Bartow—A Mechanical Cause of Homogeneity of Crystals. 677 
In this case the other constituent will consist of balls suffi- 
ciently small to be inoperative in the sense above defined,’ or at least 
so small as not to be prevented from entering the interstices of the 
solid constituent. 
Finally, in a combination of two fluid assemblages there are 
two ways in which the balls or groups forming one constituent can 
be conceived to be related to those of the other, and this is the 
case whether the intermixture is accompanied by change of state 
ornot. Hither the balls of both constituents may be operative or 
those of one constituent may be all inoperative. And a complete 
gradation is conceivable from extreme cases of the first kind, in 
which each constituent is equally concerned in determining the 
kind of arrangement to be displayed by the compound assemblage, 
to cases in which nearly or quite all the general arrangement is done 
by the balls of one constituent only, the other, in obedience to the 
principle of closest-packing, just filtering into the available inter- 
spaces without producing any material permanent effect on the 
arrangement of the balls among which it intrudes itself. 
The rate at which intermixture proceeds will, it is evident, 
largely depend on the nature of the relation just referred to. 
Local disturbances, whether brought about by external causes— 
as stirring or shaking—or by changes occurring within the assem- 
blages themselves, or arising partly from both sources, will, it is 
evident, on the one hand, promote intermixture and render its 
progress more rapid, but on the other hand, they will, if they 
continue, hinder the intermixture from becoming completely 
symmetrical or homogeneous. Disturbance followed by a gradual 
passage to tranquillity will best promote the formation of a 
homogeneous intermixture by the action of the principle of closest- 
packing. ; 
Deferring the consideration of one or two details of the in- 
termixing of assemblages for a moment, let us now compare the 
processes thus traced with actual phenomena. 
We have just seen that in the presence of the disturbances 
referred to, permanent combination in precise proportions will only 
occur where the local attainment of perfect homogeneity brought 
1See p. 547. This is more readily understood if we take mutually-repellent particles 
instead of balls. 
