RICHARDS. — SIGNIFICANCK OF CUANGING ATOMIC VOLUME. 589 



found by Biltz and Victor Meyer* to be non-volatile at 1400°, and only 

 slightly volatile at 1700°. The Iiigli boiling point of argentic chloride is 

 one of the indications of great molecular cohesion, f and the pressure of 

 this cohesion must produce great compression of the chlorine. Tlierefore 

 we may refer the unusually great contraction taking place during the 

 formation of argentic chloride chiefly to the physical intermolecular 

 pressure and not so much to the essentially chemical interatomic pressure. 



Even this, however, is not a complete treatment of the matter, since 

 the metals themselves have different boiling points, and hence are prob- 

 ably under different internal pressures. Silver itself must be more com- 

 jiressed by its own cohesion than zinc is, for Nernst has shown that the 

 former metal probablj'^ boils at a point above 1950° | This condition in 

 silver must tend to counterbalance in part the effect of the great internal 

 pressure in argentic chloride, for if silver is already more greatly com- 

 pressed, it could not be so much changed as zinc by additional pressure. 

 The much greater compressibility of chlorine is undoubtedly tlie reason 

 why the great cohesion of argentic chloride diminishes the volume of 

 a salt to a greater extent than the great cohesion of the silver diminishes 

 its own volume. 



The complication of these various conflicting tendencies is no argu- 

 ment against the validity of the reasoning; on the contrary, any pertinent 

 hypothesis which does not consider them is incomplete. The volume- 

 change occurring when two solids or liquids combine is obviously an aver- 

 age or composite change, depending upon the extent of both the chemical 

 and physical compression of the factors which must be overcome, as well 

 as ujion both the chemical and physical compression of the product, — 

 just as the heat of reaction concerns the heat absorbed in disintegrating 

 the factors, as well as the ''chemical heat" and the "latent heat" given 

 out in the act of combining and solidifying. These considerations were 

 neglected in comparing above the cases of haloid salts of the three alkali- 

 metals, since these salts are very similar, and their boiling points probably 

 not far apart. § Moreover, the metals, too, are very similar. Hence the 



* Ber. d. deutsch. ch. Ges., 22, 727 (1889). 



t Ostwald gives a table illustrating Trouton's rule (Lehrbuch, 1, 354), which 

 shows tliat with nearly related substances the boiling point is very nearly propor- 

 tional to the molecular latent heat of vaporization, and that even with very 

 different substances this is approximately true. 



t Z. Elektrochem., 9, 62 (1903). 



§ The boiling points of these salts have not been accurately determined, but 

 qualitative experiments of my own in tubes of fused quartz showed that all the 



