554 Trow. — On Fertilization in the Saprolegnieae . 
nuclei has already commenced. The most distinct and deeply stained 
nuclei are in the spirem stage (Fig. 17 a, hi) Later prophases are to be 
seen (Fig. 17 c). Some of the nuclei, e.g. those marked e, f,g, h, are pro- 
vided with deeply stained granules attached to the outside of the nuclear 
membrane. In one case, marked /, three granules were seen in this 
position, with protoplasmic threads radiating outwards from them. Such 
an assemblage of nuclei is of course very puzzling, and as this and the 
following stages are difficult to fix properly, there is room for errors in 
interpretation. Contrary to Davis’s experience, the staining is not excep- 
tionally difficult, the detail comes out perfectly with all the methods 
employed. The most natural inference to be drawn from these observa- 
tions seems to be, that a certain number of nuclei are selected to 
undergo a second mitosis, and that the remainder are doomed to de- 
generation. 
The further study of the fate of these dividing nuclei is very difficult, 
and, but for the successful fixation of one culture, might have proved barren 
of real result. One oogonium, thoroughly examined in a series of seventeen 
sections — a week’s work— -furnished examples of the metaphases and 
anaphases of this second mitosis. Some of these sections are represented, 
wholly or in part, in Fig. 18. It will be observed that the dividing nuclei 
are associated with small resting nuclei which still retain their normal 
characters and with degenerate nuclei of various types. The nuclei do not 
divide synchronously. In the late metaphases, as shown at a, b, c, d, a 
small elongated spindle is seen. It is noteworthy that this appears to 
consist always of three threads. Whether the two outer lines represent 
spindle threads or the optical sections of the nuclear membrane may be 
a matter of controversy, but an examination of the nuclear figure at a leads 
to the belief that the nuclear membrane is absent and that the chromosomes 
pass polewards along all these threads in a less regular manner than is the 
case during the first mitosis. This division, indeed, resembles in many of 
its features the corresponding division in Albugo described by Stevens (’ 99 ) 
and Ruhland (’ 03 ). At the poles of the spindles there can generally be 
detected a small deeply stained granule, which, as it is the centre of a system 
of polar radiations, must be regarded as a centrosome, the radiating threads 
constituting its astrosphere or aster. The rays of the astrosphere can be 
traced without any break of continuity right up to the centrosome. Late 
anaphases are seen in *?, /, g, hj. The two nuclei at e are daughter-nuclei, 
the connecting threads of the spindle being no longer continuous. One of 
the daughter-nuclei has turned through an angle of nearly 90°. The new 
nuclear membrane makes its appearance at this point, but is at first not 
readily demonstrable. At /, g, the centrosomes somewhat obscure their 
nuclei. These, although demonstrable, have been omitted from the draw- 
ings for the sake of clearness. The nucleus figured at^ had neither centro- 
