The question being whether the molecular weight of ammonium chloride is two vols. or four vols., an idea of the magnitude of the assumed decomposition is conveyed by the proportion of the volume of the decomposed salt to the volume of the non-decomposed, and Mr. Greene's quotation of the percentage of weight is irrelevant and misleading, and his number not even correct. A mixture containing
1.055 vols. of spec. gr. 26.75 = 28.22 and
12.32 " " " " 13.375 = 164.78
------ ------
13.375 " 193
has the spec. gr. 193 / 13.375 = 14.43. The proportion in one vol. of the undecomposed to the decomposed salt is, therefore, as 1 to 11.68 and the percentage of volume of the former 0.0789, and that of weight 28.22 / 193 = 0.146, and not 0.16.
It is not easy to imagine why a small fraction of the heavy molecules should be volatilized undecomposed, the temperature being sufficient to decompose the great bulk. Marignac assumes, indeed, partial decomposition, but the difficulties which he encountered in making the experiments, on the results of which his opinion rests, were so great that he himself accords to the numbers obtained by him only the value of a rough approximation.
The heat absorbed in volatilization will comprise the heat of combination as well as of aggregation, if decomposition takes place, and will therefore be the same as that set free at combination. Favre and Silbermann found this to be 743.5 at ordinary temperature, from which Marignac concludes that it would be 715 for the temperature 350°; he found as the heat of volatilization 706, but considers the probable exact value to be between 617 and 818.[1]
[Footnote 1: See Comptes Rendus, t. lxvii., p. 877.]
An uncertainty within so wide a range does not justify the confidence of Mr. Greene which he expresses in these words: "It is, therefore, extremely probable that ammonium chloride is almost entirely dissociated, even at the temperature of volatilization." By Boettinger's apparatus a decomposition may possibly have been demonstrated, but it remains to be seen whether it is not due to some special cause.
When Mr. Greene says that the relations between the physical properties of solids and liquids and their molecular composition can in no manner affect the laws of gases, nobody is likely to dissent; but the conclusion that their discussion is foreign to the question of the number of molecules in unit of volume does by no means follow. If the specific gravity of a solid or the weight of unit of volume represents a certain number of molecules, and is found to occupy two volumes in a compound of the solid with another solid, the number of molecules in one volume is reduced to one half. This I have shown to be the case in a number of compounds, and the decrease of the specific gravity with increase of the complexity of composition appears to be a general law, as may be concluded from the very low specific gravity of the most highly organized compounds, for instance the fatty bodies, the molecules of which, being composed of very many constituents, are of heavy weight; and likewise the compounds which occur in combination with water and without it, the simpler compound having invariably a greater specific gravity than the one combined with water; for instance:
BaH_2O_2 sp. gr. 4.495
" " + 8H_2O " 1.656
S_2H_2O_2 " 3.625
" " + 8H_2O " 1.396
FeSO_4 " 3.138
" + 7H_2O " 1.857
and so in every other case. This is now a recurrence of what takes place in gases, and proves the fallacy of the hypothesis; for if these compounds could be volatilized the vapor densities would necessarily vary in the inverse proportion of the degree of composition.