34 2 Olive. — Sexual Cell Fusions and Vegetative 
almost entirely in the lower, more swollen part of the nucleus. Whether 
this nucleus has yet passed through the equatorial-plate stage I am not 
able to state positively, as this stage is somewhat obscure for the spermo- 
gonial nuclei of Triphragmium. This phase is more clearly brought out in 
the later sporophytic divisions. 
Fig. 4 shows a somewhat later stage of division in the spermatial fila- 
ment of Triphragmium. This seems to be an early anaphase, since some 
of the chromatin still occupies a more or less central position, while other 
chromatin material has already been drawn closer to the poles. The 
central spindle is here shown to be a rather thick, achromatic strand, 
situated a little to one side of the axis of the figure, while the mantle fibres 
are also more or less clearly defined, having one end attached to the poles, 
and extending thence along the axis as granular filaments, which appear to 
be mostly chromatic in their staining reactions. An aster is shown at the 
lower pole, and also at the lower pole of Fig. 3. 
Fig. 5, from the spermogonium of Uromyces Lilii , although somewhat 
ill-defined and difficult of interpretation, apparently represents a stage next 
in order of development. The long axial strand is probably the central 
spindle, while the mantle fibres, particularly at the lower pole, suggest 
a late anaphase. It is of interest to observe in this figure that, although 
mitosis is somewhat far advanced, the dividing nucleus has not yet passed 
into the end of the filament, where the spermatial' bud is already well 
formed. Herein this form on Lilium differs markedly from Triphragmium , 
in which the nucleus, when in quite an early stage of mitosis, moves 
upward into the very tip of the budding spermatium (see Figs. 2, 3, 4, 6). 
A late anaphase of the spermogonial nuclei of Triphragmium is repre- 
sented in Fig. 6. Such a figure is perhaps most often met with, presumably 
because its sharply-defined poles so easily catch the eye. Only a few 
obscure filaments now remain of the central spindle. But the mantle 
fibres, on the other hand, are at this stage so conspicuous that they can be 
quite easily and accurately counted. In Fig. 6 a, for example, a daughter- 
nucleus in anaphase condition is drawn, in which the filaments attached to 
the poles are seen to number eight. Eight indicates also, in my opinion, 
the chromosome number for this form. 
This last-mentioned figure (Fig. 6 a), which is drawn with a very high 
magnification, shows most clearly a peculiar phenomenon. The centre 
is here plainly a double structure, and each half has four clearly-defined 
strands attached to it. The double nature of the centrosome is also shown 
at both poles of Fig. 6 , and at the lower pole of Fig. 3. Two possible 
explanations of this phenomenon are apparent. The doubling has obviously 
arisen either from a premature division of the centre preparatory to the 
next division ; or else it results from the close proximity in the same 
nucleus of the chromatin contents of the maternal and paternal nuclei, 
