Bartow—A Mechanical Cause of Homogeneity of Crystals. 679 
passing disturbances continually cause lapses to irregularity, the 
effects which must evidently be produced find a very close parallel 
in certain phenomena of diffusion where there are limits to the 
combining proportions. For in the case of the hypothetical com- 
bination, as in that of diffusion, we shall have intimate intermixture 
in practically uniform proportions throughout, without any de- 
finite orientation, and a saturation-point whose position will 
practically be definite, and depend on the amount of disturbance 
prevalent,' as well as on the nature of the ideal equilibrium ar- 
rangement. 
The rate at which intermixture takes place will, as has been 
said, largely depend on whether operative? constituents are to be 
found in both combining assemblages or not, and on the part the 
latter respectively take in determining the nature of the ideal 
equilibrium arrangement. 
We can conceive of assemblages composed of linked groups 
whose forms are such that the different groups fit together so ill 
that the law of closest-packing affords great resistance to inter- 
mixture in any proportions regular or otherwise. And this is 
paralleled by some of the cases in which liquid bodies will not 
diffuse into one another.’ 
Comparisons may be instituted with the diffusion of gases as 
well as with the diffusion of liquids. 
_ There are also resemblances to the case of a solid assemblage 
receiving within its interstices inactive constituents in obedience 
to the law of closest-packing. Thus we have the common pheno- 
menon of solid bodies taking up small quantities of water or certain 
other liquids, without material change of properties ; in some cases 
the liquid previous to its absorption being gaseous, as in the case 
of hydrogen occluded by platinum or palladium. 
In cases of the latter kind the comparison is instituted with 
the subsequent intermixture, and not with the change of state. 
Having thus indicated the general resemblances subsisting 
between actual phenomena of combination and intermixture and 
the behaviour of our hypothetical assemblages, a few additional 
1 In the case of diffusion, the vigour. of local disturbances is probably traceable to 
temperature conditions. (But see Appendix, note 1, p. 688.) 
2 See p. 547. 3 Comp. pp. 670 and 672. 
