rapidly at 20°, and at about 70° the double salt splits up into trisodium phosphate and tristrontium phosphate. When strontium chloride solution is poured into a solution of trisodium phosphate at about 10°, a gelatinous precipitate of tristrontium phosphate is first formed with absorption of heat, but this rapidly changes into crystalline sodium strontium phosphate with development of heat. The rapidity with which crystallisation takes place is doubtless due to the presence in the gelatinous precipitate of a small quantity of colloidal sodium strontium phosphate, which rapidly becomes crystalline, and thus initiates the transformation of the entire mass. Sodium strontium arsenate (loc. cit.) is more stable than the corresponding phosphate; heat of formation in the crystalline condition, +50.2 Cal. The author was unable to obtain any evidence of the existence of a double sodium calcium phosphate between 10° and 18°. The corresponding crystalline double phosphate and double arsenate of barium can, however, be readily obtained in a similar manner; their heats of formation are respectively +508 Cal. and +504 Cal. Sodium barium phosphate is less stable than the strontium compound. It is more readily decomposed by water, and the tendency to form barium triphosphate is greater. The latter compound seems to exist in a peculiar molecular condition. The phenomena observed in the case of sodium strontium and sodium barium phosphates are similar to those previously observed with ammonium magnesium phosphate. C. H. B. Heat of Combustion of Organic Compounds. By F. STOHMANN (Ber., 20, 2063-2066).—A reply to Thomsen (this vol., p. 761), in which the author discusses the difference between his results and those of Berthelot and Vieille, and points out that the differences are in the majority of cases less than is indicated by Thomsen. W. P. W. Heat Equivalents of Benzoyl-compounds. By F. SтOHMANN, P. RODATZ, and W. HERZBERG (J. pr. Chem. [2], 36, 1—16).—In the following determinations, the substances were burned in free oxygen; the substance was put into a small lamp, which in the case of compounds which melt with difficulty was provided with platinum wires, to conduct the heat from the wick to the substance. Several experiments made with each substance gave the following mean results:: The difference in heat equivalents of benzene (J. pr. Chem. [2], 33, 257) and benzoic acid is 9063 cal. This number is also obtained by adding together the differences in the heat equivalents shown in the conversion of toluene to benzyl alcohol, benzyl alcohol to benzaldehyde, and benzaldehyde to benzoic acid, and subtracting from the whole the heat equivalent of toluene less that of benzene. The conversion of benzoic acid and the formation of benzoic ether are enthodermic processes; the mean difference in the case of the alkyl ethers (obtained by comparison of the heat equivalents of benzoic acid and alcohol on the one hand, and the heat equivalent of the salt on the other) is -4740 cal. In the case of the aromatic benzoates the mean difference is -10833 cal. (compare this vol., p. 427). N. H. M. Relation between the Boiling Points of the Monatomic Alcohols and their Constitution. By F. FLAWITZKY (Ber., 20, 1948-1955). The relation between boiling points and the constitution of monhydric alcohols is discussed, and several tables given. The author considers that although it is not yet possible to show a general dependence of the boiling points on the constitution of alcohols, it is possible to predict approximately a boiling point when the constitution is known. N. H. M. Alteration of the Freezing Point. By F. KOLÁČEK (Ann. Phys. Chem., 31, 526–536).—In these remarks on Helmholtz' paper on this subject, the author defines the freezing point of a salt solution as the temperature at which the ice and salt solution have the same vapourtension, and points out that this definition was not put forward as a shrewd conjecture, as Helmholtz supposes, but was the result of general reasoning. The formula given for the freezing point is extended to the case when the heat of solution comes into play. From thermodynamic considerations he finds that where T, T, are arbitrary temperatures, pw, p the vapour-tensions of water and the salt solution, s the amount of salt dissolved. If s be small, we see that log P, and therefore P Pw P, is inde pendent of T, a law which has already been discovered by Wüllner. t be the freezing point, since (Ann. Phys. Chem., 29, log Pe we find the equation 0·00965t(1 +t 0·000531) = [log] + best course would be to take T = Since V is an unknown function, it might be supposed that the T, when the term under the integral sign vanishes, but the values of log p/p below zero are unknown. The next best thing to do is, make this term a minimum. The present research gives : R, a, b are known constants, and dt dp v may be calculated from the formula p + a/v2 = RT/(v b). The ratio cp/c is known, and cp has been found by Wiedemann. Ón substitution, turns out to be very nearly constant. Thus integrating the equation f c/v2, dt 0 we find ƒ = F(t)/v2 + tp(v). Clansius chooses (v) = 0, F(t) = 1/t, Van der Waals (v) = 0, F(t) = 1. Both choose the form of f, so as to satisfy the above equation. The present experiments, conducted at one temperature, afford no information about the form of F(t). We are therefore unable to discriminate between the hypotheses of Clausius and Van der Waals on this point, though the formula of Clausius gives results most in accordance with the numbers obtained from the experiment, unless the value of a be altered. C. S. Cooling of Carbonic Anhydride on Expansion. By E. NATANSON (Ann. Phys. Chem. [2], 31, 502-526).-The author believes that fresh measurements of the internal work of an expanding gas may supply considerable information about the forces which act between the molecules. The method adopted was the same in principle as that used by Thomson and Joule. The gas was allowed to expand through a porous plug of cotton-wool. Greater care was taken to secure the purity of the gas. This is most essential, for Thomson has pointed out that mixtures of gases behave in a very anomalous manner. Carbonic anhydride was chosen as the subject of experiment, because the intermolecular forces are large, and the many previous researches supply the requisite data for the calculations. The gas contained in a strong iron bottle escaped into a sphere (used for the purpose of equalising the pressure), and passed through a series of drying tubes filled with calcium chloride, into a number of copper tubes immersed in a water-bath of very considerable capacity. To increase the effective surface of the tubes, they were filled with metal turnings. Thus the gas on entering the nozzle containing the porous plug was at a uniform temperature equal to that of the waterbath. The nozzle was of exactly the same construction as the one employed in the experiments of Thomson and Joule. By means of stopcocks placed between the sphere and the plug and at the extremity of the nozzle, the pressure on each side of the plug could be completely controlled. Fifty-three experiments were made at pressures ranging from 2 to 25 atmospheres, with the gas at a mean temperature of 20° C. If AT, Ap be the differences of temperature and pressure on the two sides of the plug, the results of the experiments show that at a temperature of 20° C., p being the mean of the pressures on the two sides of the plug, measured in atmospheres. Taking the equation of elasticity to be p = Rt/(vb) — f(v, t), an attempt is made to determine the form of the function f. Applying the laws of thermodynamics to the present case of a gas expanding through a porous plug under small differences of pressure, we find which, in virtue of the equation of elasticity, may be reduced to Since it is our object to calculate merely the numerical value of the right hand side of this last equation, we may use Van der Waals' formula, p = Rt/(v b) a/v2, which is known to give the compressibility correctly. Expressing by this means in terms of dp dv we arrive, on substitution, at a result of the form dt dp Remembering that t is constant with respect to T, we see that when y is a minimum, s a maximum, and vice versa. Supposing y positive, the required values of T, may be found by plotting the curve ordinate and abscissa Ti, r, and selecting by inspection the least maximum value of x. Then neglecting y, we obtain from the last equation a near approximation to the value of t, the temperature of the freezing point. A table of the values of this quantity for some 20 substances is given. The greatest difference between theory and observation is about 17 per cent., a great improvement on the values given by Helmholtz, for which the discordance is as much as 40 per cent. in several cases. C. S. Phosphonium Chloride. By S. SKINNER (Proc. Roy. Soc., 1887, 283-289). In order to determine the relations of phosphonium chloride to temperature, volume and pressure, the author compresses equal volumes of phosphine and hydrogen chloride in a Cailletet's apparatus. The critical point was found at 48° under 95 atmos. The maximum vapour-pressure line lies below those of hydrogen chloride and phosphine at all temperatures. From -30° to 10° it is normal; above this combination begins. At temperatures near the critical points the volume of liquid phosphonium chloride produced is nearly one-half of that of the constituents (liquid). H. K. T. Apparatus for Determining Vapour-densities. By C. SCHALL (Ber., 20, 1827-1830).-A modification of the apparatus lately described by the author (this vol., p. 695). Determination of the Vapour-density of High-boiling Substances under Diminished Pressure. By C. SCHALL (Ber., 20, 2127-2129). An application of the form of manometer employed by Meier and Crafts (Ber., 13, 851), and by Nilson and Pettersson (ibid., 17, 987) to the apparatus described by the author (preceding Abstract). Results obtained by its use are given in the paper. W. P. W. Influence of Temperature on the Rapidity of the Action of certain Mineral Acids on Marble. By W. SPRING (Bull. Soc. Chim., 47, 927-933).-The rapidity of action of hydrochloric, hydrobromic, hydriodic, nitric and perchloric acids on marble increases with the temperature, it being about doubled for a difference of every 20° between 15° and 55°. When an aqueous solution of one of the above acids is allowed to act on marble until action ceases, the author finds that the rapidity of action decreases uniformly with the strength of the acid until threefifths of the acid has been neutralised; after this point, the decrease of the rapidity of action is rather less than the decrease in the amount of acid present; this bears out the observation of Ostwald, that the rapidity of the action of monobasic acids is increased by_the presence of their salts. A. P. Application of the Electrometer to the Study of Chemical Reactions. By E. BoUTY (Compt. rend., 104, 1789-1791 and 1839— 1841). An application of the measurement of resistances to determine |