Bartow—A. Mechanical Cause of Homogeneity of Crystals. 587 
alike, but enantiomorphously related. As they lie on axes they 
are singular points.’ 
If in the case under consideration a homogeneous arrangement 
is produced, it is probably this.’ 
Suppose next :— 
(0) That the distance between the two centres of a pair is 
relatively large, so as very nearly but not quite to allow the balls 
to take the arrangement prevailing when they are all indepen- 
dent.® 
In this case an arrangement approximating closely to the 
compact arrangement taken up by the balls when independent will 
obviously be produced ; it cannot, how- 
ever, be a homogeneous one belonging 
to the cubic system, because it is not 
possible to connect the points of the 
closest-packed cubic system above de- 
scribed, two and two in such a way 
that the arrangement of the ties shall 
comply with the definition of homo- 
geneousness and belong to the regular 
system. In some instances of this kind 
probably an arrangement of the balls 
according to the system 21 of Sohncke with centres of inversion, 
d.e., of type 49a, in my list* will result, the generic symmetry 
being that of class 12 in Sohncke’s list. The projection of 
a stack of spheres whose centres would have such an arrange- 
ment is shown in fig. 4, the balls being so placed as very nearly 
to have the alternative closest-packed homogeneous arrangement 
‘ 
1 Thid., 23, p. 60. Comp. Min. Mag., vol. xi., p. 182. 
2 Other symmetrical orientations which are consistent with homogeneousness will 
be found to give a lower degree of symmetry and would probably not give such close 
packing. 
3 A comparison of the specific weights of different bodies composed of the same 
atoms in the same proportions, but whose molecules are of different degrees of com- 
plexity, rather favours the conclusion that, in the solid and liquid states, if as the 
stereochemists suppose, the atoms have definite situations, the distances separating the 
nearest atoms of different molecules of a substance are frequently not very much greater 
than the distances separating the nearest atoms of the same molecule. 
* Zeitschr. fiir Kryst., 23, pp, 30 and 46. 
