Bartow—A Mechanical Cause of Homogeneity of Crystals. 641 
sections being intimately related to the external form, and thus 
indicating a connection with strain set up by the change.! 
In others, in which the zeolite is already twinned, the twinning 
disappears, and the higher symmetry imitated by it actually 
obtains as the water is driven out, e. g., in the case of desmine.” 
Multiple twinning by dimorphous change which involves some relative 
displacement of the different individuals—Mimetic crystals. 
Multiple-twinning of a less symmetrical character is producible 
as follows :— 
Tn cases of dimorphous change let the shearing and distortion 
which take place in consequence of the existence of some restraint 
on change of form in the way above explained, be such as to in- 
volve, when completed, a diminution of symmetry so that alternate 
individuals found ranged about an axis after the change has taken 
place do not possess the same orientation of parts. 
The most obvious way for this diminution of symmetry to come 
about is for the angular dimension of the individuals to change so 
that they are no longer capable of fitting exactly together around 
the axis of the nucleus, a condition of strain and ultimate rupture 
of the altering assemblage being brought about in consequence of 
the angular dimension of the individuals becoming either too 
great or too small for them collectively to make out the entire 360° 
about the axis around which they lie. 
The principle of closest-packing will, however, limit the amount 
of disturbance thus caused so as to make it the least possible, and 
thus when a compound-twinned nucleus is formed in an assemblage 
in this way, and the general conditions are regular and favourable, 
as many segments or individuals will continue symmetrically related as 
the change of angular dimensions will admit of, the dislocation being 
confined to one place or to two places on opposite sides of the axis. 
When a distorted solid nucleus is once formed it is easy to 
see that regular growth® of its various surfaces will perpetuate 
1 Math. und naturwiss. Mittheilungen aus den Sitzungsberichten der kéniglich 
Akad. der Wissenschaften Berlin, 1890, p. 708. 
2 Tbid., p. 717. Comp. Brauns, Joc. cit., pp. 814-316, and 320. 
3 The regularity need not be so great as that the growth is uniform at any given 
3A2 
