indeed it is improbable that any hydrate would be insoluble in a liquid which dissolves both the anhydrous substance and water. The melting-points of the anhydrous and hydrated substance were 40°48 and 45°42 respectively: Linnemann gives 35°-38° and 46°.5. Anhydrous pinacone superfuses considerably, and in the case of the experiments Nos. 32 and 33, in which the liquid substance was taken at an initial temperature of 44°, it did not solidify till after it had attained the temperature of the calorimeter: the heat evolved when it did solidify was noted separately, and gave the two supplementary measurements entered in the footnote to Table II. Octohydrate of Tin Tetrabromide.-The tetrachloride of tin forms a tetrahydrate; but as the anhydrous substance was found not to solidify till -34°15, it was not a suitable substance for the present investigation. The tetrabromide, which solidifies at the higher temperature of 29°-36, was Fig. 1.-Freezing-points of SnBr, solutions. therefore examined in order to ascertain whether it also formed a hydrate. The results of a series of freezing-points of mixtures of it with water are given in Table IV. and fig. 1. 100 The freezing-point of the tetrabromide is no doubt lowered by the addition of water, and probably very rapidly so, for with 2 per cent. of added water no crystallization was obtained at -70°; but with this amount of water, and also with amounts up to 24 per cent., mutual solution is not complete, the liquids being always cloudy. When the water is increased to about 26 per cent., the liquid is clear at high temperatures, and generally becomes cloudy on cooling (it did not do so in the experiment at 73.4 p. c. SnBr1); and the more the added water is increased the lower is the temperature which the solution will stand without becoming cloudy. Hard, well-defined, transparent crystals were obtained from solutions containing 73.6 to 66.7 per cent. of the tetrabromide; but with the strongest of these solutions the liquid was turbid when solidification occurred, and the observations were difficult and doubtful. From solutions containing less than 56 per cent. of the bromide, water crystallizes. On plotting out the results from 73.6 to 66.7 per cent. they form the intermediate curve shown in fig. 1, which on a more open scale is found to indicate a maximum at a strength of 75'0 per cent. of the bromide, and at a temperature of 19°; that is, somewhat beyond the point at which the determinations become impossible owing to the cloudiness and dissociation of the solution. This indicated that the hydrate was probably an octohydrate, which contains 75.261 per cent. To establish this more satifactorily, a quantity of a 73-per-cent. solution was allowed to deposit a few crystals, and these after being drained were analysed and gave the values Sn, 20-662, Br, 35-011, H2O (by difference), 24-327, theory for SnBr, 8H2O requiring Sn, 20-334, Br, 34.926, H2O, 24.739. This octohydrate, as may be inferred from what has been said above, cannot be melted without becoming decomposed and cloudy; but if heated to a higher temperature till clear, and then cooled without stirring, it may generally be cooled to atmospheric temperatures and crystallized without decomposing. By repeatedly crystallizing in this way a sample was prepared for the determinations. The crystals themselves, however, on being kept for any length of time, or on being scratched, generally become cloudy and dissociated. The heat-capacity and heat of fusion determinations with this substance could not be made in the ordinary manner, as it would have been impossible to see whether, on cooling in the platinum bottle, it had decomposed or not. Fortunately it easily superfuses, and by taking advantage of this property it was found possible to determine its heat of dissolution both in the solid and liquid condition at the same temperature, and from these the heat of fusion may be deduced. The substance in these determinations was enclosed in a glass bulb, which was broken under the surface of the water. The results, as will be seen, are not very concordant. This may be due to the fact that the samples used were different preparations, and may have been of different degrees of purity. The same preparation, however, was used in the first-quoted determination with the solid and in the first with the liquid, and similarly with the two second determinations, and these two pairs are concordant in the values which they give for the heat of fusion, 10,230 cal. according to the first determinations, 10,176 cal. according to the second. The anhydrous tetrabromide used was not satisfactorily pure. The commercial sample obtained was found, in spite of special care having been taken in its preparation, to contain a considerable amount of dibromide, for, on melting it, globules of the latter separated. After twelve fractionations by crystallizing, a sample was obtained with 8 per cent. of the dibromide, but further fractionation did not appear to reduce the amount of this impurity: indeed the formation of the dibromide seems to occur spontaneously in the tetrabromide, for it was often noticed that a specimen which was perfectly clear and free from any visible globules of the dibromide, would, on being melted again, leave particles of the latter adhering to the glass; yet in no case was the presence of bromine indicated either by colour or smell. When water is added to the tetrabromide, visible traces of bromine are liberated. The values obtained for the heat of dissolution of the anhydrous salt are not very concordant, but this may be due to the different proportions of water used in the two determinations. Hemiheptahydrate of Sodium Hydroxide (NaOH,3H2O).— The heat of formation of this hydrate (for a description of which see Trans. Chem. Soc. 1893, p. 893) from the monohydrate and water was determined. This does not constitute an unexceptionable instance; for one of the constituents itself being a hydrate, and not a simple compound, the results are complicated by the fact that the heat of fusion of this constituent may not represent its true heat of fusion, but may include some heat absorption due to partial dissociation on melting. In dealing with the monohydrate it was found that after it had solidified in the platinum bottle containing the thermometer, it was necessary to heat it very slowly indeed to remelt it, so as to avoid breaking the bulb of the thermometer by unequal expansion. Two thermometers were destroyed by heating too rapidly. Compound of Benzene and Azobenzene.-The values for benzene will be found in the Proc. Roy. Soc. loc. cit. The solvent used in determining the heat of dissolution was benzene. On calculating the results, the heat of combination, both in the liquid and solid condition, was found to be negative. It appears very improbable, however, that this should be a correct result in a case where we have direct combination occurring in the absence of a solvent, and without the formation of any secondary products: in every case, except one doubtful one, where the heat of formation of a hydrate has been measured, the value is a positive quantity (see Chem. Soc. Trans. 1887, p. 77). The explanation here may be that the measurement of the heat of fusion of benzene and of the compound is at fault: in both cases the heat-capacity in the solid condition is greater than that in the liquid condition, which is exceptional and which is generally taken to imply that the heat absorbed in fusion is absorbed gradually over an appreciable range of temperature, and not all at the ordinary fusing-point, so that the heat of fusion as measured at the fusing-point itself is too low, and the heat-capacity of the solid as measured in its neighbourhood is too high. Compounds of Naphthalene with Metadinitrobenzene and with Dinitrobenzene.-These compounds, and several others of a similar character, were obtained by Hepp (Annalen, ccxv. p. 379) by mixing solutions in benzene of the two constituents and crystallizing. The presence of any solvent appeared, however, to be superfluous, and preparations of them were made by mixing the substances in the proper proportions when liquid. It seemed desirable, however, to obtain more evidence than that heretofore existing as to the substances formed being really definite compounds. Series of freezingpoint determinations were, therefore, made, the results of which are given in Table V. and figs. 2 and 3. These prove the definite existence of the compounds in question: we have a figure made up of three curves (fig. 2), the first one representing the lowering of the freezing-point of dinitrobenzene by naphthalene, the last the lowering of the freezingpoint of naphthalene by dinitrobenzene; the intermediate one is evidently an independent curve with two branches and represents, therefore, the crystallization of some third substance; the maximum of this curve is situated at 57 per cent. of dinitrobenzene, which agrees well with equimolecular proportions, these requiring 56.7 per cent. The results in the case of dinitrotoluene are similar, except |