212 Proceedings of Royal Society of Edinburgh. [sess 
Dr Wemyss Fulton sent me some young sand-eels, taken in tow -net, 
15 miles off Aberdeen on 16th May 1894, and preserved in spirit, 
some young flat fish (plaice?) taken off Montrose 2'lst May 1894, 
and some very small Clupeoid fishes taken 30th March 1889. In 
all of them diatoms were present. In the sand-eels four genera, 
Skeletonema, Eucampia , Melosira , and Chcetoceros ; in the flat fish 
Skeletonema and Nitzschia (in both cases Skeletonema being 
predominant), and in the Clupeoids Coscinodiscus (abundant) and 
Melosira. All these were complete valves, which appears to 
indicate that they were eaten directly by the fish, and not within 
minute Crustacea, in which case they would have been broken 
up into fragments. 
The seasonal occurrence of diatoms has, however, its special 
botanical interest. The ordinary processes of reproduction by 
simple division and the more rare formation of auxospores does not 
appear from my opening citation of literature to have satisfied 
observers. Observations, accurate or not, of other forms of repro- 
duction have been in the air, if the expression may be permitted, 
for some years, and Dr John Murray has, with rare divination, 
always associated this seasonal occurrence with some mode of 
reproduction yet to be disclosed. At his instigation, then, I have 
made the following observations, and since the modes of reproduction 
are of diverse sorts, though all broadly the same, I will proceed 
from the simplest case to the others. 
In a tow-net haul obtained last April off Gigha there was a 
large quantity of Bidduljohia mobiliensis. Many of the specimens 
were in the state represented in Plate I. fig. 2 ; the whole of the 
cell-contents rounded off and in many cases with a large oil drop 
enclosed by the chromatophores, nucleus, &c. The same was the 
case with Ditylum Briglitwellii , and an extreme form of contraction 
of contents in this species is figured on the same plate (fig. 6). It 
was only on a subsequent and more detailed examination of this 
and other captures that the state represented on fig. 3 was dis- 
covered. There is here within the parent Biddulphia a young 
“ cyst ” with a slightly silicified wall, but without the characteristic 
spines, &c., of the parent form. I take this “cyst” of B. 
mobiliensis to correspond to the form described by Professor Cleve 
(reproduced in fig. 1) for Biddulphia aurita (which also is without 
