1122 Proceedings of Royal Society of Edinburgh. [sess. 
Both curves have two climaxes, which become an important 
support for our opinion that there are two races. The comparison 
of the two tables shows further that the drying of specimens 
causes the climax of the narrow race to be transferred from 4-5 y 
to 8 y, and that of the broad race from 8 to 12 y* It is natural 
that the narrow race has changed a little less than the broad 
one, because the change occurring is that of a cylinder being 
flattened, and this effect must be greater in a large cylinder 
than in a smaller one. The effect of drying may also more 
directly be made out. Some of the broader specimens contain 
resting spores which are much more strongly silicified than the 
other part of the cell, and therefore they do not become flattened 
in any mentionable degree. In two specimens the figures were : 
breadth of the cells, 1 4 y and 11 y, breadth of the corresponding 
spores respectively, 11 y and 8 *5 y ; the difference is then 3- 
2 - 5 y, that is, 1 y less than the comparison of the tables gives us, 
but it must be remembered that also the spore becomes a little 
flattened. 
The result arrived at through all these reflections is, that the 
narrow race, which has its normal breadth at about 4 y. must be 
reckoned as about 8 y broad, if we want to compare our figures 
with the measurements of other authors ; regarding the broad 
race the figure is, instead of 8 y 7 about 12 y. To make these 
figures as exact as possible, I have measured specimens from 
all the samples in dried condition ; the above Table II. gives 
the number of specimens measured in each sample, and in this 
table the distinction between the narrow and the broad race 
has been determined according to the above reflection and to 
the following Table V., in which the breadth of 250 specimens 
is represented. 
* An ocularmicrometer, in which each space between two lines corresponds 
to 4 /x, has been used for the measurements ; this may explain the curious fact 
that Table III. is one-sided, then the specimens, which really are 3 /x broad, 
may easily be taken as 4 ^ broad. 
