61 General and Physical Chemistry. Electrical Conductivity of Alkali Chlorides and Nitrates. By FRIEDRICH KOHLRAUSCH and MARGARET E. MALTBY (Chern. Centr., 1899, ii, 465-466 ; from Xitx. Kgl. Aknd. TViss. Berlin, 1899, 665),- With the view of testing the theoretical connection between molecular conductivity and ionic migration, the conductivities in dilute solution of the chlorides and nitrates of potassium, sodium, and lithium have been determined with the greatest care. The numbers in the follow- ing table give the equivalent conductivities a t 18" for solutions of the concentration rn, and represent the mean of several determinations, which agreed on the average to 1/4000th. m. 0~0001 0~0002 0.0005 0.001 0.002 0-005 0-01 0.02 0.05 0.1 0.2 0.5 1.0 KC1. 129.05 128.76 128.09 127.33 126.29 124.40 122.42 120.00 115.94 11 2.00 107.96 102.40 98.28 NaCl. 108.06 107-80 101.18 106.48 105.55 103.79 101.95 99.66 95.86 92.01 87.73 80.93 74.34 LiC1.98.06 9 7-78 97.13 96-45 95.55 93.86 92.08 89.88 86.22 82-35 77.9 3 70.65 63-30 KNO,. 125.49 125.18 124.44 123.64 122.59 120.47 1 18-20 115.27 11 0.09 104.7'7 98.74 89.23 S0.47 NaNO,. 204.53 106.18 103.53 102.85 101.88 100.07 93-16 95.70 91.60 87.24 82.28 74.05 65-86 LiNO,. 94.38 94.07 93.45 92.80 91.91 90.27 88.55 86.38 82.86 79.24 75.06 68.03 60.80 For solations up to and including the concentration N/500, each ion has a specific conductivity, which depends on the concentration alone; up to N i l 0 solution, the deviations from t h i s rule amount to +, 1 per cent. As a rule, the fall of ionic conductivity with increasing concentration is most rapid in the case of ions with high migration velocity, but no general and simple law can be laid down.Within the same limits, the fall of conductivity is, in the cases of potassium and sodium, more rapid for the nitrate than for the chloride; in the case of lithium, more rapid for the chloride than for the nitrate. J. C. P. Change of Free Energy in Fused Halogen Compounds of some Heavy Metals. By RICHARD LORENZ (Zeit. aizorg. Chenz., 1899, 22, 241-255. Compare Abstr., 1899, ii, 269).-In determining the E.M.F. of polarisation for a number of fused salts (Abstr., 1899, ii, 267), Czepinski found the temperature coefficient abnormally large at higher temperatures. Further investigation by Weber (Abstr., 1899, ii, 724) has shown that this coefficient d E / d T is regular, even up to high temperatures, when the electrodes are kept far enough apart ; the necessity for this is due to the diffusing and penetrating power of the halogen and the metallic vapour. VOL.LxxvIiI. ii, 562 ABSTRACTS OF CHEMICAL PAPERS. Differentiation of the equation E= Q + T.dE/dT leads to the relation - dQ/d17= 5?.d2E/dT2. Thus, when dE/dT is constant, Q must also be constant. This relation is approximately fulfilled in the case of lead chloride, so that the molecular heat of the compound is very nearly the sum of the atomic heats of the components. On the assumption that this relation holds exactly, the E.M.F. may be extrapolated for ordinary temperatures. The value thus calculated for the E.M.F.of the system P b 1 PbCl, 1 C1, is 1.65 volts, agreeing with the value 1.59 volts for the E.M.F. of decomposition of lead chloride in solution, calculated by Bodlander from its solubility. The paper also discusses the ionic concentrations in fused salts, leading to the conclusion that the dissociation of fused silver chloride is rather more than twice that of fused lead chloride. J. C. P. Coagulative Power of Electrolytes, By W. C. DAMPIER WHETHAM (Phil. h?C6g., 1899, 48, 474--477).-The coagulative power of a n electrolyte may be taken as inversely proportional to the number of gram-equivalents necessary to produce immediate coagula- tion in a given solution of a colloid. The coagulative powers of sulphates with uni-, bi-, and ter-valent ions stand in the ratio 1 : 35 : 1023 (compare Trans., 1895, 67, 6 3 ) , those of chlorides in the ratio 1 : 30 : 1650.There is certainly an intimate connection between the coagulative power and the electrical properties of a solution, and it may be supposed that to produce coagulation a certain minimurii electrical charge is necessary. Further, as the electrical charge on an ion is proportional to its valency, equal charges are carried by 293 triads, 3n dyads, or 6% monads. On this basis, i t is shown that the coagulative powers of uni-, bi-, and ter-valent ions in equivalent solution are in the ratio 1 : x : x2, agreeing well with the numbers just quoted. If this view is correct, the coagulative power of a quadri- valent metallic sulphate would be about 30,000 himes as great as that of a sulphate with a univalent ion.J. C. P. Application of the Dissociation Theory to the Electrolysis of Aqueous Solutions of Two Electrolytes with One Common Ion. By JAMES G. MACGREGOR (Chem. Centr., 1899, ii, 8 2 ; from 171-ccn.s. Roy. SOC. Canada, [ii], 4, 117--14S).--The author has cal- culated the values of the dissociation in solutions of mixed electrolytes with one common ion, and obtained from these values the ratio of the transference numbers of the other two ions, the values so resulting being compared with those obtained by Hopfgartner (Abstr., 1898, ii? 151). The agreement was satisfactory in the cases of solutions of sodium and hydrogen chlorides, potassium chloride and iodide, and barium and hydrogen chlorides. Agreement was not obtained in the case of solutions of copper and hydrogen sulphates, but this is explic- able by the dissociation of the acid into H and HSO, ions as well as into H2 and SO”, ions.Theory of the Electrolytic Solution Pressure. By R. A. LEHFELDT (Phil. Mag., 1899, 48, 430- 433).---The electrolytic solu- tion pressure, II, of a met31 is calculated from the observed E.M.F. L. M. J.GENERAL AND PHYSiCAL C€iEMISTRY. 63 between the metal and the solution in which it is immersed by the equation E= RT/E.log,II/P, where E is the quantity of electricity associated with 1 gram-equivalent, and P is the osmotic pressure of the metallic ions in the solution. For zinc, II has the very high value 9.9 x 10ls atmospheres. An expression is obtained for the tension due to the electrical double layer between the metal and the ions in solution, leading t o the equation, x = JJDII/2rc2, where x is the number of gram-equivalents that go into solution per sq.cm. of surface, and D the dielectric constant of the medium, From this equation, it may be shown that in order to produce the solution pressure attributed to zinc, 1.27 grams of the metal would have to pass into the ionic form per sq. cm. immersed, a conclusion which is not borne out by observation. Besides this difficulty with regard t o the electrolytic solution pressure of zinc, there is one of another kind i n the case of palladium, for which 11= 1.5 x ; the smallness of khis value involves the rejection of the molecular theory of fluids. J. C. P. Thermal Conductivity in Gases. By M. SMOLUCHOWSEI R.VON SMOLAN (Chem. Centr., 1899, ii, 353, from Oesterr. Chem, Zeit., 2, 385). -It has been previously shown that, as foretold by Maxwell, the thermal conductivity of a gas is independent of pressure. The author shows, however, that this does not hold for very low pressures, and in this case a temperature difference occurs a t the surface of gas and solid. When the mean wave-length is greater than the dimensions of the containing vessel, the phenomena are very complicated, Brush’s assumption of the existence of etherion is not justified, and the author considers the supposed element to be merely aqueous vapour (compare Abstr., 1899, ii, 287). Thermal Capacity and Colour Changes of Solutions of Cobalt Chloride. By M. WREWSKY (Chem. Centr., 1899, i, 1202; from J.Russ. Chem. Xoc., 1899, 31, 164--171).-0wing to difficulties in calorimetry a t high temperatures, the changes in colour were brought about by the addition of alcohol. The change from a blue to a red solution is accompanied by an increase in the difference between the thermal capacities of solution and solvent. The author considers the results support Berthelot’s hydrate theory. The decrease of thermal capacity by dissolution of salts is a function of both concen- tration and temperature, increase of both acting in the same sense. L. 31. J. L. M. J. NOTE BY ABSTRACTOR.-with regard to the comparison of the heat capacities of a solution and its components, see Tammann, Abstr., 1896, ii, 289. Thermal Capacity of Solutions of Sulphuric Acid. By EUGEN YON BIRON (Chew.Cerztr., 1899, i, 1202-1204; from %ss. Chem. Xoc.. 1899, 31, 171-203).-The author has determined the mean specific heat between 18.6 and 21.8 for solutions of sulphuric acid of the composition H,SO,,nH,O, where rz varies from 0 to 1600, the extreme values for these liquids being 0.3352 and 0.99675. The molecular heat for the monohydrate is 51.17, that calculated from its 5-264 ABSTRACTS OF CHEMICAL PAPERS. components being 50.9, and the difference is ascribed to the heat of dissociation of the hydrate. A t dilutions greater than H,SO,, 100H,O, a linear relationship is found to exist between the equivalent electrical conductivity and the difference between the found and calculated values for the molecular heat, so that it appears probable that both owe their origin to the same cause (compare Tammano, Abstr., 1896, ii, 289).Some Boiling Point Curves. 11. By JOHN K. HAYWOOD (J. Amer. Chern. sbc., 1899, 21, 994--1001).-A continuation of the author's previous paper (Abstr., 1899, ii, 632), the mixtures examined being benzene with chloroform, carbon tetrachloride, ether, acetone, and methyl alcohol ; and methyl alcohol with ethyl alcohol, carbon tetrachloride, and ether. Minima were obtained in the follow- ing cases a t the temperatures and composition given : benzene with methyl alcohol, 58*33 ; 53 to 67 per cent, of benzene ; methyl alcohol with carbon tetrachloride, 55*95', about 18.4 per cent. of alcohol. The law previously suggested that chemically similar substances yield similar boiling point curves again receives support, whilst it also ap- pears that similarity of constitution of the two mixed compounds is unfavourable to the production of a minimum even when the boiling points are close together.L, M. J. By GARNETT RYLAND (Arne?.. Chern. J., 1899, 22,384-396).-.A mixture of methyl alcohol and benzene containing 38 per cent. of alcohol distils at the constant temperature 57-57.5' ; similarly, a mixture of ethyl alcohol and benzene with 32 per cent. of alcoholj distils unchanged a t 67-68', The composition of the mixture with constant boiling point is in both cases dependent on the pressure. Numerous binary mixtures of organic liquids have been examined with the view of discovering mixtures of constant boiling point. Forty-five have been found to boil at a constant temperature a t or below the boiling point of the more volatile constituent, two above the boiling point of the less volatile constituent, and one between the boiling points of the constituents.For these mixtures of constant boiling point, the ratio of the two constituents in the distillate is approximately that of the products of their vapour density and vnpour tension at the temperature of distillation, a result in accordance with earlier researches on the subject. By H. W. BAKHUTS ROOZEBOOM (Proc. K . Akad. Wetensch. Amsteydccm, 1899, 1, 466-468). -Melting point curves for mixtures of the racemic and dextro-compounds were examined in the case of dimethyl tartrate and dimethyl diacetyltartrate. I n the first case, the racemic compound melts at 89.4' and the active compound a t 43*3", and a minimum of 41.6 is reached a t 3 per cent.of racemic compound ; in the second case, the melting point of the racemic compound is 83%" and that of the active compound 104.3'; the minimum of 83.4" is reached at about 86 per cent of the racemic compound. I n both cases, the form of the curve is almost horizontal at the racemic compound, and indicates L. M. J. Liquid Mixtures of Constant Boiling Point. J. C. P. Melting Points in Systems of Optical Isomerides.GENERAL AND PHYSICAL CHEMISTRY. 65 considerable dissociation. The results confirm the author’s previously published theoretical views (Abstr., 1899, ii, 401). By ALBERT DAIIMS (An72. Chim. Phys., 1899, [vii], 18, 140-142. Com- pare Abstr., 1897, ii, 245 ; 1899, ii, 546).-Coppet’s determinations of the freezing points of dilute acetic acid agree closely with the values previously obtained by the author, who has also anticipated the theoretical conclusions deduced by this investigator.A Reply. By LOUIS C. DE COPPET (Ann. China. Phys., 1899, [vii], 18, 142--144).-The author, whilst admitting that Dahms first determined the freezing point of the eutectic mixture of acetic acidand water, claims to have discovered the lower freezing point of a superfused mixture at a temperature about Z0 lower than the eutectic point. The author’s conclusions were originally published over 2 7 years ago. Since binary liquid mixtures having two eutectic points and three freezing points are known, it is highly probable that similar mixtures having three or more eutectic points should be capable of existence.G. T. M. L. M. J. F r e e z i n g Point of Mixtures of Acetic Acid and Water. G. T. M. Freezing Point of Mixtures of Acetic Acid and Water. Test by Freezing Point Determinations of the Dissociation Values obtained by the Conductivity Method in the case of Solutions of Potassium and Sodium Sulphates. By E. H. ARCHIBALD (Chern. Centr., 1899, ii, 7, and Chem. News, 1899, 80, 46, et sep. ; from Trans. Nova Scotia Inst., 10, 33).-Determinations were made of the electrical conductivity of solutions of each of the com- pounds at various dilutions at 1 8 O and at 0”. I n the case of potassium sulphate solutions a t concentrations above 0.35, the dissociation is slightly greater at Oo than a t 1 8 O . The cryoscopic depressions of the solutions were also determined by the method adopted by Loomis, with whose results also the values are in good accord.Comparison of these depressions with those calculated from the dissociation values show the agreement to be close, the differences never exceeding 4 per cent., being, however, for the sodium chloride always in the same directions. For mixtures of equal volumes of solutions containing molecular pro- portions of the two salts, the depression is given by the expression D = 1 sS6, (1 + a1 + a,)iV/2, where N is the number of gram-equivalents per litre of each salt in the solutions mixed, and ala, the dissociation (this expression appears t o involve the assumption that the salts are equally dissociated, which is, however, almost the case for the salts examined], and the depressions so calculated are found to agree well with those determined experimentally.Cryoscopic Behaviour of Substances with Constitutions similar to that of the Solvent. By FELICE GARELLI and F. CALZOLARI (Gaxzetta, 1899, 29, ii, 357-375. Compare Abstr., 1899, ii, 732).-The cryoscopic behaviour of various substances has been studied in solvents, from which they are derived by the substitution of an amino- or a hydroxyl group for a hydrogen atom. I n the following cases, abnormal molecular weights were obtained, indicating the forma- tion of solid solutions between the solvent and the solute; p-hydroxy- L. M. J. V.66 AnSTRACTS OF CHEMICAL PAPERS. azobenzene and p-aminonzobenzene dissolved in azobenzene ; m-nitro- and p-nitro-phenol and o-nitro-, m-nitro-, and p-nitro-aniline in nitro- benzene ; 2 : 4-dinitroaniline in nL-dinitrobenzene ; p-xylidine in p-xylene ; p-hydroxyacetophenone in acetophenone ; triphenylcarbinol in triphenylmethane ; and glycollic acid in acetic acid.With solu- tions of o-nitrophenol in nitrobenzene, 2 : 4-dinitrophenol in m-di- nitrobenzene, and p-aminoacetophenone in acetophenone, normal mole- cular weights are obtained. Influence of the Solvent on the Cryoscopic Behaviour of Phenols. By KARL AUWERS [with W. BARTSCH and H. 3X. SMITH] (Zed physikal. Chenz., 1899, 30, 300--340).-If the abnormal cryo- scopic depressions of hydroxy-compounds are due to association, then an association constant should be obtainable by a calculation similar to that used for obtaining the dissociatioiz constant, Experiments with acetoxime, benzophenoxime, Z-camphoroxime, benzoic acid, and o-bromo- benzoic acid, however, show that, except in the last case, no constant value for the association constant is obtained, a t least not on the assumption of double molecules. Determinations with other com- pounds to high concentrations further show that the molecular weight does not, in general, tend to any constant value; the greatest valueof association hitherto observed is about four, whilst cryoscopic depres- sions of ethyl alcohol in benzene give values leading to seven times the normal molecular weight.It appears, therefore, that the abnor- mal depressions are not due to association, but to some mutual influence of solvent and solute, and determinations were made of the depressions of a number of hydroxy-compounds in various solvents.The following mere employed as solvents, and the cryoscopic constant obtained for each is added :-Nitrobenzene, 70 ; nz-dinitrobenzene, 106 ; p-nitrotoluene, 78 ; 2 : 4-dinitrotoluene, 89 ; 8 : 4 : 6-trinitrotoluene, 115 ; p-chloronitrobenzene, 109 ; p-dichlorobenzene, 77 ; p-chloro- bromobenzene, 92 ; T-dibromobenzene, 124 ; benzil, 105. I n all these solvents, ortho-substitution derivatives of phenol were cryoscopically normal as in benzene or naphthalene ; with para-compounds, however, this is not the case, p-nitrophenol, methyl p-hydroxybenzoate, and p-hydroxybenzaldehyde being abnormal in the halogen conipounds, but almost normal in those solvents containing a nitro-group, so that the halogen increases and the nitro-group decreases the abnormalising influence in the solvent.Other substituted phenols dissolved in p-dic hloro-, p-chlorobromo-, and p-dibromo-benzene were then investi- gated, and it was found that, as in naphthalene, para-compounds are more abnormal than meta-, and the order of the groups arranged according to their influence in giving abnormal values is : aldehyde, cyanide, substituted carboxyl, nitro-groups, whilst the abnormality is greatest in dichlorobenzene, and least in dibromobenzene. The ab- normal cryoscopic depressions are hence the product of two factors, one dependent on the constitution of the solute, the other on that of the solvent (compare Abstr., 1896, ii, 293 ; 1897, ii, 476). Influence of the Medium on the Heats of Solution of Salts.By N. GALITZEI (Chem. Centr., 1899, ii, 469-470 ; from J. Russ. Chem. Soc., 31, 536--5PO).-The effect of increasing quantities of alcohol on T. H. P. L. M. J.GENERAL AND PHYSICAL CHEMISTRY. 67 the heats oE solution of potassium nitrate and potassium carbonate in water has been investigated. By the addition of alcohol to water, the heat of solution is diminished, the diminution ultimately reaching a minimum, which depends on the nature of the dissolved salt. No connection could be traced between this phenomenon and the electrical conductivity of the solutions. By KONRAD DIETERICI (An7~. Phys. ClLem., 1899, 69, 6S5-’705).-For most substances hitherto investigated, the ratio of the actually observed critical density to the ideal density (the density which the substance would have if the law pv = 161’ held for the critical pressure and temperature) is nearly constant, and equal to 3-75,; for ethylene, nitrous oxide, nitrogen, and oxygen, the ratio differs considerably from 3.75, but this is probably due to deficient experimental methods.I n van der Waals’ equation, the cohesion pressure is eqlial to a/v2, and the correction for the molecular volume is a constant b ; in these circumstances, the ratio wJvk (ideal volume : critical volume) has the value 2.67. This want of correspondence between experiment and van der Waals’ equation remains, even when b is regarded as a function of v. If the equation ( p + T)(V - b) = BY’ be accepted, and the cohesion pressure rr=a/vs/3 instead of a/v2, the ratio vo/vk is equal t o 3.75.This purely empirical law regarding the cohesion pressure, although it brings van der Waals’ equation into agreement with the critical phenomena, is without theoretical basis. 0 t her theoretical considerations lead t o the equation p=RT/(u - b) . e-A/ziT, where A is a function of v, and represents the work to be done by a molecule against the forces of cohesion, O K ~ the supposition that A = C/w, the ratio vJvk is equal t o 3.695, a value agreeing closely with that given by experiment. By HEINRICH LEY (Zeit.physikaZ. Chew,., 1899,30,193-%7).-1n a salt solution, besides the electrolytic dissociation, hydrolysis may occur with the formation of acid and base which further undergo electrolytic dissociation. The concentrations of the various molecular groupings present in the solu- tion are determined by the equilibrium constants for the hydrolytic dissociation and for the electrolytic dissociation of salt, acid, base, and water.The author first investigates theoretically the necessary equations for the cases where (1) either base or acid is weak ; (2) both acid and base are weak. For the determination of the hydrolysis, the method of sugar inversion was found suitable, and the author obtains the value 16.8 for the inversion constant a t looo, which is thus intermediate between the values 17.9 found by Trevor (Abstr., 1893, ii, 62) and 16.0 found by Smith (Abstr., 1898, ii, 155). The variation of the constant with temperature was also deter- mined and found to be given by the expression K T ~ = K T ~ e.A(T1- To)lTITo, where A is a constant.Some salts were found t o cause an irregular increase of inversion, apparently not due to hydrolysis, as, for example, lanthanum chloride. Lithium and magnesium salts have no effect, but beryllium and aluminium cause considerable inversion, and for these the hydrolytic constant was calculated a t dilutions from 32 to J. C. P. The Critical State. J. C. P. Hydrolytic Dissociation in Salt Solutions.68 ABSTRACTS OF CHEXJCAL PAPERS, 512. The values do not remain constant, but the author considers the agreement to be satisfactory, owing to the complicated equations and the fact that they are rigorously valid for only binary electrolytes. From the values of the dissociation, the basicity of the hydroxides may be compared, and it is thus found that beryllium hydroxide is about eleven times as strong a base as aluminium hydroxide.Salts of cerium, copper, and zinc are only very slightly hydrolysed ; lead causes greater inversion, but still far less than the alumiuiurn salts. The hydrolysis of methyl acetate was also employed for the deter- minations, and the values for the hydrolysis of aluminium chloride by this process agree well with those obtained by the inversion method. Conductivity determinations may also serve for the calculation of hydrolysis ; the acetates of manganese, cobalt, zinc, and nickel exhibit a perfectly normal change of conductivity with dilution, and are hence not hydrolysed. With acetate of lanthanum, cerium, and lithium, this is not the case, whilst acetates of lead, beryllium, aluminium, and mercury give very abnormal results, indicating considerable hydrolysis.Where quantitative results were obtained, the values of percentage hydrolysis a t v = 1024 and 25' are : beryllium sulphste, 5 per cent. ; aluminium chloride, 4% per cent.; lead chloride, 4.4 per cent. ; uranium nitrate, 5.9 per cent. ; mercuric perchlorate (v= 512), 37 per cent. (Abstr., 1898, ii, 66). Determination of Solubility Coefficients of Liquids. By A. A~GNAN and E. DUGAS (Compt. rend., 1899, 129, 643-645).-When aniline and water are mixed, there is no alteration of total volume consequelit on the reciprocal solubility of the liquids, and in this case i t is possible to determine the solubilities by observation of the volumes of the two layers before and after agitation? in two different experi- ments.The values so obtained in C.C. per cubic centimetre are : (1) aniline in water, 0-036 ; (2) water in aniline, 0.042. I n the case of water and fermentation amyl alcohol, however, the total volume does not remain constant, and very diverse values for the solubilities, sometimes even negative, are obtained, so that in this case the author considers the case is not one of simple solubility, and the results may be due to the fact that one of thealcohols present in the fermentation amyl alcohol is capable of combining with water. By W. HERZ (Zeit. ccnorg. Chem., 1899, 22, 279-284).--The results previously obtained (Abstr., 1899, ii, 752) were not in harmony with the law of mass action, but it is now found that a t lower concentra- tions the equation K= [Mn**][NH,]~//CNH4*l2 gives fairly constant values of K : in this equation, the symbols in square brackets indicate the concentrations of the respective molecules or ions.With the mean value of K = 1.6 x and the dissociation constant of am- monia = 0*000023, the solubility of manganous hydroxide is calculated t o be 0.6 x a value of the same order as that obtained by Bodlander feom other considerations. Fusion of Sodium Thiosulphate. Hydrates. By FRIEDRTCH WILHELM KUSTER and A. THIEL (Zeit. anovg. Chem., 1899, 21, 401--404).--Fueed sodium thiosulphate might be regarded either as L. M. J. I;. M. J. - Equilibrium between Manganous Salts and Ammonia. J. C. P.GENERAT, ANT) PHYSICAT, CHEMISTRY.69 a definite liquid compound, or as a reciprocal solution of t h e decom- position products of the crystallised salt. I n the first case, excess of water or anhydrous thiosulphate in the fused hydrate would give two distinct series of solutions, the properties of which might be repre- sented by two curves, cutting each other at the point corresponding with the composition Na2S,O,,5H,O. If, on the other hand, the fused hydrate is merely a reciprocal solution of water and anhydrous salt, its properties ought to be intermediate between those of solutions with excess of water and anhydrous salt respectively. The latter behaviour has been observed in the case of the conductivity. As the number of water molecules to one molecule of anhydrous salt rises from 4.69 t o 6.65, the conductivity increases proportionally, and there is no dis- continuity whatever corresponding with the composition Na2S,0,,,5H20.J. C. P. Molecular State of Ammonia and of Amines in Aqueous Solutions. By ARTHUR HANTZSCH and F. SEIZALDT (Zeit. yl~ysikccl. Chem., 1899, 30, 25S-299).-1n the hope of solving this ques- tion, the author determined the partition ratio of ammonia and amines between water and other solvents. Ammonia in water and chloroform mas first examined ; if i t exists as hydroxide, the scheme of dissociation is NH,HOSNH, + H,OZNH’, + HO’, and as ammonia is but a weak base, thelntter dissociation is very slight. The calcula- tion of the partition ratio is made on the assumption that the hydr- oxide itself is insoluble in chloroform, and as the concentration of the gaseous ammonia in the aqueous phase is proportional to that of the hydroxide, the total conceutration may be, and is, employed instead of that of the gaseous ammonia only.The value for the partition ratio, water/chloroform a t 2 5 O , is about 25, and remains constant for varying concentrations, but i t increases with fall of temperature. The values were not affected by the addition of ammonium chloride, a result the authors consider t o be surprising, although, in view of the slight electrolytic dissociation of ammonium hydroxide, i t might have been expected. The partition of piperidine between benzene and water and the effect of addition of piperidine hydrochloride, were next examined; the addition of sodium hydroxide was found to cause con- siderable decrease in t h e partition ratio, which was also found t o decrease with increase of temperature.The partition at various temperatures was also determined €or trimethylamine and triethyl- amine between water and toluene ; acetic acid, pyridine, hydrogen cyanide, and cyanacetic acid between water and benzene; and for acetic acid, cyanacetic acid, ferric thiocyannte, and mercurous chloride between water and ether. With the amines, in every case the partition ratio decreases with rise of temperature ; this was also the case in a few other systems, but in most the converse obtained. This seems t o indicate the existence of hydroxides and a n increase of molecular dis- sociation at higher temperatures, but the authors consider it more probable that hydrates, and not hydroxides, exist in the solutions.In ammonia solutions, the temperature coefficient is very small, and here especially the quantity of hydroxide must be very sinall. L. M. J.7 0 ABSTRACTS OF CHEMICAL PAPERS. Conversion of Mixed Crystals into a, Compound. By H. W. BAKHUIS ROOZEBOOM ( PYOC. K. Akad. Tetensch. Amsterdam, 1899, 2, 74-77).-No difference is observable in the melting points of dextro- Iawo-, and inactive camphoroxime, or of mixtures in any proportions ; the solid mass is further perfectly homogeneous, sa that the existence of mixed crystals is confirmed (Forster and Pope, Trans., 1897,71,1049), and the melting point CUITB for the mixture is hence a horizontal lint., The transitidn point to monoclinic crystaIs varies from 112.6' for 100 per cent. of either compound to 109.4' for t h e inactive mixture, and the curve is quite symmetrical, and here also mixed crystals are again formed.By cooling still more, a further change takes place, and this occurs at 103" for the inactive mixture, but the t'emperature is con- siderably reduced by excess of either component, and could not be observed at all when t h e percentage rose above 75. I n this case, therefore, the mixed crvstals become converted into a comDound. and the views' of Pope (Zoc.'cit.) are confirmed (see Abstr., 1896, ii, 401). L. M. J. Mixed Crystals of Mercuric Iodide and Bromide. By W. REINDERS (Proc. K. Akccd. Vetensch. Amsterdam, 1899, 2, 146- 148).- The melting point curve of mixed crystals of mercuric iodide and bromide is continuous between the temperatures 236 5' and 255*4O, the melting points of the bromide and iodide respectively. It passes through a minimum at 216.1', corresponding with a mixture containing 59 per cent.of mols. of the bromide, and at this temperature the crystals deposited have the same composition as the molten mixture; if the latter contains more (or less) bromide than the above-named quantity, the crystals deposited are richer (or poorer) in bromide than the mixture. Below 216", mixed crystals in all proportions can exist. At l27', yellow mercuric iodide is converted into the red modi- fication. By admixture of the bromide, this transition point is lowered and also widened out into a transition-interval determined by two limiting curves, one for the yellow, and the other for the red, cryst#als. The former of these runs from 127' when no bromide is present t o a point on the Oo abscissa corresponding with 33 per cent.of mols. of the bromide; whilst t h a t for the red crystals proceeds from the same starting point and cuts the 0' abscissa in a point indicating the pres- ence of 8.6 per cent. of the bromide molecules. The transition interval was determined partly by observation of the colour change, as this made i t possible to determine the composition of red crystals which a t a definite temperature change completely into the yellow variety ; further, solutions were found by trial, each of which deposits at a particular temperature, both red and yellow mixed crystals. These curves could not be continued below,OO, but their direction shows that, if there is a transition point for mercuric bromide, it must be at a very low temperature; at - 83", no indications of such transition could be found.Solid mercuric bromide and iodide diffuse into one another at ordinary temperatures, and more so when heated, and the transition temperature of a finely-powdered mixture of them agrees very nearly with that of the mixed crystals of the same composition. T. H. P.INOltGANIC CHEMISTRY, 71 Properties of Flames. By NICOLA TECLU (J. pr. Clmn., 1899, [ii], go, 396-399).-I. A stream of coal gas is allowed to issue from a platinum tube about 12 mm. in length and 2 mm. in diameter, so as to afford a flame of definite dimensions, the pressure of the gas at the point where i t enters the tube being registered by means of an alcohol manometer.It is found that on heating the platinum tube by means of a flat flame bunsen burner, if the gas supply is taken directly from the main, the flame suffers a contraction of about 70 per cent., whilst the manometer registers only R small increase in pressure (about 40 mm.). If, on the other hand, the gas is supplied from a small laboratory gas-holder, the flame suffers no appreciable contraction when the tube is heated, but the manometer registers an increase in pressure nearly three times as large as in the former case (about 110 mm.). The explanation offered is that the disturbances, set up by the sudden increase in bulk which the gas undergoes in coming into contact with the heated tube, are propagated backwards with the velocity of compression waves, and suffer reflection in the small laboratory apparatus, producing pressure in the mass of gas and thus affecting the manometer and serving to maintain the size of the flame ; where the enormous technical gasometers are involved, the energy of the waves is minimised or dissipated. 11. An apparatus is described which consists of a modification of that used in the ordinary lecture experiment for demonstratiiig the reciprocal nature of combustion. By ARTHUR JOHN HOPICIKS (Amer. Chem. J., 1899, 22, 407-410).-1n this apparatus, a glass tube passing through a doubly-bored indiarubber stopper, fitted to a tall glass cylinder, is con- nected near the bottom of the latter with one limb of a Y-tube, the lower limb of which is open, whilst the third is connected to a vertical glass tube which reaches up nearly to the stopper, and is there curved downwards. Through the second hole in the stopper passes a tube connected with a filter pump which serves to draw a current of air through the apparatus, the rate of this being regulated by a stop- cock on the tube open to the air. The crystals to be dissolved are placed a t the bottom of the cylinder ; the saturated solution formed there is withdrawn continuously by the suction in the Y-tube and discharged upon the less saturated solvent a t the top, fresh solvent thus being brought continuously in contact with the substance to be dissolved. W. A. I?. A. L. A Dissolver.