590 . Scientific Proceedings, Royal Dublin Society. 
Further, the more complicated assemblages produced in this — 
way can be subjected to symmetrical aggregation and partitioning, . 
and still more complicated groups obtained in this way; or a 
number of smaller groups of new patterns can be had by sym- 
metrically altering the linking. And then, by appropriate changes — 
of the external conditions, assemblages thus arrived at can be 
conceived to be broken up, and the various kinds of groups massed 
by themselves, all in obedience to the principle of closest-packing. 
In this way, by the action of closest-packing, with the aid of 
the additional hypothesis of symmetrical linking, very intricate 
results may be reached by a number of successive steps as the 
conditions change, and orderly complexity and highly specialized 
grouping, such as appeared at first sight impossible of attainment 
by any selective action of the principle referred to, is seen to be 
quite within the capacity of this agency for producing symmetry. 
As already noticed, a slight change in the relation between the 
balls may cause closest-packing to be attained in a different homo- 
geneous arrangement,’ and it is possible that in some cases two or 
more kinds of balls may be so related as readily to form a variety 
of different groups with but comparatively slight changes of the — 
general conditions.” Two kinds of balls may for example be so 
related as, under slightly different conditions, to form groups of 
all or most of the different kinds just referred to above, and also 
complex combinations of these groups. 
We may compare with this the complex isomerism displayed by 
the hydrocarbons, which may perhaps, as just hinted, be traceable 
to a species of polymorphism. 
It is important in this connection to distinguish between that 
kind of dimorphism or polymorphism which affects only the solid 
or completely-linked state, and dimorphism or polymorphism of the 
kind just referred to. In cases of the former the two forms 
become identical when they pass to the liquid condition, 7. e., they 
have the same grouping. In cases ofthe latter not only is the 
arrangement of the tranquil homogeneous assemblages different, but 
the grouping when they break up is different, although the consti- — 
tuents of the assemblages are the same. 
1 See p. 575. 2 Compare p. 576. 
