of slightly greater diameter, filled with the solvent used for the experiment. The cathode, a large piece of platinum foil, was fitted close to the sides of the beaker and sur- rounded the inner vessel. The supply of electricity was obtained from a storage battery having a maximum electromotive force of about 40 volts. When aqueous solutions of metallic salts were contained in the cell it was often necessary to reduce the current by placing a rheostat in the circuit. Three amperes was the strongest current used in any experiment. A more powerful cur- rent caused too much heating, even when the solution was well cooled by ice-water. 60 grams (one mol.) of methyl-malonic ether were dissolved in 75 grams of absolute alcohol containing 6.6 grams (one atom) of metallic sodium. This solution was subjected to the action of a current of i of an ampere for 28 hours, that is, for a time slightly greater than would theoretically be required completely to decompose the sodium salt placed in the cell. The porous cell served a useful purpose by preventing the saponification of the ethereal compounds, the alkali produced by the electrolysis all accumulating in the outer vessel. In the experiments on methyl-malonic ether a total of 190 grams of substance was consumed. After electrolysis the alcoholic solu- tion was neutralized with dilute sulphuric acid, the alcohol evapo- rated on the water-bath ; the residue extracted with ether and water; the ethereal extract dried with calcium chloride; the ether removed by distillation; and the resulting oil then distilled under reduced pressure. Two grams of oil were thus finally obtained which boiled under a pressure of 15 mm. at 167°. The principal portion of the oil boiled at 110° under the diminished pressure and was methyl-malonic ether. The molecular weight of the higher-boiling portion was determined by the method of Raoult : Wt. of substance. t° — t"' . Benzene. Mol. wt. found. 0-3095 gram, 0.371 13.2624 grams. 314 0-6955 " 0.801 " 321 0.9618 " 1.073 " 341 The oil was then analyzed. I. 0.2I20 gram substance gave 0.4160 gram CO2 and 0.1368 gram H2O. 580 IVeems. II. 0.4640 gram substance gave 0.9130 gram CO2 and 0.2936 gram H2O. Calculated for C ilciilated for Calculated for 2CoH2(C02C3Hj)4 + Found. (CH3)5CH3(CO,C3H6)4. CjH3(COaC„H6)4. (CH3)2C5(CO,C2Hj4. I. II. Carbon 55.49 52 68 53.71 53.63 53.66 Hydrogen 7.53 6.89 7.13 7.17 7.03 It will be seen from these figures that the oil boiling at 167° under reduced pressure contained very nearly the quantity of carbon and hydrogen that would be found in a mixture of two molecules of ethane and one molecule of dimethyl-ethane-tetra- carbonic ethers. That such a mixture was actually present was rendered probable from the result of the molecular weight deter- minations, for such a mixture would have a molecular weight of 327, while the mean result found was 325. The formation of this mixture is easily explained. The methyl- malonic ether used was prepared in the usual way from malonic ether, sodium ethylate and methyl iodide. The boiling-points of methyl-malonic ether and of malonic ether differ from each other by only about one degree, so that it is impossible to separate the two by the most careful fractionation. It is stated by MuUiken that in the electrolysis of sodium malonic ether the yield of ethane-tetracarbonic ether is 25 per cent. In my experi- ments only about 2 per cent, of the methyl-nialonic ether used was converted into the oil boiling at 167° under reduced pressure. From this experiment and that which follows there is no doubt that electrosyntheses from the homologues of the malonic ether are made less readily than when malonic ether itself is used as the starting-point. Consequently, in the electrolysis just described a considerable part of the malonic ether present as an impurity in the methyl-malonic ether was changed into ethane-tetracarbonic ether, and its quantity was greater than that of the desired product. By taking advantage of the remarkable stability of ethane- tetracarbonic ether towards weak alcoholic potash, which has already been referred to in this paper, it was found possible to separate the mixture of dimethyl-ethane-tetracarbonic and ethane- tetracarbonic ethers by partial saponification. The portion of the high-boiling oil recovered from the molecular weight determina- tions was warmed with five times the calculated quantity of a ten- per cent, alcoholic solution of potash for three hours. An insol- 0?i Elecirosyntheses. 581 uble potassium salt separated out in distinct needles. This was filtered off and washed with alcohol. Only about one-third of the oil had been saponified. The salt was dried at 160° and the potassium determined as sulphate. I. 0.1138 gram substance gave 0.1026 gram K2SO4. II. 0.1282 gram substance gave 0.1156 gram K2SO4. Calculated for Found. (CH3)2C2(COsK)4. I. II. K 40.41 4049 40.48 Electrolysis of EthyUmalonic Ether {^Diethyl- ethane-tetracarbonic Ether), OH6CCO2OHB .pXT C.HsOCONa >^(CO.C.H5). ^ (C2H5C02)2^'^ ^>^(C020H5>' This electrolysis was conducted in the same manner as that of the methyl-malonic ether. 1.5 grams {h atom) metallic sodium were dissolved in 50 grams of absolute alcohol, and when the solution had become cold, 25 grams of ethyl-malonic ether were added to it and the solution placed in a porous cell. A current of \ ampere was passed through the solution for 17 hours. The mixture was then neutralized with dilute sulphuric acid mixed with water, extracted with ether, and the residual oil obtained from the extract distilled under reduced pressure. About 0.5 gram of this oil boiled above 150" under a pressure of 15 mm. The remainder was ethyl-malonic ether. A molecular-weight determination of the oil boiling above 150°, by Raoult's method in a benzene solution, indicated that it was a mixture of diethyl-ethane-tetracarbonic ether and ethyl-malonic ether. Wt. of substance. Benzene. <° — 1°' . Mol. wt. found. 0.4025 gram. 14-45 grams. 0.45° 309 0.2583 (recovered). 14.313 0.293° 309 The molecular weight of ethyl-malonic ether is 188, and that of diethyl-ethane-tetracarbonic ether is 374. The material used in the above molecular-weight determinations was saponified by boiling with five times the calculated quantity of a ten-per cent, solution of potash, and the potassium salt obtained dried at 160° and analyzed. 582 Weems. I. 0.2121 gram of this salt gave 0.1779 gram K2SO4. II. 0.1373 gram of this salt gave 0.1152 gram KsSO* Calculated for ,H.),C,(CO,K)4. Found. I. II. 37-67 37.65 37-67 K These results show that the anticipated reaction, 2C9Hi504Na = Naa -|- CisHsoOs, takes place when ethyl-malonic ether is electro- lyzed, but that the yield of diethyl-malonic-tetracarbonic ether is very small, amounting, according to this experiment, to only 2 per cent, of the weight of the ether electrolyzed. Electrolysis of Benzyl-malonic Ether. 2.1 grams (f atom) of metallic sodium were dissolved in 35 grams of absolute alcohol, and the cooled solution mixed with 30 grams (one mol.) benzyl-malonic ether. This solution con- ducted very poorly, a current of 3^ of an ampere which gradually increased to 2V of an ampere was passed through the solution for 105 hours. The oil obtained after neutralizing with dilute sul- phuric acid and extracting with ether weighed about 25 grams, and was unchanged benzyl-malonic ether. Even the highest- boiling portion of this oil, when distilled under reduced pressure, was found to have a molecular weight, according to a determina- tion by Raoult's method, as low as 225. The calculated mole- cular weight of benzyl-malonic ether is 250. Electrolysis of Acetyl-malonic Ether. The acetyl-malonic ether used for this experiment was prepared by the method of Michael' and boiled at 120° under a pressure of 15-17 mm. It was freed from traces of malonic ether by shaking with a solution of sodium carbonate, in which, unlike malonic ether, it is soluble, and from which it is precipitated unchanged by acids. 21 grams of acetyl-malonic ether were dissolved in 126 grams of 50 per cent, alcohol, and the electric current passed through the solution for two days. The strength of the current, which at the beginning of the experiment was f of an ampere, after 20 hours had decreased to 2V of ^n ampere. The anticipated product of this electrolysis, diacetyl-ethane-tetracarbonic ether, from its con- stitution should evidently be a neutral body and consequently 1 This Journal 14, 495. On Electro syntheses. 583 insoluble in sodium carbonate, like malonic ether. The contents of the electrolytic cell were repeatedly shaken with ether and a 10 per cent, solution of sodium carbonate. All unchanged acetyl- malonic ether was dissolved by the carbonate, while the neutral compounds went into the ethereal extract. Examination of the neutral body contained in the ethereal solution proved it to be simply malonic ether. It boiled between 90°-ioo° under a pressure of 10 mm. and weighed 7 grams. No evidence that the anions, the formation of which from acetyl- malonic ether might be expected, had combined with one another was found. The malonic ether obtained from this experiment may have been formed in the cell by action of water during the electrolysis, or afterwards, while the contents of the cell were being shaken with ether and sodium carbonate. As 11 grams of acetyl-malonic ether were recovered from the solution of sodium carbonate, and 7 grams of malonic ether represented 9 grams of acetyl-malonic ether, no appreciable amount of the acetyl com- pound electrolyzed could have been changed to other compounds. Electrolysis of Aceto-acetic Ether. — {^Diacetyl-succinic Ether). 2CH3C(ONa) :CHC02GH6=: 20 grams of aceto-acetic ether were added to 20 grams of 50- per cent, alcohol in which 0.3 gram of metallic sodium had been previously dissolved. A current of one ampere was passed through the solution for two hours. By the electrolysis of a dilute alcoholic solution of pure aceto-acetic ether Mulliken^ obtained a product which gave a somewhat uncertain test for diacetyl-succinic ether by Knorr's pyrrol-reaction with a pine splinter. The experiment was repeated in order to ascertain whether a larger quantity of diacetyl-succinic ether could be obtained by using the sodium salt. At the close of the electrolysis, in which a porous cell was not used, hydrogen sodium carbonate had been deposited as an incrustation on the anode, indicating oxidation or a splitting off of a portion of the aceto-acetic ether. After neutralizing the solu- tion with dilute sulphuric acid and extracting with ether, the oil from the ethereal extract was tested for a i :4-diketone by boiling > This Journal 15, 532. 584 IVeems. a pine-splinter in its acetic-acid solution to which ammonium acetate and dilute sulphuric acid had been added. The splinter assumed a deep red color that indicated the presence of a consid- erably greater quantity of diacetyl-succinic ether than was found in the case of the electrolysis of simple aceto-acetic-ether solution. Electrolysis of Aceto7ie-dicarbonic Ether. It was thought probable, since diacetyl-succinic ether had been obtained from aceto-acetic ether, that an analogous diketone might be prepared by the electrolysis of acetone-dicarbonic ether. This expectation was not realized. The acetone-dicarbonic ether was prepared according to the directions of von Pechmann' and boiled at I45°-I50° at 20 mm. pressure. The ether was converted into the potassium salt, which is easily soluble in water without decomposition. 12 grams of this salt were dissolved in 15 ccm. of distilled water and electrolyzed in a small porous cell. The current, which was at first about one ampere, entirely ceased after a few hours, all the potassium having passed into the water of the outer vessel. The oil which remained in the cell was acetone-dicarbonic ether, which distilled below 150° at 20 mm. pressure. Knorr's pine-splinter reaction for a i : 4-diketone gave a negative result when applied either to the oil before distillation or to the different fractions of the distillate. Electrolysis of {\)-Phenyl-{^-methyl-{;^-pyrazalon. — {Phenyl-y methyl-bis-pyrazalon). CeHs CeHs CbHs N N N / \ _ / \, / \ ^ N CO — N COOC N CH3 — C — CH2 CH3— C — C— C — C — CH, + H., H H or, 2C10H.0ON2 + O = C20H.8OSN4 + HsO. 3J grams of (i)-phenyl-(3)-methyl-(5)-pyrazalon, melting at I28°-I29°, were dissolved in 20 grams absolute alcohol and electrolyzed in a small beaker. The current soon increased from \\^ to \ of an ampere. The electrolysis was continued for 24 ' Ann. Chem. (Liebig) 261, 159. On Electrosyntheses. 585 hours. The solution was then of a deep red color and contained a white precipitate, which was filtered off and thoroughly washed with alcohol. This white powder, which was quite insoluble in alcohol and ether, gave the following characteristic reactions for phenyl-methyl-bis-pyrazalon : It was readily soluble in a solution of sodium hydroxide, and such a solution containing only a slight excess of alkali gave a steel-blue precipitate with ferrous salts. When boiled with ferric chloride it was changed to pyrazalon- blue. It could be heated to a temperature above 250° without showing signs of decomposition. The yield of bis-pyrazalon in this experiment amounted to ten per cent, of the phenyl-methyl- pyrazalon. In the electrolysis of phenyl-methyl-pyrazalon we have an instance of electrosynthesis which may be explained, so far as the facts connected with it have been studied, either as the result of an oxidation or of direct union of anions. From the ease with which phenyl-methyl-pyrazalon is converted into the bis-pyrazalon by weak oxidizing agents like ferric chloride it is probable that this synthesis is, in a large measure at least, the result of oxidation. Nevertheless, the distinctively acid character of phenyl-methyl- pyrazalon renders it by no means improbable that bis-pyrazalon is at the same time formed by the second method of synthesis. Electrolysis of Cyan-acetic Ether. *Haller' has carried out numerous syntheses in which the sodium salt of cyan-acetic ether was made to play the same role as the sodium salt of aceto-acetic ether in the familiar acetoacetic-ether syntheses. Thus sodium cyan-acetic ether and acetyl chloride gave aceto-cyan-acetic ether, while sodium cyan-acetic ether and chlorocarbonic ether gave cyan-malonic ether. In the hope that the anions from this sodium salt might unite during electro- lysis, 15 grams of the pure sodium salt are dissolved in 30 grams of distilled water, and the solution electrolyzed in a porous cell for four and one-half hours with a current of ^— ^ of an ampere. The contents of this cell, after neutralization and extraction with ether, consisted of an oily residue, which was distilled under a pressure of 15 mm. The distillate consisted of cyan-acetic ether and a substance boiling at i6o°-i65°, which solidified on cooling and melted at 67°. Analysis showed it to be cyan-acetic acid. , 'Ann. chim. phys. 16, 403. Vol. XVI.-44. ^86 We ems. There was no indication that any dicyan-succinic ether had been formed. Electrolysis of Succmimide. Walden' in his work on the affinity constants of organic acids found that an aqueous solution of this typical imide is a non-con- ductor of electricity. One atom of hydrogen in succinimide can, however, be readily replaced by alcoholic radicals, by treating its sodium or silver salt with alkyl iodides,^ just as the homologues of malonic ether can be prepared by the same general method. Since Mulliken had shown that, while a solution of malonic ether does not conduct electricity, its sodium salt in alcoholic solution is a good electrolyte, it was thought possible that electrosyntheses might likewise be made from the metallic compounds of acid amides and imides. Walden in his experiments with an aqueous solution of the sodium salt of succinimide found as the value of the difference /I32-1024, 20.54, which indicates that this compound is very easily decomposed by the hydrolytic action of alkali, giving rise to the sodium ammonium salt oiz-dibasic acid, i.e. succinic acid. It was, therefore, decided to use in my experiments the easily solu- ble mercury salt of succinimide. The succinimide was prepared according to a method recently described by Verley.' This method gives a yield of 80 per cent, and may be carried out by heating a mixture of one molecule of potassium succinate and two molecules of ammonium chloride in a glass retort immersed in a bath of fusible metal. The succinimide, which melted at 125°- 126°, was converted into the mercury salt by treating a concen- trated solution with mercuric oxide and recrystallizing from alcohol.* To test the purity of the mercury salt, two determina- tions of mercury were made, with the results that the first gave 50.64 and the second 5069, instead of 50.50 of mercury as calcu- lated for the formula C8Hs04N2Hg. 10.02 grams of this mercury salt were dissolved in 80 grams of water. At the beginning of the electrolysis the current was 0.5 of an ampere, but after 22 hours, when all the mercury had been precipitated, from some obscure secondary reaction it had increased to 06 of an ampere. The solution was evaporated and 4.79 grams of pure succinimide, melting at 125°, were recovered. Since the mercury salt electro- 1 Ztschr. phys. Chem. 8, 484. * Comstock and Kleeberg: This Journal 12,493; Comstock and Wheeler : Ibid. 13, 520. > Bull. Soc. Chem. 10, 690. ■» Menschutken : Ann. Chem. (Liebig) 162, 171. 071 Elecirosyntheses. 587 lyzed contained only 5.01 grams, the electrolysis must have pro- ceeded as in the case of a mercury salt of an inorganic acid. CsHsNsO^Hg + H2O = 2aH6N02 + Hg -h O. Electrolysis of Phihalimide. Solutions of 10 grams (one mol.) of phthalimide melting at 220° were dissolved in 50 grams of alcohol, and 1.6 grams (one atom) of sodium dissolved in 20 grams of alcohol were mixed and electrolyzed in a porous cell by a current which was at first \ of an ampere. After eight hours the cell contained nothing but pure phthalimide, melting at 229°. The phthalimide used in the experiment was all recovered. Electrolysis af Acetamide and Benzafnide. Although the concentration of the solution and the density of the current in the experiments with acetamide were very favorable for the union of the anions, as in the electrolysis of the acid imides just described, only negative results were obtained. Ten grams of Kahlbaum's acetamide, melting at 83°, were dissolved in 50 grams of distilled water. The solution showed a somewhat surprising conductivity. A current of 0.8 of an ampere was passed through the solution for 14 hours. The solution was then evaporated and nearly 10 grams of acetamide with unchanged melting-point were recovered. Fearing that the high conductivity of the acetamide might be due to an impurity being present, its mercury salt was prepared by dissolving mercuric oxide in an aqueous solution of acetamide. 22 grams of the mercury salt recrystallized from water were dissolved in 30 grams of water and electrolyzed for 51 hours. The current, which was at first only -^ of an ampere, increased as the mercury separated out at the cathode and free acetamide was formed in the solution, until it amounted to 0.1 of an ampere. The mercury salt electrolyzed corresponded to eight grams of acetamide. Seven and one-half grams were recovered. An attempt was made to electrolyze an alcoholic solution of benzamide and an aqueous solution of its mercury salt ; but the first solution was such a poor conductor, and the mercury salt so slightly soluble in water, that no satisfactory results were obtained. 588 Weems. Summary. The most important results of this investigation may be briefly summarized as follows: 1. The action of various oxidizing agents on malonic ether, methin-tricarbonic ether, acetyl-acetone, and aceto-acetic ether, doesnot give the compounds obtained from these bodies by electro- lysis. In other words, these electrosyntheses are best explained as the resuh of the direct union of anions and not of an oxidation. 2. Pairing of anions has been shown to be of particularly common occurrence in the electrolysis of sodium compounds of derivatives of malonic ether. Dimethyl and diethyl ethane-tetracarbonic ethers were thus obtained from methyl and ethyl malonic ethers. The yields from these homologues were, however, much smaller than from malonic ether itself, or from methin-tricarbonic ether. The last-named compound was proved by saponification to have the constitution C2(C02C2H6)6. No indications of a pairing of anions in the electrolysis of benzyl and acetyl malonic ether could be detected. 3. Pairing of anions occurs in the electrolysis of certain com- pounds of the aceto-acetic-ether type (aceto-acetic ether and acetyl-acetone'), but failure to secure such synthetic products in the experiments with acetone-dicarbonic ether and acetyl-benzoyl- methane showed that this reaction can hardly be called a general one. The formation of bis- phenyl-methy 1-pyrazalon in the electro- lysis of phenyl-methyl-pyrazalon may be partly an electrolysis of this type or may be due wholly to oxidation. 4. The electrolysis of acid amides (acetamide, benzamide, suc- cinimide and phlhalimide), or their sodium or mercury salts, yields the original amides or imides without any pairing of anions. Chemical Laboratory, Clark University, ytine, 1894. 1 The electrical conductivity of a very pure specimen of acetyl-acetone lately prepared by Dr. Mulliken and measured by Mr. Alfred Wakeman, of the Sheffield Scientific School, was found to be no greater than that of pure water. It is, therefore, quite likely that the real electrolytes in Miilliken's electrosynthesis of tetra-acetyl-ethane were traces of salts of acetyl- acetone derived from bases in the glass of the electrolytic cell, or from the dilute alcohol used as a solvent instead of acetyl-acetone itself. Pentosans in Plants. 589 PENTOSANS IN PLANTS. By G. de Chalmot. I. — Pentosans in Seeds. Are there pentosa^is in the plants zvhich are formed from seeds which germinate in darkness ? It has become necessary to answer this question since Cross, Bevan and Beadle' in a recent article assert that the furfurol- yielding compounds in the sprouts which are developed from barley in darkness are so-called oxycelluloses, and are not pentosans. " These sprouts yield 5 per cent, of furfurol, but they do not show the usual reactions that indicate pentosans." If this assertion be true it would be of fundamental importance. For the furfurol-yielding bodies in barley grown in light are, at least for the greater part, pentosans. The straw of small grain is in fact one of the very best materials for obtaining abundant xylose. I found, however, that the furfurol-yielding compounds in corn and barley plants which have been developed from seeds in darkness do only very partially belong to the celluloses ; far the larger part belongs to the nitrogen-free extract. Thus I found in corn plants (seeds, stems and roots) 12 days old and which had grown in darkness, total furfurol-yielding bodies, 13,9 per cent.,^ of which are contained in the raw fiber, 1.5 per cent. In the stems of barley plants 7 days old, from seeds grown in darkness, I found: total furfurol-yielding bodies 10. i per cent., of which are contained in the raw fiber, 2.0 per cent. The raw fiber was obtained by the usual Weende method. Messrs. Cross and Bevan communicated to me in a private letter that they do not use the Weende method, but use another, the main features of which are a long digestion in cold i-per cent, sodium-hydroxide solution, and after washing out the sodium hydroxide, a second long digestion in i-per cent, hydrochloric acid in the cold. It is very probable that they therefore do not extract as much matter as is done by the Weende method (digestion with boiling i i-per cent, hydrochloric acid for half an hour, and afterwards for the same time with boiling ij-per cent, sodium-hydroxide solution). Consequently their fiber may also 1 Ber. d. chem. Ges. 27, 1065. 2 Furfurol x 2. 590 JDe Chalmot. contain more furfurol-yielding bodies. I shall, however, prove that pentosans are contained in sprouts developed in darkness. Stems of barley germinated and grown in darkness for seven days were twice digested with 2-per cent, ammonia solution for two hours, and after removing the liquid as much as possible, they were digested for sixty hours in the cold with 5-per cent, sodium- hydroxide solution. The solution was separated from the stems and hydrochloric acid and alcohol were added. A gum separated, which was collected, washed with alcohol, and dried over caustic potash. The dried gum was grayish in color and looked like impure wood-gum. The very smallest amount of this gum gave a distinct color reaction with the phloroglucol reagent. The absorption-line between the lines D and E of the spectrum,' which is character- istic for the pentoses, was very plainly discernible. I heated a small amount of this gum (4 grams) for six hours on the water-bath with 50 cc. of 3-per cent, sulphuric acid. The liquid was neutralized with barium hydroxide. The barium sul- phate was filtered off. Phenylhydrazine hydrochloride and sodium acetate were added to the liquid, which was heated at 70° C. in the water-bath. A yellow osazone was formed after some time, which was collected on a filter. This osazone was then pressed between layers of filtering paper, purified by washing with small amounts of alcohol, and dried over sulphuric acid. This osazone melted at i40°-i45° C. (uncorr.). Since I was not sure whether it was simply an impure osazone of a pentose or a mixture of osazones, I boiled 0.77 gram of the osazone with 50 cc. of water for two hours. A small part was dissolved and separated by cooling. Its melting-point was 153° (uncorr.). The rest of the osazone was once more boiled with 40 cc. of water, and the undissolved part was dried and then boiled with so small an amount of alcohol that it was nearly but not quite fully dissolved. The alcoholic liquid was poured off and allowed to stand over night. It then deposited a small amount of a tarry matter, from which it was poured off again and mixed with a small amount of water. An osazone was precipitated, which was washed with water and alcohol, and melted at 153° C. (uncorr.). The original osazone seems therefore to have consisted of the osazone of a pentose only. The melting-points of both xylosazone and arabin- osazone are isS^-ieo® C. > Tollens : Landw. Vers. Stat. 39, 443. Pentosans in Plants. 591 I could not find out which of the pentoses was present, since the quantities with which I worked did not admit of further tests. The stems of barley plants grown in darkness, therefore, do contain pentosans, and once more it has been proved that vege- table matters which yield furfurol contain pentosans. It is of course not impossible that besides the pentosans there are other substances which yield larger amounts of furfurol. However, the presence of these other substances has not been proved, and I therefore for the present hold that the essential part of the fur- furol which is yielded by the hydrolysis of vegetable materials with hydrochloric acid (of 12 percent.) originates from pentosans. The observation of Cross, Bevan and Beadle that young barley plants grown in darkness do not show the color-reaction with phloroglucol required further explanation. When I made extracts of the barley stems above referred to with diluted hydro- chloric acid or diluted sulphuric acid, and when I discolored the filtrates with animal charcoal and heated them with the phloro- glucol reagent, the liquid became brown, and a precipitate was formed which made invisible any red color which may have been formed. It has already been known for a longer period that hexoses form such brown dyestuffs when heated with the phloro- glucol reagent. Soluble hexoses are contained in large amounts in the young barley plants, and to them this brown color and pre- cipitate is due. It moreover seems to be mainly fructose or cane sugar which makes it impossible to recognize the pentose reaction. I took the gum which I extracted from the barley plants, and which showed the pentose-phloroglucol reaction very strongly, and heated it with the phloroglucol reagent and a little starch. I was then able to obtain a cherry-red color by heating very gently. In this case the pentose reaction (red) set in before the glucose reaction (brown). When I used instead of starch cane-sugar, I did not obtain any red color, but the liquid at once became brown. The fructose (of the cane-sugar) therefore reacts with the phloro- glucol at the same time as, or earlier than, the pentose. The Conduct of the Pentosans by the Germination of Seeds in Darkness. In a former paper' I gave a preliminary report on this subject. The study of the changes in the amount of vegetable matters 'This Journal 15,276. 592 De Chalmot. during the germination of seeds in darkness is of fundamental importance for obtaining an insight into the physiological value of these matters. I have therefore continued my investigations in this direction, and the facts which I have pointed out confirm and enlarge my former results. In my former paper I communicated : 1. That when corn and peas germinate in darkness, the young plants contain more pentosans than the seeds. 2. That the amount of pentosans in the seeds proper decreases. 3. That the young plants of Trapceolum majus contain consid- erably less of the pentosans than the seeds. 4. That it was not decided whether the increase in pentosans in corn and peas was due to a new formation or whether these substances were taken up from the soil. For when peas were grown in asbestos, the young plants contained exactly the same amount of pentosans as the original seeds. I have in my later experiments always planted the seeds in a medium which did not contain pentosans. In most instances I have taken pumice-stone broken to pieces of about i cm. in diameter. I planted about 100 seeds in flat, round dishes of 25 cm. diameter and 7 cm. depth. A hole in the bottom of the dishes allowed the water to flow off. In the experiments in which I have made use of especial solutions for watering, I put the dishes on big funnels and collected the liquid that ran through. This liquid was afterwards made up to the original volume by adding pure water, and was used again for watering. This second solu- tion was somewhat more diluted than the original solution, not only because the plants absorbed a part of the nutritious matter, but also because the pumice-stone retained about 0.5 liter of liquid. Since possibly the same amount of liquid was applied to all the experiments of one group, the solutions used for the different experiments were diluted in about the same proportion and at the same time. The reaction of the solutions was always tested before watering. In all cases, however, it was found to be neutral. I watered every morning and evening. In a few experiments I made water-cultures. I used for these larger flat dishes of 30 cm. diameter and holding about 4.5 liters. Wire-nets with large meshes and thinly covered with paraffin were laid over these dishes. The seeds were put between layers of filter-paper which had been made wet with the solutions to be used. The first roots Pentosans in Plants. 593 were allowed to develop to a length of about 2 cm. The seeds were then fastened in the meshes of the nets so that the roots reached the solution in the dishes but that the seeds proper could not touch it. I have compared either the seeds with the young plants, or the young plants of two or more experiments made side by side in the same darkened room, with each other. In the latter case all the circumstances but one, the influence of which I wanted to find out, were the same. Experiments where the plants had not developed normally have not been reported. I have made in most instances duplicate analyses. Where only a single analysis has been made, I daresay that the variation from the truth will not have been more than 0.3 per cent, of pentosan ; for the largest variation of two duplicate analyses has been 0.5 per cent. My conclusions, however, are based on larger differences than 0.3 per cent, of pentosan, and in far the majority of cases on much larger differences. I first wished to find out whether the increase of pentosans in the young corn plants could have been due to an absorption out of the soil. Table I. 250 seeds of corn were laid between layers of wet filtering paper and, after the first root appeared, were transferred into ground glass. They were watered with well-water, of which i liter did not contain a trace of pentosans. Since the amount of nutritious matter, especially of potash, in the ash did not seem quite suffi- cient for the development of the plants, I watered them on the I2th of December with a solution which contained per liter 2 grams potassium chloride and i gram secondary sodium phos- phate. This application benefited the plants materially. The experiment lasted 18 days (from November 28 to December 16, 1893). 250 plants were harvested. The seeds contained before the germination 4.87 (4.92 ; 4.77 ; 4.91) per cent, of pentosan. The seeds contained after the germination 7.93 (8.01 ; 7.85) per cent, of pentosan. The stems -f- roots contained 12.51 (12.49; 12.53) per cent, of pentosan. 594 ^^ Chalmot. Before germination. After germination. Gain or loss. Gram. Gram. Gram. I seed r= 0.3483 dry matter, i seed =0.1998 dry matter. I stem -|- root -zz. 0.0925 " I plant =10.2923 " — 0.0560 I seed r:: 0.0170 pentosan. i seed ^ 0.0158 pentosan. — 0.0012 I stem 4- root iro. 01 16 " I plant =: 0.0274 " -|-o.oio4 Pentosans are therefore formed in corn plants which grow in darkness, and the increase of pentosans in young plants is not due to an absorption of these substances from the soil. This cannot be wondered at, since I have now proved that pen- tosans are not formed by the assimilation-process, as I have been inclined to believe heretofore. Cross, Bevan and Beadle' com- municated in a recent article that they have found that by the germination of barley in darkness the amount of pentosans also increases. When we compare the results in Table I with my former results with corn grown in garden earth, we perceive that the increase in pentosans is much larger, and that the amount in the seeds proper has diminished much less (0.0012 gram against 0.0039 gram) in the experiment in ground glass. The latter fact is probably due to the plants grown in garden earth being able to take up nitrogen compounds, as we shall see later on. That the increase of pentosans in the young plants was so much larger than I had found before, I brought into connection with the fact that these plants were further developed. For the loss of dry matter was much larger (0.0560 gram against 0.0278 gram). The experi- ments reported in Table II confirm this belief; they show that the young corn plants are subject to the same natural law which I found last summer, viz. that the amount and the percentage of pentosans increase during the life of the organs of the plants." Table II. Seeds of corn were allowed to swell up in a solution which contained per liter i gram potassium chloride and 0.67 gram secondary sodium phosphate. They were planted in pumice- stone and watered with well-water. Two experiments were made : the first lasted 15 days (firom January 3-18, 1894); the other 24 days (from January 18 to February 11). The temperature of the > Ber. d. chem. Ges. 27, 1065. " Probably organs where large amounts of reserve substances (not pentosans) are deposited are exempt from this law. Pentosans in Plants. 595 room in both instances was about the same ; in both experiments ICO seeds were planted. In the first experiment 93 plants were harvested ; in the second, 98. The average seed taken for the first experiment contained 0.3243 gram of dry matter ; for the second, 03238 gram. I have, however, calculated the results below as if in both instances seeds of 0.3243 gram of dry matter had been taken. This facilitates making the comparisons. I have followed the same practice in the later tables. The seeds contained before the germination 6.42 per cent, of pentosan. The plants of experiment i contained 10.88 per cent, of pen- tosan. The plants of experiment 2 contained 13.00 per cent, of pen- tosan. I. Before the germination. Gram. I seed 1=0.3243 dry matter. i plant: I seed zr 0.0208 pentosan. i plant; After the germination. Gram. 10.2586 dry matter. : 0.0281 pentosan. Gain or loss. Gram. — 0.0657 -j- 0.0073 seed: seed; : 0.3243 dry matter. : 0.0208 pentosan. plant 33:0.2377 dry matter, plant 330.0309 pentosan. — 0.0866 -j-O.OIOI The experiments with peas formerly reported on were doubtful, for I found that the amount of pentosans increased when the plants were grown in garden-soil, and that the amount remained the same when they were grown in asbestos. That this difference is due to an absorption of pentosans from the soil becomes very doubtful when we look at the results obtained from corn. I have, moreover, found that the amount of pentosans decreases when peas are grown for a longer time in soils which contain much humus and pentosans. Table III. Peas were grown in wood-soil for 10 days (from June 24 to July 4, 1893); 125 seeds were planted and 98 plants were harvested. The seeds contained 6.44 per cent, of pentosan ; the plants, 8.47 per cent. Before germination. Gram. After germination. Gram. Gain or loss. Gram. I seed =: 0.2681 dry matter. I plant 1=0.1746 dry matter. — 0.093s I seed = 0.0173 pentosan. I plant =0.0148 pentosan. — 0.0025 596 De Chalmot. Table IV. Peas were grown in garden-soil for 10 days (from June 25 to July 5, 1893); 125 seeds were planted and 108 plants were har- vested. The seeds contained 6.62 per cent, of pentosan ; the plants, 8.06 per cent. Before the germination. After the germination. Gain or loss. Gram. Gram. Gram. I seed zz 0.2557 dry matter. 1 plant ■=. 0.1828 dry matter. — 0.0729 I seed =1:0.0169 pentosan. i plant rz 0.0147 pentosan. — 0.0022 To explain these differences I made the following experiments : Table V. Peas were grown in pumice-stone. Three experiments of 100 seeds each were started on May 20, 1894. Experiment I was harvested on May 25, experiment II on May 29, and experiment III on June 5. Of experiment I, 83 plants were harvested ; of II, 86, and of III, 79 plants. The average seed of I contained 0.2465 gram of dry matter ; of II, 0.2488 gram ; and of III, 0.2476 gram. The seeds contained before the germination 6.8 (6.8 ; 6.8) per cent, of pentosan. The plants of experiment I contained 7.8 (7.7 ; 7.9) per cent, of pentosan. The plants of experiment II contained 8.5 (8.4 ; 8.6) per cent, of pentosan. The plants of experiment III contained 8.55 (8.4; 8.7) per cent, of pentosan. I. Before the germination. After the germination. Gain or loss. Gram. Gram. Gram. I seedzr 0,2465 dry matter, i plant =0.2244 dry matter, — 0,0221 I seed =z 0.0168 pentosan, i plant rz 0.0175 pentosan, +0-0007 II. 1 seed zr 0.2465 dry matter, i plant zz 0,2105 dry matter. — 0,0360 I seed =z 0,0168 pentosan, i plant =: 0.0179 pentosan. +0,0011 III. I seed z= 0.2465 dry matter. i plant=zo.i703 dry matter, — 0.0762 I seed =: 0.0168 pentosan, i plant ::= 0.0146 pentosan. — 0.0022 The amount of pentosan in peas grown in darkness therefore first increases and afterwards decreases. Since we know that the Pentosans in Plants. 597 pentosans in the seeds of Trapaolum niajus decrease largely during the germination, it cannot be wondered at that the pento- sans in peas also can be converted into other substances. It becomes then, however, doubtful whether the decrease of pento- sans in the seed proper is at all due to a removal of pentosans. On the other hand the supposition gains ground that the pento- sans in the seeds are converted into other substances, while the pentosans in the young stems and roots are formed from other substances. I have been able to trace only very small amounts of soluble pentosans in germinating seeds,^ which in connection with other facts makes it appear that the pentosans, as such, are only removed in very small quantities. We have observed that the amount of pentosans in peas decreases when they grow for a longer period in darkness, and we found that in corn the pentosans increase under these circumstances. This difference is not only due to the pentosans in the pea seeds being more easily converted into other substances, but also to the increase of pentosans in the young pea stems and roots being much less pronounced than in corn plants. The stems -|- leaves of peas sown on July 25, in the open ground, contained on Aug. 2, 4.8 per cent, of pentosan; on Aug. 7, 4.8 per cent, of pentosan; on Aug. 19, 7.3 per cent of pentosan. Also in other plants of the leguminous tribe the pentosans seem to increase only slowly. The stems -j- leaves of red clover sown on March 31, in the open ground, contained on May i, 5.6 per cent, of pentosan ; on May 30, 7.0 per cent, of pentosan. In these cases the plants had the benefit of the sunlight, but we shall see later on that this makes little difference in the changes of the amount of pentosans. After having considered the changes of the pentosans in repre- sentatives of the GramincB and Legiiminosce, I made one experi- ment with seeds that contained much fat. These seeds are very poor in pentosans, and if our former deductions are correct we should observe a large increase of pentosans when these seeds germinate in darkness. S. Frankfurt^ observed an increase of pentosans in the germinating seeds of the sunflower {Helianthiis annuus')^ where it was in the ratio of 2.75 : 5.86. I obtained similar results with pumpkin (^Cucurbita Pepo). iThis Journal 15, 278. *Landw. Vers. Stat. 43, 176. 598 De Chalmoi. Table VI. 350 seeds of pumpkin were allowed to swell up in water and were transferred into pumice-stone. The germinative power of the seeds was very small. They were planted on July 12; 75 plants were harvested from July 20-28. The shelled seeds contained 1.6 (1.6 ; 1.6) per cent, of pentosan. The plants contained 4.3 (4.2 ; 4.4) per cent of pentosan. Before germination. After germination. Gain or loss. Gram. Gram. Gram. I seed (shelled) 0.1409 dry matter, i plant =20.1361 dry matter. — 0.0048 I seed (shelled) 0.0022 pentosan, i plant == 0.0059 pentosan. + 0-0037 Before closing this chapter I wish to give a list containing the percentage of pentosan found in the dry matter of different seeds, which may be of use in later investigations : Corn, 4.65*;' 4.07 ; 3.95 ; 6.42 ; 4.87. Barley, 10.6 " (furfurol 5.3). Pea, 3.81* ; 4.27* ; 6.8 ; 6.44. Acorn (^Quercus /alcaia, shelled seeds), 1.34*. Nasturtium (^Tropcsolum majus, shelled seeds), 11.97*. Peanut, 2.76*. Pumpkin, 3.45 (shelled 1.6). Hemp,' 11.02. Sunflower,^ shelled seeds, 2.74 (fruits, 12.69). The Influence of Light. Since the absence of light might have some influence on the amount of pentosans in the plants, I made the following experi- ments : Table VII. Peas were grown in pumice-stone from June 28 to July 12. Three experiments were started. In experiment I the plants were grown in darkness; in II and III, in light. The experiments stood in the same room. The light was rather diffuse, so that the plants in experiments II and III, though green, had not 'Figures marked with * have been obtained by the method of slow distillation and are therefore too low. *Cross, Bev.Tn and Beadle: Ber. d. chem. Ges. JJ7, 1064, note. 3 Frankfurt : Landw. Vers. Stat. 43, 143. « Ibid. Pentosajis in Plants. 599 developed normal leaves. It was the intention to allow the plants of experiment III to grow for a longer period, but owing to their abnormal appearance this experiment (III) was discontinued. The plants were watered with a solution which contained per liter 0.5 gram ammonium nitrate, 0.5 gram primary potassium phosphate, 0.25 gram magnesium sulphate, MgS04.7H20, 0.25 gram potassium chloride, i gram calcium carbonate.' The solution was shaken before it was used. Of experiment I, 62, and of II, 82, plants were harvested. The seeds of experiment I contained 0.2628 gram of dry matter; those of II contained 0.2636 gram of dry matter. The seeds before the germination contained 6.45 (6.4 ; 6.5) per cent, of pentosan. The plants of experiment I contained 6.95 (6.8; 7.1) per cent, of pentosan. The plants of experiment II contained 6.65 (6.5; 6.8) per cent. of pentosan. I. Before the germination. After the germination. Gain or loss. Gram. Gram. Gram. I seed 3=0.2628 dry matter. i plant 1=10.1923 dry matter. — 0.0705 I seed = 0.0170 pentosan. i plant zz 0.0134 pentosan. — 0.0036 II. I seed =10.2628 dry matter, i plant 1=0. 1879 dry matter. — 0.0749 I seedz=o.oi7o pentosan. i plant =10.0125 pentosan. — 0.0045 Table VIII. Seeds of corn were grown in pumice-stone from July 18-28. Two experiments of 100 seeds each were started. In experiment I the dish with the plants was put in a box. The plants therefore grew in darkness. The dish with the plants of experiment II was placed on the box. The experiments were placed in the open air under a porch, which shut off the direct rays of the sun. The plants were watered with well-water; on July 25, however, there was applied to the plants of both experiments 0.45 gram potassium chloride and 0.45 gram calcium chloride dissolved in 250 cc. of water. This materially increased their development. The plants of both experiments developed very well ; those of II being quite normal. Of experiment I, 95 plants were harvested; of II, 92 ' This solution was used with a view to experiment III, which, however, failed. 6oo De Chalmot. plants. The seeds taken for experiment I contained 0.3485 gram dry matter, those of II, 0.3479 gram. The seeds contained before the germination 4.4 (4.4 ; 4.4) per cent, of pentosan. After the germination the seeds of experiment I contained 10.45 (io-4; lO-S) P^'' cent, of pentosan. The stems + roots of experiment I contained 12.3 per cent, of pentosan. The seeds of experiment II contained 10.3 rio.2; 10.4) percent, of pentosan. The stems + roots of experiment II contained 12.5 (12.3; 12.7) per cent, of pentosan. I. Before germination. After germination. Gain or loss. Gram. Gram. Gram, I seed zr 0.3485 dry matter, i seed no. 1297 dry matter. I stem + root zr 0.1236 " I plant z=o.2533 " — 0.0952 I seed rz 0,01 53 pentosan, i seed =0.0136 pentosan. — 0,0017 I stem + root =30.0152 " I plant =0.0288 " +0.0135 II. I seedz=o.3485 dry matter, i seed =10.1342 dry matter. I stem + root ir 0.1282 '• I plant =10.2624 " — 0.0861 I seed =.0.0153 pentosan. i seed 3=0,0138 pentosan. — 0.0015 I stem-j-root ^o.oi6c " I plant =0.0298 " +0,0145 The experiments with peas (Table VII) have not succeeded very well, for the plants of experiment II did not receive sufficient light to allow them to develop normally, I should have made this experiment over if I had found time. As it is, I believe that in connection with the experiments with corn (Table VIII) I may draw the general conclusion that the influence of the light during the time that plants can grow in darkness is only slight. The composition of the tissue, so far as pentosans are concerned, is little affected. The Influence of Nitrogen. I have tried to find out whether the amount of nitrogen com- pounds at the disposition of the young plants has any influence Pentosans in Plants. 6oi on the amount of pentosans which are formed, or which are dissolved during the germination of seeds in darkness. I made these investigations with corn seeds, which contain few nitrogen compounds, and which, therefore, are more apt to be influenced by nitrogen in fertilizers than seeds richer in nitrogen compounds. It had been found by former investigations that the amount of ash constituents in corn seeds is not sufficient to enable them to use their reserve substances to the best advantage. I therefore made two parallel experiments, and watered the first with a solution containing all plant-nutritious matters, and the other with a solu- tion containing all but nitrogen. Nitrogen compounds were applied in the most easily available form, viz.^ nitrates. Table IX. Two experiments of loo corn seeds each were made in pumice- stone, April 5-28, 1894. They were watered with the following liquids : Liquid for Experiment I Liquid for Experiment II containiug nitrogen. without nitrogen. Nitrogen-free water, 2.25 liters. 2.25 liters. Sodium nitrate, 2.0 grams. ., Potassium nitrate, 2.0 " ,, Primary potassium phosphate, , ,, KH.PO, ^-S i.S grams. Magnesiumsulphate,MgS04.7H20, i.o " i.o " Sodium chloride, 0.5 " 0.5 " Potassium sulphate, .. " 1.75 " Calcium carbonate, i.o " i.o " Of experiment I, 95 plants were harvested; of 11,92 plants. The seeds for experiment I contained 0.3855 gram of dry matter; those of II, 0.3854 gram. The seeds contained before the germi- nation 3.95 (3.9; 4.0) per cent, of pentosan, 1.70 (1.72; 1.68) per cent, of total nitrogen, and 10.50 (10.63; 10.38) per cent, of albu- minoids. The plants of experiment I contained 6.5 (6.7 ; 6.4 ; 6.4) per cent, pentosan; 2.30 (2.25 ; 2.34) per cent, of total nitrogen, and 8.56 (8.56; 8.56) per cent, of albuminoids. The plants of experiment II contained 8.45 (8.4; 8.5) per cent, of pentosan, 1.98 (2.01 ; 1.94) per cent, of total nitrogen, and 8.60 (8.25; 8.94) per cent, of albuminoids. Vol. XVI.-45- Sain or loss. Gram. — 0.0475 + 0.0068 + 0.0012 — O.OI16 — 0.0580 + 0.0125 — O.OOOI — 0.0123 602 De Chalmot. Before germination. After germination. Gram. Gram. I seed := 0.3855 dry matter. i plant n: 0.3380 dry matter, I seed ::= 0.0152 pentosan. i plant nz 0.0220 pentosan. I seed ^0.0066 total nitrogen, i plant =: 0.0078 total nitrogen. I seed 1^:0.0405 albuminoids, i plant 3:0.0289 albuminoids. II. I seed ^0,3855 dry matter, i plant rz 0,3275 dry matter, I seed =10.0152 pentosan. i plant =r 0.0277 pentosan, I seed 1=0,0066 total nitrogen, i plant =10.0065 total nitrogen, I seed =0.0405 albuminoids. i plant =1:0.0282 albuminoids. We observe that the young plants of experiment I have taken up considerable quantities of nitrates. The percentage of albu- minoid in the plants of both experiments was, however, about the same. The plants to which nitrates had been applied contained considerably less pentosans than the others. But this might be due to another reason than the influence of the nitrogen com- pounds. For these plants probably, owing to the higher concen- tration of the solution, had lost less in dry matter. This indicates that they were not as far developed, and the amount of pentosans in corn increases when the development of the plant advances. Therefore new experiments were started in which I wanted to find out at the same time whether the pentosans will decrease in young corn plants if large amounts of nitrates are applied, and if the plants are allowed to develop as far as possible. In this case sodium sulphate was added to the nitrogen-free solution in order to make up for the smaller concentration caused by the absence of nitrates. Table X. Four experiments with corn in pumice-stone were started on June 9. In experiments I and III the plants received nitrogen in fertilizers, in experiments II and IV they did not. The solutions for watering contained: I and III, II and IV. Nitrogen-free water, 2,25 liters. 2.25 liters Ammonium nitrate, 1,5 grams. Primary potassium phosphate, I.O " 1,0 gram. Magnesium sulphate, MgS04,7H30, 0,5 " 0,5 " Potassium chloride, 0.5 " 0.5 " Sodium sulphate. 1.75 " Calcium caibonate, 2,0 " 2.0 •' Pentosans in Plants. 603 After 5 days there was added to the solution containing nitro- gen I gram of ammonium nitrate, and to the nitrogen-free solu- tion 2.25 grams of sodium sulphate. Experiments I and II were harvested on June 19, experiments III and IV on June 24. The plants of experiment III were dying when harvested ; their roots were dead and so were the under- parts of the stems. The leaves, however, were still quite fresh. The water which had run through the pumice-stone of this exper- iment had dissolved substances from the roots, for it had become slightly colored and had the odor of the crushed plants. Of experiment I 100 plants were harvested, of II, 99 plants, of III, 83 plants, and of IV, 93 plants. The seeds of experiment I contained 0.2789 gram of dry matter; those of II, 0.2802 gram ; those of III, 0.2802 gram ; and those of IV, 0.2823 gram. The seeds contained before the germination 4.07 (4.3 ; 3.8; 4.1) per cent, of pentosan. The plants of experiment I contained 12.5 (12.5 ; 12.5) percent, of pentosan. The plants of experiment II contained 13.45 (i3'3; 1 3-6) per cent of pentosan. The plants of experiment III contained 14.4 per cent, of pento- san. The plants of experiment IV contained 14.75 (i4-7 ! i4-8) per cent, of pentosan. I. Before germination. After germination. Gain or loss. Gram. Gram. Gram. I seedzi: 0.2789 dry matter. i plant z= 0.1862 dry matter. — 0.0927 I seedz=o.oii4 pentosan. i plant n: 0.0233 pentosan. -f-0.0119 II. I seedi30.2789 dry matter. i plant 1=0.1931 dry matter. — 0.0858 I seedzzo.0114 pentosan. i plant =: 0.0260 pentosan. -f- 0.0146 III. 1 seedzr 0.2789 dry matter. i plant z=o.i6o8 dry matter. — 0.1181 I seedrzo.0114 pentosan, i plantrr 0.0232 pentosan. +0.0118 IV. I seed=:o.2789 dry matter. i plant = o.i9i4 dry matter. — 0.0875 I seed =: 0.0114 pentosan. i plant rz 0.0282 pentosan. +0.0168 The loss in dry matter of the plants of experiment I was larger than that of the plants of II. The former were practically 6o4 ^^ Chalmot. exhausted. Still they contained less pentosan than the plants of experiment II. The plants of experiment III were partly dead. They did not contain more pentosan than those of I, for they seemed not to have developed any further. They contained, however, much less dry matter, which probably had been washed out of the dead parts. The plants of experiment IV were farther developed than those of II, but not even quite as far as those of I. They contained more pentosan than those of II, and considerably more than those of I and III. These results fairly prove that the nitrogen compounds tend to decrease the formation of pentosans, or to facilitate their decomposition. The amount of pentosans, however, does not decrease in corn plants even when nitrogen compounds are liberally applied. They even do not decrease after parts of the plants have died off and decomposition has set in. This latter fact cannot be wondered at, for we know that the pentosans of dead corn leaves are also much less readily decomposed and washed out of the leaves than other vegetable matters. It was further of interest to find out whether the nitrogen com- pounds only decrease the formation of pentosans in the stems and roots or whether they also facilitate the decomposition of the pen- tosans of the seeds. The experiments reported in Tables XI and XII prove that they do both. Table XL Two experiments of ICO corn seeds each were started on July 31. The seeds were put between layers of wet filtering paper, and after they had developed the first root, were suspended in wire-netting over flat dishes so that the roots entered into the following solu- tions: « Solution containing n Experiment I trogen. Nitrogen-free solution Experiment 11. Nitrogen-free water, 4.0 liters. 4.0 liters. Ammonium nitrate, 0.5 gram. Potassium nitrate, 1.5 " Primary potassium phosphate, i.o " 1.0 gram. Potassium chloride, 0.25 " 1-35 " Magnesium sulphate, 0.5 " 0.5 " Calcium carbonate, i.o " 1.0 " Sodium sulphate, 0.9 " On August 4 there was added 0.5 gram of ammonium nitrate to solution I. The volume of the solution was kept stationary by Pentosans in Plants. 605 adding nitrogen-free water. Every da};- 2 liters of air free of nitro- gen compounds were driven through each of the solutions. Of experiment I 93 plants were harvested on August 9 (9 days). Experiment II was harvested 12 hours later, on the morning of August 10 ; it yielded 94 plants. The seeds of experiment I contained 0.3499 gram of dry matter; those of 11,0.3491 gram. The seeds before the germination contained 4.4 (4.4 ; 4.4) per cent, of pentosan. The seeds of experiment I contained after the germination 6.0 (6.1 ; 5.9) per cent., and the stems + roots 9.8 per cent. The seeds of experiment II contained after the germina- tion 5.65 '(5.6; 5.7) per cent., and the plants 11.8 per cent, of pentosan. Before germination. After germination. Gain or loss* Gram. Gram. Gram. I seed=:o.3499 dry matter, i seed =10.2222 dry matter. I stem -J- root zz 0.0762 " I plant =10.2984 " — 0.0515 I seed z= 0.0154 pentosan. i seed z= 0.0133 pentosan. — 0.0021 I stem -|- root z= 0.0075 " I plant =: 0.0208 " -(*0'0054 II. I seed = 6.3499 dry matter, i seed =0.2515 dry matter. I stem -j- root := 0.0607 " I plant =0.3122 " — 0.0377 I seed=io.oi54 pentosan. i seed zzo.0142 pentosan. — 0.0012 I stem-]- root rz 0.0072 " I plant zz. 0.0214 " +0.0060 Table XIL Two experiments of no corn seeds each were started on August 2. The plants, which grew in pumice-stone, were watered with : Nitrogen-free water, Solution containing nitrogen. Experiment I. 2.25 liters. Nitrogen-free solution. Experiment 11. 2.25 liters. Ammonium nitrate, 1.5 gram. Primary potassium phospl Potassium chloride, late, 1.0 •♦ 0.5 " 1.0 gram, o.s " Magnesium sulphate, Calcium carbonate, 0.5 » 0.5 '• 0.5 " 0.5 " Sodium sulphate, 1.75 " On August 8, I gram of ammonium nitrate was added to solu- tion I. Of experiment I, 102 plants were harvested on August 13 6o6 De Chalmot. (ii days); of II, 91 plants were harvested on the morning of August 15 (36 hours later). The seeds for experiment I contained 0.3328 gram of dry matter; those of II, 0.3315 gram. The seeds contained before the germination 4.4 (4.4 ; 4.4) per cent, of pentosan. The seeds of I contained after the germination 10.15 (iO'4 ', 9-9) per cent, of pentosan. The stems + roots of I contained 10.55 (io-5; 10.6) per cent, of pentosan. The seeds of II contained 13.8 (13.7; 13.9) per cent, of pentosan. The stems + roots of II contained 13.6 (13.5 ; 13.7) per cent, of pentosan. The stem + roots of I contained 15.92 per cent, of albuminoids. The stem + roots of II contained 1 1.69 (i 1.63; 1 1.75) per cent. of albuminoids. I. Before germination. After germination. Gain or loss. Gram. Gram. Gram. I seed =10.3328 dry matter, i seed zr 0.0939 dry matter. I stem -|- root — 0.1440 " I plant =10.2379 " — 0.0949 I seed 1=0.0146 pentosan, i seed zro.oo9S pentosan. — 0.0051 I stem-}- root =10.0152 " I plant =10.0247 " -(-o.oioi I stem -f- root =: 0.0229 albuminoids. II. I seed 1=0.3328 dry matter, i seed z= 0.0959 dry matter. I stem + root zr O.I 350 " I plant z= 0.2309 " — 0. 1019 I seed =10.0146 pentosan. i seed 3=0.0132 pentosan. — 0.0014 I stem -j- root =30.0184 " I plant =10.0316 " -f-0.0170 I stem + root z=. 0.0158 albuminoids. The plants of experiment I (Table XI) have been much further developed than those of experiment II ; hence the results are less obvious. The results reported in Table XII are very convincing. Though the plants of experiment I were a little less developed than those of II, this difference is only small, while the difference in the amount of pentosans, both in the stems -|- roots, and in the seeds, is very pronounced. The stems + roots of experiment I (Table XII) also contain considerably more albuminous matters than those of experiment II. Pentosans in Plants. 607 Resum^. 1. Pentosans are present in the stems which are developed from barley seeds that germinate in darkness. There is no reason to assume in these stems other substances which yield furfurol or so-called oxycelluloses. Far the larger part of the furfurol-yield- ing compounds in these stems is, moreover, soluble in strongly diluted acids and alkalies, and therefore does not belong to the celluloses proper. 2. The amount of pentosans increases in corn which is allowed to germinate in darkness, and this increase is not due to an absorption of pentosans from the soil, but to a formation of pentosans from other substances. For such an increase of pentosans takes place when the seeds are planted in pumice-stone which does not contain pentosans. 3. The more the development of the young plants is advanced, the more pentosans do they contain. 4. When peas are allowed to germinate in darkness, the amount of pentosans first increases and then decreases. 5. The facts lead to the conclusion that the pentosans in the seeds are decomposed, and those in the stems and roots are newly formed, but that the transportation of pentosans as such is only of subordinate importance. 6. Seeds which contain much fat contain only a small amount of pentosans, but this increases materially during the germination in darkness. 7. The absence of light has only little, if any, influence on the conduct of the pentosans during the short time that seeds can develop in darkness. 8. When besides all other nutritious matters, nitrates are applied to seeds which are germinating in darkness, the pentosans of the seeds are more readily decomposed, and less pentosan is formed in the young stems -f- roots. The latter, on the other hand, contain more albuminous matter. II. — The Formation of Pentosans in Plants. In my paper on " Soluble pentoses in plants '" I put forward the following general proposition : 1 This Journal 15, 21. 6o8 , De Chalmot. " Pentoses are either formed along with hexoses by the assimi- lation process, or the plants have the power to form pentoses from hexoses." In my subsequent investigations I have shown that pentoses are not formed by the assimilation process.' Hence pentoses are elaborated from hexoses. In the first part of this paper it is, more- over, shown that the pentoses increase to a very great extent in corn which is allowed to germinate in darkness. I shall now make an attempt to solve the question, in what manner the pen- toses are formed from hexoses, and shall therefore make use of the chemical and physiological facts which have come to my knowledge. I am sorry to acknowledge, however, that my solu- tion of this problem must still be regarded as an hypothesis. By hydrolysis of the pentosans two pentoses have been found to result, viz., /-xylose and /-arabinose. Xylans (pentosans which yield xylose) are in almost every instance intermixed with, or chemically united with glucosans. Wood-gum, for example, is always found beside glucose-cellulose. In several celluloses, xylose molecules are so strongly united with glucose molecules that they cannot be separated without breaking up the cellulose.^ Arabans (pentosans which yield arabinose) are in most instances found beside galactans — in fact so frequently that arabinose and galactose were formerly often confounded. They are beside each other in arabinic acid,' in the gums that are exuded from the barks of the cherry* and the peach trees,^ in gum tragacanth," in coffee-beans,' and in the paragalactan which is found in seeds of many leguminoses.* Galactose molecules were never found in celluloses,' and neither were arabinose molecules. It is true that in some exceptional cases xylose was found beside little or no glucose, and arabinose without galactose. Thus Stone'" found a gum in the fruits of Gymnocladus Canadensis « J. Am. Chem. Soc. 16, 6i8. 'E. Schuize: Ber. d. chem. Ges. a*, 2277; Winterstein : Ztschr. physiol. Chem. 17, 382; ToUens and Schuize: Ann. Chem. (Liebig) 371, 55. ' ToUens : Kohlehydrate, 214. < Loc. cit. » Stone: This Journal 12, 435. • ToUens : Kohlehydrate, 217. ' E. Elwell : This Journal 14:, 473. "E. Schuize: L^ndw. Vers. Stat. 41,223; Ztschr. physiol. Chem. 14, 227. » Ber. d. chem. Ges. 24, 2281. 10 This Journal 15, 660. Pentosans in Plants. 609 which is probabl)' composed of glucose and arabinose molecules. Winterstein' found in amyloid galactose, xylose and but little glucose. E. Schulze and Steiger''' and C. Schulze and Tollens' found in bran of wheat and rye a little arabinose along with xylose and glucose, but they did not find galactose. If we compare the formulae proposed by E. Fischer^ for a?-glu- cose and /-xylose, and those for d^-galactose^ and /-arabinose, it must strike us that a great similarity exists : H H OH H CH2(0H).C . C . C . C .C.OH, ^/-glucose; OH OH H OH H OH H CH Ber. d. chem. Ges. 84, 524. ^xhis Journal 15, 21. ' Ibid. 15, 276. 4 Ber. d. chem. Ges. 25, 2569. <> Ibid. 27, 1063. « E. Schulze : Ibid. 34, 2283. On Chemical Equilibria as Temperature- Functions. 6ii both show these reactions. I shall endeavor to obtain additional support for my hypothesis, but I am aware that it alone can never entirely explain the formation of pentosans in plants, for this formation takes place only in living cells.' III. — Pentosans in Pine Wood. Some time ago I found that the unfinished oak wood contains less pentosans than the wood of which the formation is accom- plished. According to Wieler,* unfinished wood oi Pinus sylvestris contains 14-15 per cent, of wood-gum. Since I found no more than 10 per cent, of pentosan in pine wood, it seems as if either the pentosans decrease instead of increasing by the formation of this wood, or that the wood-gum of Wieler does contain other substances than pentosans. I have not found time to investigate this subject more fully, but I proved that in pine wood also the amount of pentosans increases during its formation. A pine, Pinus taeda, of about 40 years was cut down in the commence- ment of June and was investigated immediately afterwards. The wood contained, unfinished young wood, 7.65 (7.5; 7.8) per cent, of pentosan; mature wood of 10 and more years old, 9.85 (9.8 ; 9.9) per cent, of pentosan. Johns Hopkins University. ON CHEMICAL EQUILIBRIA AS TEMPERATURE- FUNCTIONS. By J. E. Trevor and F. L. Kortright. The consideration that the only things of which we can take direct cognizance are manifestations of energy, and that the total energy of an isolated system remains constant in amount or, in other words, subsists, leads at once to the idea of the substantiality' of energy and to the conception that all natural processes are in their ultimate nature energy transformations. This makes the 1 Sec my investigation of the wood-formation : This Journal 16, 218. " Landw. Vers. Stat. 88, 307. 6i2 Trevor and Kortrighi. equilibrium of any material system subject to the condition that a possible variation of any one of the forms of energy which it includes must be exactly compensated by the corresponding variation of the forms correlated with this one; the criterion of equilibrium being that the algebraic sum of the energy changes involved in a virtual displacement of the equilibrium must vanish. Denoting any energy form by E, this condition is expressed analytically by I^E=o, (i) the summation being taken over all the correlated energies of the system. The proposition in this general form is known as the Principle of Virtual Energies; its first clear and comprehensive statement is due to Ostwald.' The resolution of an energy-form into its intensity-factor /, the quantity which measures its tendency for transference, and the capacity-factor C, the quotient of the energy by its intensity, shows that in general the above equation is to be employed in the more special form, ICdI—o, (2) to represent the equilibrium of the virtual energy-changes about a displaceable equilibrium. When this criterion is applied to the equilibrium between heat Q and work W, it furnishes the fundamental differential equation for the equilibria of thermodynamics, in the form, IdW=.-^,-dT, (3) Z" representing the absolute or energetic temperature of the heat quantity Q. This equation contains the theory of all displace- able chemical equilibria, because such equilibria always involve heat and some one or more entirely convertible forms of energy ; for isothermal equilibria it becomes, of course, IdW-=o, (4) the Principle of Virtual Work. In all customarily considered types of cases the chemical equi- libria which appear involve thermal and volume energies, and the influence of temperature-changes upon them must be regarded as displacements due to the addition of heat of an increased intensity ' Ztschr. phys. Chem. 10, 363 (1892). On Chemical Equilibria as Temperature-Functions. 613 (temperature), or the reverse. There are two cases to be con- sidered, the Complete and the Incomplete Equilibria of the Phase Rule. I. Complete Equilibria. — When n substances compose a chem- ical system of w + i heterogeneous phases, one of which is vapor, the equilibrium is a function of but one independent variable, and its equilibrium-equation, the differential equation of the " Bound- ary curve," is ^-^.dT=vdp, (5) RTd\o^,p — -Pj^.dT=o. Here p is the equilibrium-pressure at the temperature T, p is the total molecular heat of change, i. e. the heat Q absorbed at con- stant volume plus RT the volume energy (normally) obtained, and R is the constant of the gas-equation /z^ = ^7". To get rid of the volume energy term we subtract — -~ — ~~~^ from both members, and there results RTd\ogeK--9^.dT=o, ' (5a) log, A' being the characteristic constant of the reaction-isotherm. For, ^=r = c, the molecular concentration of the saturated vapor, and is therefore proportional to K. The relation (50;) is the famous law of van't Hoff' for the displacement of a movable chemical equilibrium. II. Incomplete Equilibria. — When n independent components make up a system of less than n-\- 1 heterogeneous phases the isothermal equilibrium of the system depends, to a greater or less degree, upon the relative concentrations of its components, and the balance existing between the respective volume-energies of the latter will be disturbed by a possible addition ot heat. The sum of the virtual changes then becomes, for a final equilibrium, ( Ivdp — 2' VdP) — -^ .dT=o, (6) P, VfPand V representing the partial pressures and molecular volumes of the original reacting substances and of those produced •HandlingarderStockholmer Akad. ai.No. 17, p. 18(1886). 6 14 Trevor and Kortright. by the reaction, respectively. Qt is, as before, the molecular heat of the reaction considered as occurring at constant volume and at the equilibrium -temperature 7.' It has been shown elsewhere' that the bracketed member of this equation becomes RT^Xogef^ on the elimination of the pressures, so we have RTd\ogeK—Mf.dT=io, (6a) identical with (5a) and representing, therefore, the effect of changes of temperature in displacing all types of chemical equi- librium under discussion. This remarkable equation is usually written in the form a log. K _ Qr . dT ~~ RT' ' (7) by writing the differential coefficient as -j^ . ~. „ , i. e. the per- K c J centage rate of change of the equilibrium-constant per degree, its meaning is at once seen to be that the temperature-coefficient of reaction depends upon the heat of reaction and is a function of the absolute temperature. The qualitative result, that a rise of temperature will displace a chemical equilibrium in the direction in which the reaction absorbs heat, is familiar to every one. To find, from the rate ~^-^. -^ with which an equilibrium R 1 1 changes with the temperature, the amount of change correspond- ing to a finite interval of temperature, the expression (7) must be integrated between the temperature limits; this is not a simple matter, for the reason that Q depends upon the specific heats of the reacting substances and the specific heats are unknown func- tions of the temperature. In the customary applications which are made to illustrate van't Hoif 's law, the heat of reaction is Xvdp — 2 VdP— -^ . dT= o , • In the case of the virtual displacement in a rigid shell, Qt T the equilibrium is between the volume-energies on each side, and their displacements mutu- ally cancel — those in one sense are exactly balanced by the corresponding ones in the Other. In the case of a vaporization, or the like, vdp--^^.dT^o, the change of volume-energy is, on the contrary, not counterbalanced, and p is therefore the heat of reaction under constant pressure, /. e., py = Qt-\- R T. " Trevor: J. Am. Chem. Soc. 15, 14 (1893). On Chemical Equilibria as Temperaiure-Fuyictions. 615 assumed to be constant over a very small range of temperatures;' but this procedure greatly restricts such applications and in general leaves much to be desired, so that in presenting the matter in the course of the regular lectures upon Physical Chem- istry given in Cornell University a more rigid integration has been employed, and with success. The heat of reaction Qt absorbed at the absolute temperature 7", being the energy-difference of the reacting systems, is that, Q^, at some arbitrarily selected standard temperature T^ increased for each degree by the excess of the total molecular heat of the system II which is produced by the reaction over that of the original system I. In symbols ^^-|f-)(7^-7;). (8) L e., Q^-=zQ,\ (c„ — cO( T— To) . (8a) With this value of Qt, van't Hofi's equation becomes ig>o + (- |;aIog.y.= ^^^:|-^|V-^a7^+^jV-^a7^. (i^ making the permissible assumption that the difference Ac of the specific heats is constant between wide temperature-limits. The integration gives or, for the heat of reaction, To illustrate the application of this relation let us consider the simplest case conceivable, the vaporization-equilibrium of water, and calculate the heat of vaporization from the vapor-pressure curve. For the equilibrium Liquid ^ Vapor, the reaction-isotherm is K.C=:c, 1 See, for example, van't Hoff : Etudes, 128-139; ^nd Nernst : Theor. Chem. 514-521. 6i6 Trevor a7id Kortright. Cand c representing the respective molecular concentrations of the two phases, and hence K=:. -^ .€■=. const X -^ from the relation />:=ci?7' obtaining for the vapor. We have therefore A The formula (iia) being applicable to any temperature-interval, we may select widely separated temperatures, say 50° and 100° C, and thus introduce large and accurately known vapor-pressures. The most accurate available data for the vapor-pressures and the specific heats at constant volume are : ^1 = 323° ^.= 373° pi-=gi.gS mm. Hg /a = 760* Ci=i8 X i.02fa/'' Cii = 18X0,47* Ac'=. — 9.90 cal. These values substituted in (iia) give j^o= 10201 rai/as the heat of vaporization of water at 0° C, reduced to constant volume. Addition of the volume-energy of the mol^ of expanding vapor, pv=z2 Tz=. 2 X 273 = 546 cal, gives />o = 10201 -\- 546 = 10747 ^«^. while the two experimental determinations which have been made in recent years are (when expressed in the unit employed above, the zero-degree calorie) Dietericl^ 10742 cal Hartog and Harker*^ 10722 " . 1 Regnault. *The mean of the best determinations near 75° C, i- e. of the nine values cited in Laiidolt and Bornstein's "Tabellen," 133-134, with the exception of the first two, which are obviously faulty. ' Given by Gray as 0.3787 at constant pressure at ioo°-i25°, which is lower than that found by Regnault for higher temperatures, so for 5o°-ioo° we round off the value to 0.37, and con- vert to the condition of consiant volume by the accepted ratio A = 1.274 at 78°. * For the employment of this term to express the gram-molecular weight, see Ostwald : Ztschr. phys. Chem. 13, 375 (1894). "Ann. der Phys. Wied. 37, 494 (1889). •Abstracted in Beib!. zu den Ann. der Phys. Wied. 18, 902 (1894). On Chemical Equilibria as Temperature- Functions. 617 The theoretical result is thus seen to differ from the average of the best experimental ones by only fifteen gram -calories, or less than two-tenths of one per cent. (0.14 per cent.), an agreement which very satisfactorily speaks for the correctness of the theory. No such result can possibly be reached on the assumption of a heat of change which is independent of the temperature. One other curious relation may be worthy of mention in this connection ; since the displacement of an equilibrium is determined by the temperature-change, the heat of reaction and the difference of the molecular heats of the reacting systems, it follows that this difference of the molecular heats is a function of the other quanti- ties named. Equation (11), in fact, when solved for Jc gives Ac= f (12) T,T,\oge-4^-— T,{T, — T,) -'I which permits a calculation of the differenceof the total molecular heats (considered as independent of the temperature) from a temperature change, the corresponding displacement of the equi- librium, and the heat of the reaction. The relation is entirely general. It may, further, be worth while to emphasize that in an entirely strict theory, the dependence of the molecular heats upon the tem- perature may not be omitted from consideration. When this is considered, the approximate form becomes the exact one in which -HM^z=- J is the increment of the characteristic constant of the logarithmized reaction-isotherm over the temperature- interval 7^2 — ^1 • In order to be able to eflfect the indicated integration, the heat of the reaction — the Heat of Dissociation of carbon dioxide — must first be determined as a function of the temperature ; we shall then be in position, from any known state of the equilibrium to calculate those at all accessible temperatures and pressures. The heat of dissociation of one mol of carbon dioxide at 18° C, the negative value of the heat of combustion of carbon monoxide, is, according to the best determinations, 68170 cal^ at constant pressure, or, at constant volume, by subtracting the energy Rl due to the change of volume and doubling the result for the two mols of carbon dioxide concerned, ^291 = 2.(68170—2 X 291); ' Ztschr. phys. Chem. 2, 782 (i888). = Berthelot : Ann. chim. phys. (5) 13, 11 (1878). On Chemical Equilibria as Temperature- Functions. 619 in general, therefore, Q at the absolute temperature T is given by the expression Q^— 2.(68170—2 X 291) + Ac{ r— 291), (14) in which Ac is the difference of the total molecular heats.' This difference, for the reaction 2CO + O, =2C02 is Ac-=. 2Cco -{-Co^ — 2CCO3 or Jc = 2 X 6,79 + 6.96 — 2.(6.8 -f- 0.0072 T) at constant pressure f at constant volume it becomes Ac= 2 X 6.79 I 6.96 _ 2.(6.8 +0.00727^) 1.4 "^ 1.41 1.3 adopting the ratios of the specific heats as given in the new edition of Landolt and Bornstein's tables, p. 340. It is to be noted that the specific heats of carbon monoxide and of oxygen are independent of the temperature within wide limits. Con- densing the above expression, Jc = 4.i7 — 0.011077 T. (15) When this value is substituted in the formula (14) for i^2.,and reduced, we obtain Qt= 133963 + 7-393407 7^— o.oi 1077 T' , (16) which is the function Qt=z > In the complete dissociation under constant pressure the volume is increased in the ratio 3 : 2, the work pv = iT, being therefore done per two mols of the dioxide. With one mol but half this work is to be subtracted, so that the expression preceding (14) should read Qm=^ (68170-291). The oversight in the text involves an unimportant error in Q of less than 0.5 per cent, at 18° C. * For CO, E. Wiedemann, from Landolt and Bornstein, p. 339 : From 23°-99° C, o.24«; cal/gram, at constant pressure ; " 26='-i9S° C, 0.2426 " " " " For Oq, Regnault : loc. cit. : From i3°-207° C, 0.2175 cal/gram, at constant pressure. For COj, E. Wiedemann and H. Le Chatelier : Zeitschr. phys. Chem. 2, 783 (1888). 620 Trevor and Koriright, whereby it is overlooked that the two oxides participate in the reaction in bimolecular quantities. This oversight invahdates Le Chatelier's results. Having now in (i6) the heat of dissociation as a function of the temperature, the complete expression (13) for /J log A' can be written log.A;-log.A; = Jg|^^(i3^3 + 7,39|407_^,o„^^^^ = 29122.3 (y-— ^) + 3-696703 logioy^ +0.002408(7;— 7;). (17) This expression gives the characteristic equilibrium-constant for any assigned temperature, when some one A^ for a corresponding 7; is known. For this latter we may suitably select the deter- mination of Deville, cited by Le Chatelier,' that the percentage dissociation a is forty per cent, at 3000° C. under atmospheric pressure. The formula for K^ i. e. the reaction-isotherm, for the dissocia- tion reaction 2CO2 t- 2CO + O2, is K.C=:cAc, for the respective molecular concentrations ; this may be written or ^=-7 ^2— ' in which v is the volume of the system, but since for the total pressure J?v •=.{2-\- a) RT, and R = 0.0819 liter X atmospheres, we obtain ^~ (I— a)^ (2 + a) 0.08197^ ' ^^^^ Inserting the values above given, = 0.40, T-=. 3273, J*=: i.oo, we have as starting-point for the calculation of equilibria, logio K^ = 4.441436 ; with this value in (17) we get our fundamental function in the desired form representing log K^ as a function of the temperature alone. The result of the calculation for a series of temperatures is as follows : J Ztschr. phys. Chem. 3, 782 (1888). On Chemical Equilibria as Temperature- Functions. 621 Temperature. /^^,oA-. Temperature. log.^fT. 1000° c. 15.762175 3000-C. 4.441436 1300 11-742573 3300 4.60691 1 1500 9-541564 3500 4.644804 2000 S-349543 3555 4.646496 (max.) 2500 5-774939 4000 4-543778 This change of the equilibrium-ratio jSfwith the temperature is more readily seen from the accompanying curve, Fig. i. 1000 1500 moo 80003500 IfiOO Fig. I. Log iqK iog,„A=^(r). A pronounced maximum of the function near 3500° C. is shown by both the table and the plat, its appearance is due to the fact that at this temperature the value of Q, the heat of dissociation, passes through zero and changes its sign. This is clearly shown by the calculation of Q in its dependence upon T, from equation (16), whereby the following series of values is obtained : Successive Values of the Heat of Dissociation. T. )iQ. T. }iQ. o°C. + 67578 cal. 2000° C. + 46767 cat. 18 67588 2500 34640 500 66508 3000 19742 1000 62662 3555 ± 1500 56125 4000 — 18343 622 Trevor and Kortright. The relation of the heat of reaction to the temperature is seen more readily from the curve of (^ = ^^(7') here appended, see Fig. 2. Since Q vanishes at 3555° C, the value of the differential coefficient ^ ^' — also vanishes, and consequently both log K and the degree of dissociation at any given pressure become maxima simultaneously. From the above A'-values the percentage dissociation a for any given total pressure at any assigned temperature is directly calculable by employment of (18), which must for the purpose be solved for a. We thus obtain, setting for convenience the symbol © for the quantity RTK © = ■ RTK—T a = [© + V®'' + (5^]« + [© — x/©^ + ©']^. (19) The final values, obtained by the use of this formula, for most of the above temperatures and a series of total pressures are collected in the following table : 071 Chemical Equilibria as Temperature- Functions. 623 Dissociations, in Entire Percentages. Temp. o.ooi atm. o.oi o.i i.o lo.o loo. 1000° C. O.I I 0.05 0.024 O.OI I 0.005 0.0024 1500 9.5 4.6 2.2 1.0 0.5 0.2 2000 57.7 34.7 18.3 9.0 4.3 2.0 2500 87.0 69.6 46.0 25.7 13.0 6.3 3000 93.9 83.4 62.7 (40.0) 21.6 10.8 3500 95.4 87.0 69.7 46.1 25.7 13.0 3555 954 87.1 69.8 46.2 25.8 13.1 4000 95.1 86.4 68.6 45.0 24.9 12.6 The maximum values at 3555° C, as shown by the table, are very marked. The possibility of their existence was recognized by Le Chatelier, although his error in calculating Q=.

o.2, Deville. 2C + etc., where p +pi +p2 + etc. = i and represent the percentage of the several components of the mixture. Since F is a factor of both members of the equation, it may be written in the form (M-i) _ p(m-i) p,Cm,-i) p,(^2~i') . ^,^ ,x MB md ^ m,d, ^ fn,d, "^ ^ ' ^^J I. — Table showing the specific refraction of some substances when the density is changed by changing the temperf^tui e, for the formula MD t «j '8 ^ e •t.- 3 " u-a 1 Substance. 1 o^P* '^l - MD From the data of 1 "^ CO III 0° Water. Liquid. 0.99987 1.334" 0.2504 Ketteler. 20 " " 0.99824 1-33327 0.2503 " 40 " " 0.99233 1.33093 0.2504 " 60 " " 0.98331 1-32753 0.2509 " 80 " " 0.97 191 1.32330 0.2513 " 100 " " 0.95863 I. 31843 0.2519 " 100 " Vapor. 0.C00S06 I.OC025 0.3101 Lorenz. Ethyl alcohol. Liquid. 0.8056 1.36962 0.3339 Ketteler. 10 " " " 0.7972 1.36565 0-3347 " 20 " " t' 0.7887 1.36164 0.335s " 100 " " Vapor. 0.00206 1.000873 0.4234 Lorenz. 10 Chloroform, Liquid. i.i;i6 1.4490 0.2044 Gladstone. 20 " " 1.4898 1.4467 0.2072 Haagen. 30 " " 1.479 1.4397 0.2064 Gladstone. 100 " Vapor. 0.00535 1. 00144 0.2691 Lorenz. 10 Carbon disulphide. Liquid. 1-2793 1.6344 0.3034 Gladstone. 24.5 " " " '•2593 1.6235 0.3049 " 30 " " " 1.2494 1. 62 1 3 0.3067 " 100 " " Vapor. 0.0034 1. 00148 0-4347 Lorenz. 2 Benzene. Liquid. 0.8979 1.5122 0.3772 Gladstone. 10 " " 0.8868 1.5029 0.3772 " 21. S " " 0.8773 1.4979 0.377S " 628 Edwards. It will be observed that there is quite generally an increase in jM-i) MD as the temperature rises, and that there is a marked in- crease for the vapor in every case. This increase with rise of temperature and change of physical state will be the subject matter of a future paper which is already outlined. II. — Table comparing the two members of the equation j^jt) i>{vi—\) A(^i— i) c J- 1 .• = ^ciA — i 4. ^^^ ' i- for some ordmary aqueous solutions. md ntidi ^ ^ J-^, j2_ " .; Per cent, of substance dissolved. Specific Index of re- fraction for (M-i) 1 1 ^T «"eS From the gravity (£>). sodium light MD. 1 s" m data of Glyce . T5° 1.000 15° 00 1.3330 0.2498 0.2498 00 Skalweit. " 10 1.0240 1.3452 0.2506 0.2502 04 " " 20 1.0490 I.3581 0.2513 0.2506 07 " " 30 1.0750 I-37IS 0.2520 0.2510 10 " " 40 1. 1020 1.3854 0.2524 0.2513 11 " " 50 1. 1290 1.3996 0.2529 0.2518 II « •' 60 1. 1 570 1.4144 0.2533 0.2521 12 " " 70 1.1855 1.4295 0.2534 0.2525 09 " " 80 1. 2125 1.4444 0.2537 0.2529 08 " " 90 1.2396 1.4595 0.2540 0.2533 07 « " 100 1.2650 1.4742 0.2537 0.2537 00 " Acetic £ 4 0.9982 20 00.00 1.33327 0.2492 0.2492 00 Ketteler. 21.5 4.65 1.0065 1-33554 0.2496 0.2493 03 Edwards. 9.26 1. 0130 1.33934 0.2501 0.2501 00 " " 13-97 1.0203 1.34221 0.2499 0.2502 03 " 20 20.72 1.0296 1.34637 0.2499 0.2508 09 Damien. 21.5 29.45 I. 0418 1.35222 0.2500 0.2517 17 Edwards. 20 41.49 1.0532 1.35878 0.2524 0.2524 00 Damien. •< S1.81 I. 0601 1.36374 0.2516 0.2532 16 " " 62,50 1.0640 1.36835 0.2529 0.2541 12 " " 68.97 1.0672 1.37120 0.2537 0.2546 09 « " 76.92 1.0690 1.37443 0.2549 0.2552 03 " " 89.96 1.0673 1.37624 0.2561 0.2561 00 " 21.5 94.33 1.0664 1.37442 0.2554 0.2566 12 Edwards. " 99.44 1.0554 1.37029 0.2561 0.2570 09 " 20 100.00 Ammo 1.0507 nia '^: 1. 0000 1.37022 0.2571 0.2571 00 Damien. 16 00.00 i.333'i 0.2498 0.2498 00 Edwards. " 4.20 0.9816 1-33546 0.2559 0.2553 c6 " 10.27 0.9587 1. 33817 0.2636 0.2632 04 " " 14.63 0.9434 1.34071 0.2693 O.2711 18 " " 19.99 0.9256 1.34397 0.2766 0.2757 09 " 23.98 0.9130 1.34597 0.2615 0.2807 08 00 TOO. 00 0.00077 1.000384 0.4982 0.4982 00 Dulong. New Formula for Specific and Molecular Refraction. 629 ii Per cent, of Specific Index of re- fraction for (^-i) TLL ;. ill. From the I' dissolved. gravity (Z>). sodium light MD. 1, si 5" iir- data of Hydrochloric i6° acid ' 4° i6 00.00 0.9981 1-333" 0.2498 0.2498 00 Edwards. " 5-31 I. 0261 1.34522 0.2502 0.2497 05 " " 10.62 1.0527 1-35725 0.2501 0.2496 05 " " 16.55 I. 0814 1.35067 0.2501 0.2495 c6 " " 23.28 1.1160 1.38723 0.2508 0.2490 18 " " 30.45 I. 1530 1.40350 0.2500 0.2488 12 «» " 37-17 I. 1887 1.41858 0.2484 0.2490 06 <« 100.00 Alcoh I. 00162 01-^ i5°.5 1.000447 0.2764 0.2764 00 Mascart. 18 00.00 0.9981 1.33300 0.2495 0.2495 00 Edwards. " 3.82 0.9922 I-335I2 0.2529 0.2537 08 •' " 12.00 0.9801 1.3408S 0.2593 0.2596 02 '« " 20.00 0.9695 1-34683 0.2656 0.2661 05 " " 25.60 0.9627 1.35000 0.2693 0.2708 15 Landolt. " 28.63 0.95697 1-35231 0.2722 0.2733 n Edwards. " 37-44 0.9408 1-35697 0.2796 0.2807 II " " 46.90 0.9212 I-3S970 0.2872 0.2S85 13 " " 50.70 0.9168 1.3609 0.2893 0.2917 24 Landolt. " 56-77 0.8991 1-36233 0.2966 0.2970 04 Edwards. " 67.67 0.8737 1. 36381 0-3053 0.3059 06 " " 74.90 0.8603 1.36350 0.3099 0.3148 49 Landolt. " • 79-12 0.8464 1. 36451 0.3166 0.3154 02 Edwards. " 92.15 0.8128 1.36283 0.3275 0.3253 22 " " 98.41 0.7948 1.36095 0.3337 0.3365 28 " 22 100.00 0.7964 1.36060 0.3328 332 8 00 Landolt. In Table II the data of Damien have indices of refraction for hydrogen a instead of sodium light. III. — Table comparing the specific refraction ^"^—njfy^ and the molecular refraction — ^t/TS — of some organic compounds. Index of "r r ¥ Substance. Specific gravity. refraction for sodium light. {M—i) -WJT 1 % Stn From data of Hydrocarbo ns 'T . methane. 0.000718 1.000443 0.6106 9-77 Dulong. ( Bartoli and 1 Stracciati. 16° pentane. 0.6174 1-35830 0.4270 30-74 5-24 " hexane. 0.6765 1.38547 0.4112 35.36 4.62 " heptane. 0.7144 1.40362 0.4025 40.25 4.89 " " octane. 0.7290 1.41202 0.4002 45-b2 5-37 " " nonane. 0.7457 1.42073 0.3971 50.82 5.20 " " decane. 0.7561 1.42574 0.3949 S^ 07 5.25 " 630 Edwards. — _• „ ■ Index of =-- Substance. Specific gravity. refraction for sodium light. (M-i) i C 8 15 From data of Hydrocarbons i-j^ . 16° undecane. 0- 0.7673 i.43t36 0.3928 61.27 5.20 ( Bartoli and ( Straceiati. .. dodecane. 0.7769 1.43676 0.3913 66.52 5-25 '• tridecane. 0.7881 1.44227 0.3891 71.58 5.06 " « tetradecane. O.7S24 1.44809 0.3919 77-59 6.01 «' •< pentadecane. 0.8096 1.45319 0.3852 81.86 4.27 " " hexadecane. 0.8167 1.45697 0.3840 86.78 4.92 " Alcohols 20° 4° for Ha. 20 methyl. 0-7953 1.3279 0.3105 9.93 Landolt. " ethyl. 0.8000 1.36050 0.3312 15-23 5.30 " " propyl nor. 0.8044 1-3835 0.3446 20.67 5.44 Briihl. " butyl nor. 0.8099 I-397I 0.3509 25.96 5.29 " " amyl ferm. 0.8123 1.4057 0-35S3 31.26 5.30 Landolt. Aldehydes 20° 4° 0.7799 20 acetic. 1.32980 0.3180 1399 Landolt. " propylic. 0.8066 1.3616 0.3292 19.10 5-" Bruhl. " butyl. 0.8170 1.3822 0.3384 24.36 5.26 " «' valeric. 0.79S4 1.3861 0.348S 29.99 5-63 Landolt. " heptoic. 0.8495 1.4234 0.3501 39.91 4.96 Briihl. Acids - 20° 4°^" 20 formic. I. 2188 1.3693 0.2213 10.28 Landolt. " acetic. 1.0495 1.3699 0.2573 15-44 5.16 " " propionic. 0.9946 1.3846 0.2793 20.66 5.22 " " butyric. 0.9594 1-3955 0.2954 25.99 5-33 '< '« valeric (iso.) 0.9298 I 4022 0.3085 31-45 5.46 " " heptoic. o.gi6o 1. 4192 0.3224 41.91 5-23 " Esters _ 20° 20 f methyl ) \ valerate. / 0.8795 1.3927 0.3207 37-20 Landolt. ' /ethyl \ \ valerate. / 0.8661 1.3950 0.3270 42.51 5. 31 ' • r amyl \ \ valerate. / 0.8568 1.4098 0.3392 58.34 S-27 • ' fmethvl \ \ acetate. J 0.9039 1-3592 0.2923 21.63 5-25 ' j methyl \ \ butvrate. j 0.8962 1.3869 0.3113 31.75 5.06 ' ' j methyl S \ valerate. J 0.879s 1.3927 0.3207 37.20 5.45 ' j ethyl 1 ( formate. ( 0.9064 1.3580 0.2908 21.52 5-23 ' ' j ethyl 1 / acetate. / 0.9007 1.3707 0.3003 26.42 4-93 ' ' j ethyl ) Ibntyrate. \ 0.8892 1.3940 0.3177 3685 5-23 ' • J ethvl \ (valerate. / 0.8661 1.3950 0.3270 42.51 5.66 ' New Formula for Specific and Molecular Refraction. 631 . Index of 'ZTi g . i^ Substance. Specific gravity. refraction for sodium light. MD. •^ ^ 1 From data of Esters-^. 14° f ethyl i \ iodide. \ 1-9313 1.5067 O.T741 27.16 Gladstone. 16 f propyl \ \ iodide. / 1.7508 1.4979 0.1899 32.28 5.12 « « f isobutyl \ \ iodide. / I.59S2 1.4874 0.2039 37.52 5.24 " 14 j amyl 1 iodide. ) 1.5048 1.4884 0.2180 43-56 5.64 « A • 20° Amines —5- . 4 14 diethyl. 0.7092 1.3824 0.3887 17.49 Gladstone. " triethyl. 0.7317 1-3957 0.3861 28.18 5-34 " 23-5 propyl. 0.7140 1.3827 0.3907 23.05 5-13 " 23.2 dipropyl. 0.7356 1-3983 0.3907 39-46 547 " 22.8 tripropyl. 0-753S 1.4121 0.3905 55.84 5.28 " Aromatic Compounds : 10 benzene. 0.8903 1.5078 O.3711 28.94 " " toluene. 0.8704 1.4982 0.3754 34.53 5-53 " 18 xylene. 0.8632 1.4966 03773 40.00 5-47 " 19 mesitylene. 0.S632 1.4960 0-3759 45-" 5.10 " 12 cumene. 0.8432 1,4801 0.3785 45-42 5-42 " 21 phejiol. 1.0598 1.5509 0.3282 30.8s " " cresol. 1.0390 I. 5419 0.3318 35-83 4.98 " " metacresol. 1-033 T.5364 0.3313 35-78 4.93 " The numbers representing the molecular refraction — n^r) — of the paraffins show for each addition of CH2 an average in- crease for the Hydro- carbons, All taken together. Alcohols. Aldehydes. Acids. Iodides. Sahsr Amines. 5.1 1 5.33 5.24 5.28 5.33 5.31 5.24 5.28 This general average difference is very nearly that found from heptane to tridecane among the hydrocarbons, although the average difference for the hydrocarbons is lower than in any other group. The average difference for the aromatic compounds is 5.24. In the case of the ethereal salts it will he noticed that if we compare the numbers for the molecular refraction with corres- ponding ones for the acids themselves, the difference for CHs is greater than the average found from the table. For example, methyl acetate, methyl butyrate, and methyl valerate give, re- spectively, the numbers 21.63, 31-75 ^"^ 37-2o; while the corres- ponding acids give the numbers 15.44, 25.99 ^^^^ 31-45 \ the 632 Edzvards. difference between the salt and acid being 6.19, 5.76 and 5.65, respectively — an average of 5.86 for all. As another example we may take ethyl formate, ethyl acetate, ethyl butyrate and ethyl valerate, which give the numbers. 21.52, 26.42, 36.85 and 42.51 ; while the numbers for the corresponding acids are 10.28, 15.44, 25.99 and 31.45 ; showing differences re- spectively of 5.62, 5.49, 5.43 and 5.53— an average of 5.52. It will be understood that the average 5.52 corresponds with the „^ , . (5.26 + 5.86) average 5.86, bemg — . The numbers found for molecular refraction by using the Lorenz-Lorentz formula . }. — ^-75 shows a greater difference for a difference of CH2 when acids and ethereal salts are com- pared than is otherwise found, as the following numbers taken from Landolt's paper' show: Formic acid, 8.52; ethyl formate, 17.93; difference for CHj, 4.71. Acetic '* 12.93; methyl acetate, 18.03; " " " 12.93; ethyl " 22.14; " Butyric " 22.01; methyl butyrate, 26.79 ; " " " 22.01; ethyl " 31.20; " Valeric '* 26.72; methyl valerate, 31.46; " " " 26.72 ; ethyl " 35-98; " An average difference of 4.75, instead of 4.56 which Landolt finds for CH2 by comparing groups of alcohols, aldehydes, acids and ethereal salts in the usual way. Landolt, using the Lorenz-Lorentz formula, found for the four groups mentioned an average difference of — 4.61 for an increase of CHj among the alcohols, 4.57 " " " " the aldehydes, 4.56 " " " " the acids, 4.48 " " " " the ethereal salts. The general average for the four groups is 4.56. There is a difference of 0.13 between the highest and lowest average differ- ence, which is 2.85 per cent, of 4.56, the general average differ- ence. For corresponding groups in Table III, the average dif- ference for CH2 is 5.29, with a difference of 0.09 between the highest and lowest average difference, which is 1.7 per cent, of 5.29, the general average difference. The numbers in Table III have been taken from the same data and therefore show less per- centage error than the Lorenz-Lorentz formula, which Briihl has shown to give less percentage error than the Gladstone formula. ' Ann. Chem. (Leibig) 213, loi. " 4-61. " 4.78. " 4.58. " 4-74. •' 4.63. New Formula for Specific ayid Molecular Refraction. 633 IV. — Table comparing the specific refraction mh ^"*^ P(M— i) molecular refraction — ^>75 — , of some isomeric organic com- pounds. Substance. Specific Index of refraction {M-i) 1 C) i From the gravity. for Ha. MD ^ ^ ^ data of Allyl alcohol. 0.8563 I.4IH 0.3402 19.73 0.46 Gladstone. Acetone. 0.7920 1.3572 0.3323 19.27 0.46 Landolt. Propyl aldehyde. 0.8066 1.36.3 0.3292 19.10 0.17 Briihl. Methyl acetate. 0.9039 '•3592 0.2923 21.63 Landolt. Ethyl formate. 0.9064 1.3580 0.2908 21.52 0.1 1 « Normal propyl alcohol. 0.8044 '.3835 0.3446 20.67 Briihl. Isopropyl alcohol. 0.7887 1.3757 0.3465 20.79 0.12 " Butyric aldehyde. 0.8170 1.3822 0.3384 24.36 " Isobutyric aldehyde. 0.7938 1.3709 0.3408 24.53 0.17 " Normal butyric acid. 0.9587 1.3958 0.2957 26.02 " Isobutyric acid. 0.9490 1.3909 0.2961 26.05 0.03 - -isobutylsalicyl aldehyde. The quantitative formation of this condensation-product is a further proof of the fact already pointed out by Bradley and one of us, that a hydrocarbon group does not interfere with the con- densation of orthohydroxy aldehydes with acetyl chloride. The compound, is a very stable one ; boiling with a strong alkaline solution of potassium permanganate does not affect it. Bromine derivatives of paraisobutylsalicyl aldehyde. — One molecule of the dry sodium saltwas suspended in carbon bisulphide and one molecule of bromine added. The reaction was completed by heating on a water-bath. The solution was filtered from the sodium bromide, the carbon bisulphide distilled off, and the solid residue dried on a porous plate. It was then dissolved in alcohol ; from this solution water precipitates a nearly pure product. It was further purified by crystallization from light-boiling petro- leum. Analysis. 0.2030 gram substance gave 0.1491 gram AgBr. Calculated for CjjHjsBrOj. Found. Br 31.13 31.25 fOH (I) Paraisobuiylorthobroinsalicy I aldehyde, C6H2 S p tt / ^ , crys- I Br ' (6) tallizes in large, slightly yellow monoclinic plates that exhibit pleochroism. The angle /5 is 71° 15'. Its melting-point is 86°-87°. With ferric chloride it gives a violet color. When oxidized with potassium permanganate there is formed a crystalline product that inelts at 208°. This is possibly the/>-isobutyl-(?-bromsalicylic acid, but the quantity obtained was too small for analysis. Efforts to condense this bromaldehyde with the aid of acetyl chloride to a^z!)-diisobutyl-t?-dibromdisalicyl aldehyde were fruit- less. No alkali-insoluble product could be obtained. This would seem to indicate that the presence of a negative bromine atom together with a hydrocarbon group will prevent condensation. 644 Dains and Rothrock. { OH Paraisobutylorthobromsalicyl aldoxime, CeHs \ n y\ ' — I Br ' 5 grams of the bromaldehyde were dissolved in alcohol and the calculated amounts of sodium carbonate and hydroxylamine hydrochloride added. After standing for 24 hours the reaction was complete, and the product was dried and crystallizes from a mixture of benzene and petroleum-ether. The yield is good. Analysis. 0.2546 gram substance gave 0.1758 gram AgBr. Calculated for CiiH]4BrN03. Found. Br 29.41 29.38 The new body crystallizes in white, tabular, monoclinic plates that melt at 163°. It is easily soluble in ether, carbon bisulphide, hot alcohol, and hot benzene; difficultly soluble in cold alcohol, cold benzene, chloroform, or petroleum-ether. Dibenzoyl ether of paraisobutylorthobromsalicyl aldoxime, f O . CO . C6H6 CeHs i (- XT •—1-25 grams of the bromaldoxime were [ Br dissolved in caustic soda and the calculated amount of benzoyl chloride added. The new compound separates out in hard lumps which are very difficultly soluble in the hot alcohol from which it was crystallized. A7ialysis. 0.4220 gram substance gave 0.1642 gram AgBr. Calculated for CjsHjoBrNO^. Found. Br 16.66 16.58 The compound crystallizes from hot alcohol in fine white mono- clinic needles that melt at 189°. It is easily soluble in benzene, chloroform, and carbon bisul- phide; less so in ether and in light-boiling petroleum. fOH Paraisob7itylorthobromsalicylhydrazone,Q.6\{A^^ ' * \ ier — To an alcoholic solution of the bromaldehyde a little phenyl Parapropionic and Metapropionic Aldehydes. 645 hydrazine was added. The reaction goes slowly, but on standing there separate out the long needles of the hydrazone. These were purified by crystallization from light-boiling petro- leum. Analysis. 0.2183 gram substance gave 0.1180 gram AgBr. Calculated for C,,H,,BrN20. Found. Br 23.05 23.00 The hydrazone forms long monoclinic crystals that melt at 152°. It is easily soluble in ether, chloroform, and benzene; moderately soluble in alcohol, and difficultly soluble in light-boiling petro- leum. The authors desire to thank Professor Erasmus Haworth of the University, who has very kindly identified for them the crystalline forms of some of the preceding compounds. Contributions from the Chemical Laboratory of Cornell University. THE POLYMERIC MODIFICATIONS OF PROPIONIC ALDEHYDE: PARAPROPIONIC AND META- PROPIONIC ALDEHYDES. By W. R. Orndorff and Miss L. L. Balcom. It has been shown in a previous paper ' by one of us that the polymeric modifications of acetic aldehyde, paraldehyde and metaldehyde have the same molecular weight (three times that of the simple aldehyde), and that metaldehyde is the unstable form changing very readily into a mixture of paraldehyde and a solid product having a molecular weight four times that of the acetic aldehyde. In this paper attention was also called to the fact that paraldehyde and metaldehyde differ from each other only in their physical properties, such as solubility, crystal form, volatility, melting-point, etc., and that their entire chemical con- duct could be accurately represented by the same structural for- iThls Journal 16, 43- 646 Orndorff and Balcom. mula in a plane. These facts led to the idea that the two sub- stances were stereoisomers, paraldehyde the stable form being best represented by the trans-formula and metaldehyde by the cis-formula : H3C H HsC H >C< >C< 00 00 HbC>. L i .CHs HaC^ I I CHi Paraldehyde.! Metaldehyde.i These two formulas were then shown to be in accord with the entire conduct of the substances. Polymerization, it is well known, is a characteristic property of the aldehydes, many of the well-known aldehydes occurring in polymeric modifications and some of them in more than one variety. Propionic aldehyde (C2H6CHO), for example, has been shown by one of us^ to polymerize and form a parapropionic and a metapropionic aldehyde. The first product is a liquid which resembles paraldehyde very closely, the second is a solid anal- ogous in every respect to metaldehyde. This resemblance being so striking, it was thought that perhaps these two substances were also stereoisomers, and that all aldehydes (with the exception of formic aldehyde) would be found to yield two isomeric trialde- hydes which were stereoisomers. Parapropionic and metapro- pionic aldehydes were therefore made and carefully studied in order to determine their molecular weights and the relation exist- ing between them and the simple propionic aldehyde from which both are made. Parapropionic Aldehyde. This substance was made from the purest propionic aldehyde by the polymerizing action of hydrochloric acid gas, the hydro- chloric acid removed by means of dry barium carbonate, the pro- duct dried with calcium chloride and subjected to careful fractional distillation, using a Hempel distilling-bulb. The parapropionic aldehyde thus made and purified was found to boil, with slight decomposition into propionic aldehyde, at i69°-i7o° C. (uncorr.) ' The broken lines represent the groups below, the unbroken lines those above, the plane of the paper, a This Journal 12,353. Parapropionic and Meiapropionic Aldehydes. 647 and to solidify completely at — 20° C, to a mass of needle-shaped crystals. A determination of its specific gravity showed that as compared with water at 4° C. its specific gravity at o" C. is 0.9549. It is only very slightly soluble in water, though freely soluble in ether and alcohol. It resembles paraldehyde closely, havifi^ a very similar odor, and being converted into the simple aldehyde (propionic aldehyde) when distilled with a small quantity of sul- phuric or hydrochloric acid. It reacts neither with fuchsine sul- phurous acid nor with hydroxylamine, and hence does not con- tain the aldehyde group CHO. Determinations of the molecular weight of parapropionic alde- hyde by the cryoscopic method, using the Beckmann apparatus and thermometer, gave the following results : Parapropionic aldehyde (C3H60)3=r 174. Solvent, phenol'; molecular depression, 75 ; depression coefficient, 0.4310. Mol. Wt. JVt. Solvent. Wt. Sub. Concentr. Depress. Dep. Coef. Mol. Dep. found. 26.8345 0.2100 0.7826 o°.345 0.4408 76.7 170 0-3445 1.2834 0^.565 0.4402 76.6 170 0.5125 1.9099 0^835 0.4372 76.0 172 26.8345 0.1360 0.5068 0°.220 0.4341 75-5 173 0.2890 1.0770 0^470 0.4364 75-9 172 These results show that the molecular weight of parapropionic aldehyde in solution in phenol is about 174 or three times that of propionic aldehyde (58), and that consequently its formula is (C3H60)3 or C9H18O3. In this connection it should be stated that S. Reformatsky " recently obtained parapropionic aldehyde by the action of zinc and ethyl monochloracetate on propionic aldehyde. Analyses of the substance showed that it had the formula (C3H60)^, and a vapor density determination by the Victor Meyer method in the vapor of boiling naphthalene gave its specific gravity as referred to air ^s 5-95, while that required by the formula (C3H60)3 is 6.02. Meiapropionic Aldehyde. This substance was made according to the directions already given .' The substance is formed with much greater difficulty than ' The phenol used in these determinations was prepared from the c. p. article by distilla- tion and melted at 4i°-42° C. (uncorr.) In working with phenol care was taken to exclude moisture by means of the device already described. See this Journal 15, 353. * Jour. Russ. Chem. Soc. 32, 197. ' Loc. cit. 648 Oryidorff and BaJcom. the metaldehyde, and unless the directions are carefully followed and the purest propionic aldehyde is used, the amount of meta- propionic aldehyde formed is liable to be very small. The meta- propionic aldehyde was filtered from the parapropionic aldehyde always formed simultaneously with themeta-product, washed with water, drained thoroughly, and dried first in the air on drying paper and then in a Hempel desiccator containing sulphuric acid. The metapropionic aldehyde thus prepared resembles metalde- hyde very closely, and like metaldehyde it is converted into the simple aldehyde from which it is derived, by heating with sul- phuric or hydrochloric acid. With fuchsine sulphurous acid and with hydroxylamine it gives no reaction and hence does not con- tain the aldehyde group CHO. Determinations of the molecular weight of the freshly prepared metapropionic aldehyde by the cryoscopic method, using the Beckmann apparatus and thermometer, gave the following results : Metapropionic aldehyde, (CEH60)a=: 174. Solvent, phenol'; molecular depression, 75 ; depression coefficient, 0.4310. Vt. Solvent. \Vt. Sub. Concentr. Depress. Dep. Coef. Mol. Dep. Mol. Wt found. 24.9265 0.1035 0.4152 o°.i8o 0.4335 754 173 O.I9IO 0.7662 0°.320 0.4176 72.7 179 24.9265 0.0840 0.3370 o°.i50 0.4451 77-4 169 0.1445 0.5797 o°.255 0.4399 76.5 170 0.2148 0.8613 o°.375 0.4354 75-7 172 These results certainly show that the freshly prepared metapro- pionic aldehyde has the same molecular weight (174) and the same formula as parapropionic aldehyde, viz. (CsHeOjs. After the metapropionic aldehyde had stood for some weeks at the temperature of the laborator)^ (about 22° C.) it was found to give a strong odor of the parapropionic aldehyde when powdered. Determinations of the molecular weight in solution in phenol of the solid product remaining (after the odor of parapropionic aldehyde had disappeared from the powdered product) gave the following results : Metapropionic aldehyde, (C3H60)s= 174 ; (CsH60)4 = 232. Solvent, phenol. Vt. Solvent. Wt. Sub. Concentr. Depress. Dep. Coef. Mol. Wt Mol. Dep. found. 21.4055 O.IOOO 0.4671 o°.i6o 0.3426 79.4 218 0.1945 0.9086 iSee footn o°.3io oie, page 647. 0.3412 79.1 219 Parapropionic and Metapropionic Aldehydes. 649 These results show that the molecular weight of this substance is intermediate between that required for the formula (C3H60)s and (CsHeO)*, and would seem to indicate that the solid product remaining is a mixture of a tetramolecular product (C3H60)4 and some of the unchanged metapropionic aldehyde. After the metapropionic aldehyde had stood for about two months it was found to closely resemble metaldehyde that had stood for some time. The crystals, which had become dull and opaque, were much more easily powdered, and gave off at the same time a very strong odor of parapropionic aldehyde. Mole- cular weight determinations in solution in phenol gave the follow- ing results : Solvent, phenol ; (C3H60)3= 174 ; (CaHsO)*^: 232. Wt. Solvent. Wt. Sub. Concentr. Depress. Dep. Coef. Mol. Wt. Mol. Dep. found. 27-3995 0,0836 0.3051 o°.i05 0.3441 79.8 218 O.I716 0.6263 0°.2I5 0.3433 79.6 218 0.2441 0.8909 o°.305 0.3424 79.4 219 showing again a mixture of thetetrapropionic aldehyde (CsHeO)* and some of the unchanged meta-product. Metapropionic aldehyde like metaldehyde then decomposes on standing into the para compound and a tetramolecular pro- duct, and is hence the unstable form. It should be noted, however, that metapropionic aldehyde does not decompose so readily as metaldehyde does. From these results it must be inferred that parapropionic alde- hyde and the freshly made metapropionic aldehyde have the same molecular weight and the same empirical formula. Now as they both easily revert to the simple aldehyde by the action of the same reagents which bring about the polymerization of that aldehyde, and as neither contains the aldehyde group CHO, it is quite likely that the oxygen atoms here play the part of connect- ing atoms, and that both substances are to be represented by the same structural formula in a plane : OH5 H >C< o o GHs I I GH5 Vol. XVI.-48. 650 Reviews and Reports. which formula is in strict accord with the entire chemical conduct of the two substances. A careful study of the chemical conduct of these two sub- stances leads inevitably to the conclusion that they are stereo- isomers. The simultaneous formation of both by the action of the same polymerizing agents on the simple aldehyde, the conver- sion of the metapropionic aldehyde into the para-product and the solid polymer, and the fact that the chemical conduct of both substances is absolutely identical with the same reagents, can only be explained on this assumption. Referring to the structural formula, it will be seen more clearly by using models' that it admits of two configurations in space, and only two, as follows : C2H5 H >C< O O Trans-form. Cis-form. The trans-formula is generally assigned to the more stable modification, so that parapropionic aldehyde is probably best represented by that formula, while the less stable metapropionic aldehyde in all probability has the cis-formula. Further work on the polymerization of the aldehydes and the relation between these polymeric modifications is in progress in this laboratory and will be reported on at an early date. Cornell University Chemical Laboratory, October, 1894. REVIEWS AND REPORTS. Select Methods in Chemical Analysis. By William Crookes. Third edition, rewritten and enlarged. The revised edition of this work will be welcomed by the many who have found the previous editions so useful in their laboratory practice. The work is so well known among chemists that an > The Kekule models are excellent for this purpose. Reviews and Reports. 651 extended explanation of the author's object in writing it is unneces- sary. The views and purposes which have guided him in the preparation of the last revision will be best presented by means of liberal quotations from the preface : "A third edition of ' Select Methods in Chemical Analysis ' hav- ing been called for, advantage has been taken to go over the whole work and remove some of the processes to make room for others which have been proposed and found to be successful during the eight years which have elapsed since the second edition was published. It must, however, not be assumed that the processes so discarded are of little value. Indeed, some are at the present time in constant use, having taken their position as regular laboratory processes, and have only been removed from these pages because their value is now too well known to make it advisable to retain them in a book which the author wishes to be looked upon as mainly a collection of novel or little-known pro- cesses. As soon as a new process takes its place among ordinary Jaboratory processes there is no reason for its retention here. Other processes have been omitted because further experience with them has shown the author that they are not so trustworthy as other newer processes which have taken their place. Others again have been omitted to prevent the book becoming of an unwieldy size. Thus most volumetric operations have been omitted, as there are now several standard works which are devoted to this branch of analysis. Then, mere detections which are not separations, and processes of only technical importance, have to a great extent been left out ; the latter are well provided for in the technical literature. It has also been thought advisable to omit many purely assaying and furnace operations, as not exactly coming within the scope of the book and being more fully treated in special works on assaying " The space gained by these omissions has been partly filled up by new processes which the author has considered worth intro- ducing; but chiefly it has been utilized in giving to the chemical world a series of electrical separations and other processes from the standard work of Dr. Classen " The author wishes to point out that this book is not to be looked upon as an encyclopaedia of chemical analysis, in which is laid down every good method for the qualitative and quantitative examination of every known substance under every possible com- bination of circumstances. The author has merely given such methods as have been proved in his own laboratory. Others — possibly no less efficient — have been passed over because he can- not vouch personally for their value. A main object has been to bring into notice a number of little-known expedients and pre- cautions which prevent mistakes, insure accuracy or economize time." H. N. M. 652 Reviews atid Reports. Quantitative Chemical Analysis by Electrolysis. By Dr. Alexan- der Classen, Professor of Chemistry and Director of the Inorganic Laboratory in the Royal School of Technology at Aachen. Authorized translation. Second English from the third German edition. Re- vised and greatly enlarged by William Hale Herrick, A. M. The work consists of three parts. Part I is general and treats of the various sources and the control of the current, and of the methods of determining and separating the individual metals. Part II is special, and deals with the analysis of alloys, minerals and metallurgical products. Part III contains tables designed to facilitate computation, directions for the preparation of reagents, and a record of analytical resuhs. The original work is necessa- rily, to a very great extent, a record of the author's own valuable contributions to this branch of analytical chemistry, though the results and observations of others are not neglected. The claim upon the title-page of the translation, that the original has under- gone a considerable enlargement, is scarcely warranted by the jfacts. The translatoi's contributions consist in the main of a description of the gravity cell, which is recommended as a substitute for the Daniell and the Meidinger; an account of the Paget ther- mopile, which, it appears, is more efficient than that of Clamond or of Noe ; a brief discussion of the secondary battery ; a descrip- tion of Smith's current-reducer and Malapert's stand ; and. finally, of some half a dozen references to the work of Smith and Knerr, and one to that of Chittenden and Blake. h. n. m. Lessons in Qualitative and Volumetric Chemical Analysis for the Use of Physicians, Pharmacists and Students. By Dr. Charles O. Curtman, Professor of Chemistry in the Missouri Medical College, including Lessons in Qualitative' Chemical Analysis, by Prof. F. Beilstein of St. Petersburg. Fourth edition, revised and enlarged, with many illustrations. John L. Boland Book Co., St. Louis. 1894. In taking Professor Beilstein's book as basis of a manual for medical students, Prof. Curtman made a wise choice. Beilstein's method of training the beginner by the careful study of common salts and compounds is excellent. The student learns to think, to observe, and to remember, and to rely on his knowledge and memory rather than on a " table." Prof. Curtman has built on the basis of Beilstein's manual in the same spirit. Organic compounds of particular interest to the medicaf student are carefully studied, and the reactions by which they can be detected are given. Among the substances studied we note salol, salicin, pyrogallol, acetanilid, antipyrin, other new organic remedies, and the vegetable alkaloids. The volumetric methods are the best in use. Particular attention is given to the analysis of drinking-water and to urine-analysis. This book can Reviews and Reports. 653 be recommended as a valuable laboratory manual, and equally valuable as a book of reference for the physician, pharmacist, or chemical student. e. r. A Brief Introduction to Qualitative Analysis, for Use in Instruc- tion IN Chemical Laboratories. ByLuowiG Medicus, Professor of Chemistry in the University of VVlirzburg. Translated from the fourth and fifth German editions, with additions, by John Marshall, Assis- tant Professor of Chemistry in the Medical Department of the Univer- sity of Pennsylvania. Third edition. Philadelphia : J. B. Lippincott Company, 1894. The laboratory manual of Prof Medicus has earned a widespread reputation as a good and thorough book of that school which trains the beginner largely through use of tables, in which the name, formula and appearance of the expected precipitate are described. Prof Marshall has added to his translation new tables, including a table of solubilities, and has amplified the text to the extent of forty pages. e. r. A Manual of Microchemical Analysis. By Prof. H. Behrens, of the Polytechnic School in Delft, Holland. With an introductory chapter by Prof. John W. Judd, F. R. S., of the Royal College of Science, London. With 84 illustrations drawn by the author. Macmillan & Co., New York. $1.50. Prof Behrens' book is the first laboratory manual on this com- paratively new branch of analysis, and is for this reason of great interest. In the first Part, the apparatus, the reagents, and the reactions of the single elements, are treated of Numerous illus- trations show the appearance of the crystalline precipitates as seen through the microscope. In Part Second methods are given for the systematic analysis of water, ores, rocks, alloys, and combinations of rare elements. The chapter on the microchemical examination of rocks, first by study of slides, secondly by study of powdered rocks, is particularly valu- able. Of almost equal interest and value is the chapter on iron and steel, on copper and its alloys, on the alloys of lead, tin, anti- mony, and on those of the precious metals. For petrographic research this manual is invaluable, but, to quote from Prof Judd's introduction, " It is evident that these methods may be often employed with advantage in the ordinary chemical laboratory. . . . Archaeologists and metallurgists, too, will find the methods for examining alloys of great service, especially in cases like those of manufactured articles or objects of art in which only very minute quantities of the material are available for analysis." That Prof Judd revised the manuscript of this book and wrote the introduction is in itself evidence of the value of the work. E. R. 654 Reviews and Reports. Manual of Physico-Chemical Measurements. By W. Ostwald. Translated by J. Walker. Macmillan & Co., 1894. This laboratory manual, the original of which appeared about a year and a half ago (see this Journal 15, 677), is now accessible to those who use only English. This work is not for the beginner, but, in the words of the author, " for the chemist or the physicist who has already gone through the greater part of his special course and recognizes the necessity of making himself acquainted with the borderland between the two sciences, which has of late attained to such importance." The subjects treated are : Calcu- lation ; measurement of length ; weighing ; measurement and regulation of temperature; thermostats; glass-blowing; measure- ment of pressure, volume and density; boiling-point; vapor- pressure ; calorimetry ; optical measurements ; viscosity and sur- face-tension ; solubility; determination of molecular weights in solutions ; electrical measurements ; chemical dynamics. This English edition has appeared at a time when there is a growing interest manifested in physical chemistry, and it will doubtless contribute to work in this line in English and American laboratories. By accomplishing this the translator will have worked to good purpose. The translation is admirable, being true to the original and at the same time written in clear and concise English. h. c. j. Laboratory Manual and Principles of Chemistry, for Beginners. By George M. Richardson, Associate Professor of Chemistry in the Leland Stanford Junior University, xii+233 pp. i2mo. New York : Macmillan & Co., 1894. Dr. Richardson's book would perhaps be an ideal one in an ideal laboratory. He wishes to dispense with all teaching by text- book or lecture until the student has learned in the laboratory "the value of careful observation," "with such aid as he may get from his teacher for his chemical facts." It is a fair question to ask how long will this take ; how long is the bewildered student to collect experimental data before he may be said to be in a posi- tion to be told the connection between his work and the facts of chemistry? In an ideal laboratory each student will have at his command ample room, all the apparatus desirable, unlimited time, and the constant attention of a thoroughly competent instructor. Under such conditions lectures and text-books are not necessary, nor are they desirable; but such conditions do not exist. As the world is at present, a student must be taught the reason for doing a given thing before he can be expected to do it well. A series of quantitative experiments requiring abundant time and carefully- prepared — sometimes complicated — apparatus, is, in the experi- Revieivs and Reports. 655 ence of the present writer, of doubtful value when the very prin- ciples those experiments are intended to establish are as yet unknown to the student. It may be noticed that many of the quantitative experiments described belong strictly to an elemen- tary course in physics and have no place here — not because they are not of great value, in fact essential, to a student of chemistry, but for the reason that every such student should have a good working knowledge of physics and should perform these experi- ments in connection with his physical course, especially as most chemical laboratories are not so arranged and equipped that such work can be done in them to proper advantage. While fault may be found with Dr. Richardson's book on account of the character of the experiments described and of the order in which the various subjects are taken up, thoughtful teachers must in general commend the emphasis which he lays upon the value of laboratory work well performed. With this book as a laboratory manual supplemented by a parallel course of instruction in the fundamental facts of chemistry, excellent results should be obtained. The character of the experiments to be performed should then, of course, be varied somewhat to meet the capacity of the individual student. w. w. r. An Elementary Manual of Chemistry. By F. H. Storer and W. B. Lindsay. Pp. 436, 8vo. The American Book Company, 1894. This new revision of Eliot and Storer's Manual is an enlarge- ment upon that of Prof. Nichols. One hundred pages have been added to the book; the number of experiments detailed has been increased and the work of recent years has been incorporated. One or two points connected with the order in which the sub- jects treated of are taken up deserve, in the opinion of the present writer, adverse criticism. The Periodic Law is apparently not referred to until the last chapter in the book is reached. Again, the study of the hydrocarbons and their derivatives, " radicals " and constitutional formulae, is taken up before the student has become acquainted with any of the metals and even before the salts of ammonium have been more than casually referred to. In this connection it may be noted that the statement on page 203, that the methyl group may be compared with "the group NHs (ammonia) " as similar " radicals," is probably a misprint. The book is full — too full, many may think — of detailed fact, let alone hypothesis. Still, great care seems to have been exer- cised to make as clear a presentation of the subject as the space at hand would allow. w. w. r. 656 Reviews and Reports. EinfOhrung in das Studium der qualitativen chemischen Analyse. Von Carl Friedheim, Ph. D., Privat-Docent an der Universitiit Berlin. Berlin : Carl Habels Verlag, 1894. Pp. 335, 8vo. Dr. Friedheim has written on the basis of Rammelsberg's " Leitfaden " a good and thorough elementary manual of quali- tative analysis. The reactions of the metals and acids are more fully explained than is usual in such manuals : for example, he explains the behavior of the hydroxides of zinc, copper and nickel, and of silver oxide toward ammonia; the action of potas- sium cyanide on salts and sulphides, and similar points obscure to the beginner. For the student of chemistry in Germany, who hears lectures on inorganic chemistry in his first semester, and does not begin laboratory work till the second, the interval of time between lectures and laboratory tends to blunt the memory of the lecture — the more so as there are generally no stimulants provided in the form of our " quizzes," and no examinations till the " Doctor Examen." In the laboratory, qualitative analysis is taken up at once, without previous experiment in general chemistr)'. In the preface Dr. Friedheim deplores this state of things at length, and speaks openly of the ignorance of general chemistry shown by the average German student ; he hopes that study of his book will better matters. The present writer, from experience in teaching in Germany and here, endorses all that Dr. Friedheim says in his preface of the state of things in Germany, but is of the opinion that in the better American colleges general chemistry is so thoroughly taught, not only in lectures, but in laboratory, " quiz " and examina- tion, that there is no need here of Dr. Friedheim's book. It is too large for a college student, too brief and too explanatory for a specialist. However, the book is written for German students, and it is to be hoped that it will gain entrance into many of the German university laboratories. e. r. Inorganic Chemistry for Beginners. By Sir Henry Roscoe, assisted by Joseph Lunt. Pp. 241, i2mo. New York : Macmillan & Co., 1894. In this book the field is limited to the elements oxygen, hydrogen, nitrogen, chlorine, sulphur and carbon, and their most important compounds. At the same time the treatment is much fuller than in the author's well known " Lessons." The physical properties also of the substances dealt with are discussed with some degree of detail. Professor Roscoe is still of the opinion that the instruction of a student in the elements of chemistry should begin with a consid- eration of the Atomic Theory. w. w. r. INDEX VOL. XVI. ACETACETIC ETHER, the condensation products of the aromatic hydrazides of . 430 Acetamide, electrolysis of 587 Acetic acid, action of water and ethyl alcohol on the freezing-point of . . . 15 " " pure, preparation of 5 " " pure, table of the lowerings of the freezing-point of, by small quantities of water 6 " " pure, table of the o werings of the freezing-point by small quantities of sulphuric acid . 7 " " pure, table of the lowering of the freezing-point by sulphuric acid and water 8 Acetic aldehyde, the polymeric modifications of, paraldehyde and metaldehyde . 43 Acetic and propionic acids, affinity-constants of some sulphur substitution products of 559 Acetic ether, an isothermal curve of the solubility of mercuric and sodium chlorides in 214 Aceto-acetic ether, electrolysis of 583 Acetobromamide, action of sodium methylate on 372 Acetone dicarbonic ether, electrolysis of 584 Acetyl-acetone, oxidation experiments with 572 Acetylene, phenomena of oxidation 164,181 " qualitative reactions for 340 Acetyl ether of^-isobutylsalicylonitrile 640 Acetyl malonic ether, electrolysis of 582 Acids, affinity-constants of 318 " organic, electrosyntheses by the direct union of anions o 569 " weak, affinity-constants of 313 Address of Emil Fischer before the Berliner Akademie der Wissenschaften . . 159 Affinity-constants of bases 473 " " " some sulphur substitution-products of acetic and propionic acids 559 " " " weak acids and the hydrolysis of salts 313 Alcohol and water, action on the freezing-point of acetic acid 15 Alcohol vapor, phenomena of oxidation 183 Aldehyde, acetic, polymeric modifications of, paraldehyde and metaldehyde . . 43 Aldoximes, velocity of transition of into acid nitriles 551 Alkyls and olefines, halogen compounds of 364 Allylene, qualitative reactions for 342 Amalgam of ammonium, note on the influence of certain metals on the stability of . 488 Amines, nitrites of some 449 1-4-Aminohydroxycyclohexane chloroplatinate 453 Amino-lauronic acid 506 " " " anhydride 507 Ammonium amalgam, note on the influence of certain metals on the stability of . 488 Animals, action of definitely related chemical compounds upon 443 Antimony and potassium, mixed double halides of 490 Aromatic isocyanides, addition-products of 372 ArrheniuSfS. Hydrolysis of salts of weak acids and weak bases .... 558 BALCOM, Miss L. L. See Orndorff, IV. R. Ballard, H. H. See Noyes, W. A. 658 Index. Baric dibromdinitrophenylate 34 " tribromdiiiitrophenylate 31 Barium benzene-sulplion-periodide 121 Bases, affinity-constants of 473 " organic, a new class coataining iodine but no nitrogen 233 Betson, J. /.. A study of the action of the salts of diazobenzene on methyl and ethyl alcohols under different conditions 235 " A study of the constituents of the nodes and internodes of the sugar cane 457 Benzamide, electrolysis of 587 Benzene, phenomena of oxidation of 182 Benzenesulphonic acid, action upon potassium iodide 116 Benzoin, two stereoisomeric hydrazones 108 Benzoyl bromamide 217 " chloramide 218 Benzoylformoparatoluide 383 Benzoyl halogen amides 216 " iodoamide 218 Benzyl ether of paraisobutylsalicylic aldehyde . . 641 Benzyl-malonic ether, electrolysis of 582 Blalock, T. L. See Morse, H. N. Bodenstein, M. Dissociation cfhydriodic acid gas 474 BrauH, F. Continuous conduction of electricity by gases 471 Bredig, G. Affinity-constants of bases 473 " Stoichiometry of the ion-velocities 473 Breed, Mary B. See Keiser, E. H. a-Bromcrotonolactone 277 /3-BromcrotonoIacton 211 CALIBRATION and graduation of volumetric apparatus, instruments for . . 479 Campholytic acid and salts 505 Camphoraminic acid, ortho methyl ester of 308 a-Camphoraminic acid 502 p-Camphoraminic acid 309, 503 Camphoric acid 307, soo " " preparation 501 " anhydride 502 " imide 502 Canadian sulphur-petroleum, examination 89 Carbon monoxide, action on methane (natural gas) at high temperatures . . . 275 " " phenomena of oxidation 164,183 " " qualitative reactions for 272 Carbon oxysulphide, qualitative reactions for 344 Chemical compounds, action of definitely related, upon animals 443 Chemical equilibria as temperature-functions 611 a-Chlorcrotonolactone 290 )3-Chlorcrotonolactone 288 Chlorformanilide 71 Chlorine, substitution-products of the action of upon methane 361 Ciamician, G. The influence of chemical constitution of organic compounds upon their capacity to form solid solutions ......... 560 Clement, A. A. Ste JVoyes, A. A. Color of salts in solution 326 Condensation-products of aromatic hydrazides of acetacetic ether .... 430 Cooke, J. P. Obituary note 566 Cornelison, R. IV. See Hill, H. B. Cream of tartar, solubility in alcohol 4^4 Cresols, action on animals 44S Index, 659 Crotono-lactones and mucobromic acid, substituted 188, 277 Cyanacetic ether, electrolysis of 585 DAINS, F. B., and Rothrock, I. R. On paraisobutylic-salicyl aldehyde and some of its derivatives 634 De Chalntot, G. Pentosans in plants 21S, 589 " Note on pentosans in soils . 229 Dennis, L. M., and Kortright, F. L. Upon the separation of thorium from the rare earths of the cerium and yttrium groups by means of potassium hydronitride . 79 Des Coudres, T. Electromotive force between two differently curved mercury elec- trodes in a solution of a mercury salt 561 " Non-polarizable electrolytic cells under the influence of centrifugal force 561 " Velocity of polarization in closed concentration cells . . . 562 Diacetyl-paraisobutyl-salicyl hydrazone 637 Diacetyl-succinic ether, electrolysis of 583 Diamido-orthophosphoric acid . 128 Diamidophosphoric acid, action of nitrous acid on 132 " " salts of 133 Diamidotrihydroxyl phosphoric acids 123 " phosphates 140 2-5-Diaminohexane nitrite 456 Diazo-alcohol reaction, effect of the presence of water upon 243 Diazobenzene nitrate, action on ethyl alcohol in the presence of alkalies . . . 247 " '* " " methyl alcohol in the presence of alkalies and of calcite 246 " " " " methyl alcohol 235 " " " " methyl and ethyl alcohol in presence of an excess of zinc dust ......... 252 " " " " sodium ethylate 2^7 " • " " " sodium methylate 244 Diazobenzene, salts of, a study of the action of, on methyl and ethyl alcohols under different conditions 235 Diazobenzene sulphate, action on methyl alcohol 250 " " action on methyl and ethyl alcoho in presence of excess of zinc dust 251 " " action on sodium methylate and ethylate .... 250 Diazo compounds, on the decomposition of 235 Dibenzoyl ester of paraisobutylsalicyl-aldoxime 639 " ether of paraisobutylorthobromsalicyl-aldoxime 644 a;8-Dibromcrotonolactone ............. 200 ajS-Dibromcrotonolactone to crotonolactone 283 Dibromdinitrophenol, properties 33 Dibromfumaric aldehyde 203 a(3-Dichlorcrotonolactone 285 Dielectric constants and chemical equivalents SS4 " " methods of estimating S54 " Die Lehre von der Elektrizitat " (C IViedemann), review ..... 563 Diethyl-elhane-tetracarbonic ether ........... 581 " Die wissenschaftlichen Grundlagen der analytischen Chemie " ( JV. Ostwald], review 564 Dihydro-amino-campholytic acid 310, 503 " " " " anhydride of 504 " " " " methyl ester of . . 308 Dimethyl-ethane-tetracarbonic ether, electrolysis of 578 Dinitrophenol, action on animals 448 Dissociation of hydriodic acid gas 474 Dissociation pressures, small, estimation of, especially in salts containing water of crystallization 55^ 66o Index. Double halidesofantimony and potassium 490 Duhem.P. General theorem upon the state of a body in solution .... 557 EDWARDS, W. F. A new formula for specific and molecular refraction . . 625 " Einfuhrung in das Studiumder, qualitativen chemischen Analyse" (C Friedheim), review 656 " Electrochemie, ihre Geschichteund Lehre " (ff'. 'f«j), review 653 Molecular and specific refraction, anew formula for 625 " refractions of some organic compounds 557 " weight estimations in solid solutions 553 Moller, G. See Jahn, H. Monamidophosphoric acid, note on 154 a-Monophenylhydrazone 113 P-Monophenylhydrazone . 112 Morse, H. N., and Blalock, T. L. Instruments for the graduation and calibration of volumetric apparatus 479 Mucobromic acid 297 " " substituted 188, 277 Mucobromamide 302 Mucobromoxime 298 " anhydride 298 " methyl ester _ 300 Mucochloramide 305 Mucochloric acid ^ 303 Mucochloroxiine 304 " methyl ester 304 Mucophenoxybromic acid 306 Mucophenoxybromoxime 306 Mucophenoxychloric acid 306 Mucophenoxychloroxime ■ 306 NATURAL GAS, action of carbon monoxide on, at high temperatures . . . 275 Natural gas "and petroleum, origin of 417 " " composition of 406 " " quantitative analysis of 412 Nernst, W. Dielectric constants and chemical equivalents 554 " Methods of determining dielectric constants 554 Nitranilines, action on animals 448 Nitrites of some amines 449 Nitrobenzol, action on animals 448 Nitrogen halogen compounds 37° " reactions for 357 Nitrophenols, action on animals 447 Nitrosomenthene 396, 397 Noyes,A.A. The splitting off of hydrogen from acid potassium tartrate . . . 558 " and Clement, A. A. The electrolytic reduction of paranitrobenzoic acid in sulphuric acid solution 511 " and Clement, A. A. Solubility of the acid potassium tartrate in the presence of other salts 558 Noyes, W. A. Camphoric acid 307.500 " 3SiA Ballard, H. H. The nitrites of some amines .... 449 OHIO PETROLEUM, sulphur compounds in 83 Olefines, halogen compounds of 364 " phenomena of oxidation 164 " qualitative reactions for 265 Orcinol, action upon ortho-sulpho-benzoic acid 524 Orcin-sulphon-phthalein 528 Organic bases containing iodine but no nitrogen, a new class of . . . . 233 " compounds, volatile, on the determination of sulphur in 544 664 Index, Organic periodides, a new class of ii6 Orndorff, W. R., and Balcom, Miss L. L. The polymeric modifications of pro- pionic aldehyde, parapropionic and metapropionic aldehydes . 645 " and White, J , The polymeric modifications of acetic aldehyde, paraldehyde and metaldehyde 43 Ortho-sulpho-benzoic chloride, a reduction-product o 366 Ortho-sulpho-para-toluic acid, phthaleins of 513 i-OrthotoIyl-3-methyl-s-pyrazolone 442 Oxygen, qualitative reactions for 358 PALLADIUM, atomic weight of 20 Palladium chloride, distillation of in a current of chlorine 21 " diammonium chloride, analysis of . 33,26 " purification in the wet way 24 Parachlor-metasulpho-benzoic acid and some of its derivatives 530 Paradiisobutyldisalicylic aldehyde 642 Paraffins, gaseous, phenomena of oxidation 164 Paraisobutylorthobromsalicyl aldoxime . 644 " hydrazone 644 Paraisobutylsalicylic aldehyde and some of its derivatives 634 " " bromine derivatives of 643 " aldoxime . ., 638 " hydrazone 637 Parisobutylsalicylidene aniline 638 Paraisobutylsalicylonitrile 640 Paraldehyde and metaldehyde, the polymeric modifications of acetic aldehyde . 43 Paramethyl-sulphon-gallein 526 Paranitrdbenzbromamide, action of sodium methylate on 370 Paranitrobenzoic acid, electrolytic reduction of in sulphuric acid solution . . . 511 Parapropionic and metapropionic aldehydes 645 Paratolylimidoformic ethyl ester 377 Paratolylimido thioformic ethyl ester 377 Paratolyl isocyanide, preparation ........... 373 " " reduction of 378 " " and silver cyanide 387 Peddie, W. Law of transformation of energy and its applications .... 560 Pentane, phenomena of oxidation 175 Pentosans in pine wood 611 " " plants 218, 589 " " seeds 589 " ** soils .............. 229 " quantitative estimation of 218 Periodides, a new class of organic 116 Petroleum and natural gas, origin of ...... . ... 417 " Ohio, sulphur compounds in 83 " oils and refining residues, composition of 83 Phenol, action on animals 443 Phenol-para-methyl-sulphon-phthalein 514 a-Phenoxy-(3-bromcrotonolactone 291 a-Phenoxy-|3-brom-y-oxycrotonic acid and salts 293 a-Phenoxy-)3-chlorcrotonolactone 295 o-Phenoxy-^-chlor-y-oxycrotonic acid and salts 296 Phenyl-diamido-phosphate 126 Phenyl-dichlor-phosphate, action of dry ammonia on 127 i-Phenyl-3-4-dimethyl-s-pyrazolone-2(?)-sulphonic acid 439 Phenylimidocarbonic diethyl ester . 300 " dimethyl ester 392 Phenylimidochlorformic ethyl ester 388 Index. 665 Phenylimidochlorformic methyl ester 391 " phenyl ester 399 i-Phenyl-3-inethyl-4benzyl-5-pyrazolone 442 " -2(?)-sulphonic acid 440 i-Phenyl-3-methyl-s-ethoxypyrazol 436 i-Phenyl-3-methyl-s-pyrazolon, electrolysis of 584 Phillips, F. C. Researches upon the phenomena of oxidation and the chemical properties of gases 163,255,340,406 Phospho-hydrocyanic acid 229 Phosphorus pentachloride, action on urethanes 70 Photochemical actions in solutions 557 Phthaleins of ortho-sulpho-para-ioluic acid 513 Phthalimide, electrolysis of 587 Physical chemistry, recent progress in 470, 551 " Physikalisch-chemische Tabellen "(//. Z.a«rf<7// und ^. >5(»«Ji^i«), review . . 564 Pinene derivatives, ketones from 404 Pine wood, pentosans in 611 Plants, pentosans in 218, 589 Polarization, velocity of in closed concentration cells . 56a Potassic dibromdinitrophenylate 34 Potassium .md antimony, mixed double halides of 490 " benzenesulphon periodide 118 '' hydrazoate, preparation 82 " hydronitride, preparation 82 " " separation of thorium from the rare earths of the cerium and yttrium group by means of 79 " iodide, on the action of benzenesulphonic acid on. A new class of organic. periodides 116 " sulphate and magnesium chloride, saturated solutions of ... . 5=6 " tartrate, acid, solubility in the presence of other salts .... 558 " " " the splitting ofi of hj'drogen from ...... 558 , " tri-iodide, dissociation in aqueous solution 554 " Principles and Practice of Agricultural Analysis "(//. W. Jf'zVrv), review . . 156 Propane, phenomena of oxidation 174 " qualitative reactions of , . 264 Propionic aldehyde, the polymeric modifications of, parapropionic and metapropi- onic aldehydes .............. 645 Propylene, phenomena of oxidation .......... 179 " qualitative reactions for 268 Pyrazol and indol derivatives 430 " derivatives .............. 436 Pyrazolone derivatives ............. 439 Pyrocatechin, action on animals 444 Pyrocatechin-sulphon-phthalein 518 Pyrogallol, action upon ortho-sulpho-benzoic acid 526 Pyruvic paratoluide .............. 384 " " action of phenylhydrazine on 385 " " phenylhydrazone 386 "QUALITATIVE Analysis for Use in Instruction in Chemical Laboratories '• (/.,. Medicus, translated by J . Marshall), review 653 " Qualitative Chemical Analysis " (/I. .rl. A'elofsen,y.A. Solubility of cream of tartar in alcohol of various strengths and at various temperatures . . . . , . . . . 464 " " The iodine absorption of some of the rarer fatty oils . . . 467 Roloff, M. Photochemical actions in solutions . . . ... . . . 557 Rothrock, I. R. See Dains, F. B. SALTS, hydrolysis of 313.331 Salts in solution, color of 326 Saponification velocity of some esters .......... 559 Sea-weeds, gases from 427 Seeds, pentosans in 589 " Select Methods in Chemical Analysis" (Jf. Crookes),xcy\e.\f 650 Shober, VV. B., and Spanutius , F. W. On phospho-hydrocyanic acid . . . 229 Smith, A. W. See Mabery, C. F. Smith, A., and Ransom, y. //. On two stereoisomeric hydrazones of benzoin . . 108 Sodium acetate, dry, and water_ action on the freezing-point of acetic acid . . 17 " and mercuric chlorides, an isothermal curve of the solubility of, in acetic ether 214 " benzenesulphon-periodide 120 " ethylate, action of diazobenzene nitrate on 247 " methylate, action of diazobenzene nitrate on 244 Soils, pentosans in 229 Solid solutions, influence of the chemical constituiion of organic compounds on the capacity to form 560 " " molecular weight estimations in 553 Solution, bodies in, general theorem upon the state of 557 Solutions, thermal expansion and compressibility of . 471 Spanutius, F. IV. See Shober, IV. B. Specific and molecular refraction, a new formula for 625 Stereoisomeric hydrazones of benzoin, . 108 Stieglitz, y. See Ungfeld, F. Stokes, H. N. On diamidoorthophosphoric and diamidotrihydroxylphosphoric acids 123 " " Note on monamidophosphoric acid ....... 154 Succinimide, electrolysis of 586 Sulphon-phthaleins, investigations on 513 Sulphur compounds, gaseous, qualitative reactions for ...... 344 " " in Ohio petroleum 83 " determinationof in volatile organic compounds 544 Sulphuric acid, colfibination with water in presence of acetic acid .... i " " hydrates of i Sulphur-petroleum, Canadian, examination ......... 89 Sugar cane, constituents of the nodes and internodes of . . . . . . 457 "TABELLARISCHE Uebersicht der Naphthalinderivate" (F. Reverdin und H. Fuldci), review .............. 233 Tatnmann, G. The thermal expansion and compressibility of solutions . . . 471 " " and Hirschberg, W. The thermal expansion of some solutions in alcohol, ether, benzene and carbon bisulphide 559 Tartar, cream of, solubility in alcohol 464 Tartaric and other organic acids, solubility of metallic oxides in normal potassium salts of 94 A*-Tetrahydroaniline chloroplatinate 453 Tetrahydroaniline nitrite 454 Index. 667 ^-Tetrahydronapluhylamine nit/ite (alicylic) ' 455 Tetraldehyde 60 " Theory of Heat " (y. /'r«'-^5*%^f^ 4.^- H- ^ , ' v^:^3^.it-« Ui'X^. V^ .^v^4Ki'^*^'A^?^»■: