RICHARDS. — SIGNIFICANCE OF CHANGING ATOMIC VOLUME. 587 



Thus, the diminution of the quotient always accompanies a diminution 

 in compressibility, as would be expected on the basis of the hypothesis. 

 The nature of the case prevents a more exact mathematical treatment of 

 these expected variations, but the unanimous verdict of this approximate 

 treatment of them advances the agreement of fact and hypothesis to an- 

 other order of approximation, in which not only the qualitative succession 

 of magnitudes possess significance, but also the deviations from direct 

 proportionality. Thus is afforded a strong presumption in favor of the 

 important significance of changing atomic volume. 



Many other cases of the same kind might be cited, but occasionally an 

 apparent exception appears. For example, silver can hardly be much 

 more compressible than zinc, and yet during the union of a mol of 

 chlorine with the former metal a contraction of 19 milliliters occurs, 

 while with the latter the contraction is only 13 milliliters, although the 

 heat of formation in the latter case is almost twice as great as in the 

 former.* 



At first such an exception as this seems a serious impediment to the 

 acceptance of the underlying hypothesis of compressible atoms, but fur- 

 ther consideration shows at once the reason of the seeming anomaly, and, 

 instead of being an exception, the case becomes a supporting example. 



It will be apparent below that such cases are easily explained by 

 the plausible assumption that more than one kind of internal pressure 

 exists in a solid compound, — not only are elements probably compressed 

 by their chemical attraction or affinity, but also the smallest particles of 

 the resulting compound exert pressure in cohering under the influence 

 of a cohesive attraction of less intensity. 



In endeavoring to imagine the mechanism of the action of these 

 forces, it is necessary to amplify the ordinary conception of atoms, and 

 endeavor to imagine the behavior of each single atom under the action 

 of the stresses to which it must be subjected. Probably each atom iu 

 such a solid as argentic chloride would be bound on one side by chemical 



rection applies so nearly equally to all that it may be neglected for the present, 

 however. 



* The molecular volume of argentic and zincic chlorides are respectively 25.8 

 and 46.9, and those of the metals 111.:: and 9.5, while the respective heats of forma- 

 tion (corrected), arc 1 1 1 and :is 1. Since the contraction of the chlorine is probably 

 the chief factor in the question, equivalent quantities of the two materials must he 

 compared ; therefore, in the calculation double the data of silver should be tak< a 

 to correspond to one of zinc, because zinc is bivalent. 



In the previous paper where this was discussed (the second of the series) a care 

 mistake in calculation of the silver chloride exaggerated the discrepancy. 



