1896.] The Classification of Diatoms. 535 
ceeded still further, the cuneate form would become rostrate 
and a still further narrowing would give a capitate form. So 
form and size, although they have a certain significance, are 
not to be considered infallible characters. 
The geological records throw no light upon the relationship 
of the Bacillariacex, for when this family first appeared, we find 
the same genera, and largely the same species as in our modern 
ones. This is probably due to the fact that their ancestors 
lacked the siliceous covering, and hence were not preserved. 
Diatoms evolved the same as all other plants until they devel- 
oped their shells, but these put a stop to their further evolu- 
tion, at least they show no trace of evolution since their first 
appearance. So the question arises whether the Diatoms repre- 
sent the ends of several closely related genetic lines the further 
development of which was stopped by their siliceous shells, or 
whether we may trace the development of one form from an- 
other. The former supposition is the more probable, for the 
form of the earliest fossil specimens is identical with that of 
modern specimens of the same species; and the same genera 
are found among fossil as among modern Diatoms. If one 
genus of Diatoms developed from another, we ought to find the 
more primitive forms in the earlier strata, for there is little 
chance that their remains would not be preserved had they 
existed. But instead of this, Diatoms of all forms appear 
almost simultaneously. We may conclude then that the 
Bacillariacex represent the silicified ends of several closely allied 
genetic lines and that they have not changed in form since 
they acquired their siliceous covering. The structure of the 
valves it follows will tell us practically nothing of their rela- 
tionship. 
There are five methods by which auxospores are formed”. 
In the first the protoplasm of one frustule simply escapes from 
the valves, grows to a certain size, and then invests itself with 
new valves. In the second, two auxospores, instead of one, 
are formed in the same way by the dividing of the protoplasm 
of a single plant. In the third, the protoplasm of two Diatoms 
unites to form an auxospore. In the fourth, the protoplasm of 
18 Murray, 1. c. 
