KIRKWOOD : POLLEN-FORMATION IN CUCURBITACEAE 229 
neither of these works, however, does it appear that they bear any 
relation to the extranuclear nucleoli. Duggar refers to the almost 
simultaneous appearance of the fibers and darkly-staining granu- 
les, and regards both as due to the nuclear changes in synapsis. 
Lloyd regards the appearance of such cytoplasmic fibers as pos- 
sibly due to the streaming of hyaloplasm or kinoplasm, from the 
fact of their similarity in appearance to features described by 
Wilson * and others working upon other material. That such a 
kinoplasmic streaming toward the nucleus occurs in Micrampelis 
is strongly suggested by the marked radial elongation of the cyto- 
plasmic meshes during the prophase of the first division (FIGURE 
25), but this feature is uniform throughout the cell. 
A careful study of the material at hand has convinced the 
writer that in this case the extranuclear nucleoli arise from the 
conspicuous cytoplasmic fibers which make their appearance dur- 
ing the growth of the pollen-mother-cells. These fibers, at first 
thin and delicate, become thicker and stain more darkly, sometimes 
appearing as a series of nodules strung together, and again as thick 
crooked masses. As their development proceeds the fibers dis- 
appear and the round cytoplasmic nucleoli multiply and for a time 
appear in the area occupied by the fibers (FIGURES 23-68). It is 
recognized, however, that the phenomena observed may be open 
to more than one interpretation. The fibers observed lie in the cen- 
ter of the largest amount of cytoplasm in the cell and they must be 
taken as evidence of a cytoplasmic activity of some sort. It is 
possible that, these being the centers of special activity, they are 
the points at which the disposition of certain cytoplasmic products 
first takes place. 
If we are to regard the spherical bodies in the cytoplasm of the 
Spore-mother-cells of Micrampelis as extranuclear nucleoli, accord- 
ing to Strasburger’s view, we must account for their persistence 
in the cytoplasm through all stages from one generation of cells 
to another. It does, nevertheless, seem that they are associated 
with the kinoplasm, inasmuch as they appear in the rod-like form 
to bea part of the cytoplasmic reticulum. That they are actually 
connected with the spindle at any stage of its development could not 
be determined accurately, though their position would often favor 
Such a view, During the metaphase certain fibers may be seen 
