252 TRANSACTIONS OF THE ROYAL CANADIAN INSTITUTE [VOL. XI 
for the general distribution of pits it will be necessary to arrive at a 
knowledge of the conditions which cause the selection of one point for 
the location of a pit in preference to another. 
Before formulating any new theory it is necessary to review briefly 
those already suggested and to point out in each case wherein it has 
failed to satisfy the test of experimental verification. It has been sup- 
posed that the presence of extraneous particles such as dust upon the 
surface would provide points of attack for the corrosive. The experi- 
ments in which dusty crystals were exposed without corresponding 
results, and also the fact that the irregularity is as striking on fresh 
cleavage faces, show that this idea is quite untenable. But it is possible 
that there may be fracturings or scratches on the surface in which the 
corrosive could find a lodgement, unhampered by currents, and so a 
start might be made. Direct experimental disproof has already been 
offered in the account of the colemanite experiments, and Daly™ has 
shown that such surface fracturings cannot be the universal cause, for 
in the case of an alum crystal growing in a saturated solution of the 
salt, a slight dilution of the liquid results in the immediate formation 
of pits. Under these conditions the crystal faces must have been per- 
fectly unbroken. His conclusion is that the arrangement is “structurally 
accidental’’. The possible connection between pits and inclusions 
which the experiments with colemanite seemed to indicate, was on 
further consideration seen to be one of a common cause and not of cause 
and effect. There remains now, in addition to Goldschmidt’s theory 
as outlined in the opening pages of this paper, another, suggested by 
von Ebner™, which supposes that the pit originates at the point where 
particles of the corrosive of greater concentration than the average come 
in contact with the crystal. Both of these theories are put to the test 
when action takes place in boiling solution and in a shaken test-tube, 
and also when the corrosive is applied in the form of a jet. Under such 
conditions there can be neither eddies nor differences of concentration 
in the solution. 
Since all of these theories fail when put to experimental test, one 
must look for a clue leading to a new explanation of this unexpected 
state of affairs. We must bear in mind however, that whatever the 
conditions may be that allow a selective attack by the corrosive, they 
must be common to adjacent sides of each cleavage plane, since on 
these the arrangement of the pits is the same. A suggestive fact in 
this connection is the growth of pits in overlapping parallelism in more 
8 op. cit., page 389. 
4 Sitzber. d. Akad. d. Wissensch. zu Wien, 89: 368. 1884. 
