Bartow—A Mechanical Cause of Homogeneity of Crystals. 591 
3. As to the effects produced by the presence of linked 
groups. 
We see that in most respects the groups will behave much as 
single unlinked balls would do—homogeneity and close packing 
will, as we have already concluded, frequently go together whether 
the balls of an assemblage are aggregated in groups or not. 
The same balls may, it is evident, under the same external con- 
ditions, take up two or more different relative arrangements in 
seeking equilibrium if some of them are linked together to form 
groups, and the grouping is different in different cases; and this 
will apply whether the balls are all of one kind or not, and whether 
in the different cases the groups contain the same sets of balls or 
not. 
Marked difference of behaviour of the same set of balls may be 
expected in different cases, especially when we compare a case in 
which few links subsist with @ case where there are many. Thus 
it is evident that the presence of links between balls will commonly 
prevent them from packing as closely as they otherwise would do, 
and that the more they are interlinked, the less free they will be to 
accommodate themselves to their surroundings, and the more the 
closeness of the packing will be likely to be impaired. And when 
a number of assemblages having similarity of composition, but 
differing in the amount of linking which obtains in them, are com- 
pared, the differences in the degree of closeness of the packing 
- which are thus occasioned may be revealed by differences in the 
susceptibility of the assemblage to some external influence. 
To compare with this we may cite the following from Lothar 
Meyer :—“ Experience has shown that the normal compounds 
always have a higher boiling point than those with side chains,’ 
and that the boiling point of the latter falls as the number of side 
chains increases. In the case of bodies having a similar constitu- 
tion, the addition of CH, raises the boiling point by from 18° to 
po 2 2 
1 Ifthe resemblance here pointed out has any significance, it would seem to furnish 
some additional evidence as to the nature of these ‘‘ side chains ’’—to show that they 
have some of the properties of the links referred to in the text. 
* Lothar Meyer’s ‘‘Grundziige der theoretischen Chemie,” p. 84; or Bedson and 
Williams’ translation, p. 91. 
SCIEN. PROC. R.D.S., VOL. VIII., PART VI. 2X 
