Trow . — The K ary o logy of Saprolegnia . 639 
haematoxylin. Later, during the examination of S. Thureti 
and 5 . mixta , an attempt was made to obtain better prepara- 
tions by using gentian violet and eosin, but this proved 
entirely unsuccessful. 
The difficulty of following the processes connected with 
the maturation of the oospore is extreme, not only because 
of the changes in the cell- wall of the oospore, which prevent the 
efficient action of the fixing reagents, but also by reason of 
the appearance in the ripening oospore of entirely new and 
puzzling structures. Fig. 23 a, , b, c f d and e represent succes- 
sive stages in the development of an oospore. The nuclei 
long remain distinct, but ultimately come together and lie 
closely side by side or end to end as represented in b. The 
histological details are difficult to follow, but it is easy, up to 
this point, to satisfy oneself of the indubitable nuclear char- 
acter of the structures by a study of successive stages. In e 
is represented a uninucleate oospore, but as I have not been 
able to observe those stages which illustrate the actual fusion, 
I do not attach much value to it. At this stage it would be easy 
to overlook one of the nuclei under certain circumstances, as, 
for example, when side by side and in the same vertical plane. 
In the stage represented at b the protoplasm is seen to 
contain deeply stained granules like large microsomata, lying 
in the knots of the protoplasmic meshwork. These appear 
to correspond to the drops of glycogen observed by Dangeard. 
They rapidly increase in size, as may be seen by comparing 
b, c , d and e. Ultimately they are seen, especially clearly in 
the haematoxylin preparations, to consist of hollow spheres, 
rings, or crescents. Finally, they increase further in size and 
diminish in number by fusing together in pairs until, at 
a late stage in the maturation of the oospore, they are 
reduced to a few rings or crescents, or sometimes a single 
hollow or more or less chambered sphere. That the rings 
are not optical sections of hollow spheres is proved by the 
fact that they frequently overlap each other, as well as by 
the appearances presented by optical sections from various 
points of view. 
