68 COMPARATIVE EMBRYOLOGY OF THE RUBIACEAE 
by inference, has nothing to do with the formation of the mitotic 
figure. At the beginning of mitosis the cytoplasm shows again a 
uniform reticular structure, and the spindle of the first division 
takes its origin in radial or more or less irregularly placed fibers 
which penetrate the nuclear cavity, and form a multipolar spindle 
primordium. 
With regard to the disappearance of the earlier formed thick 
fibers I am not able to establish a parallel in the Rubiaceae, for, in 
the embryo-sac-mother-cell in the plants studied they persist until 
the earlier stages of mitosis, at least through diakinesis. That 
they are in some way connected with the kinoplasmic figure dur- 
ing mitosis would appear from the circumstance that their ends on 
reaching the nucleus bend and run as figured in Mottier's ۸ 2, 
fig. 17, of the paper referred to. This condition, as seen in Cru- 
ctanella macrostachya is shown in pl. 11, fig. 37. From this and 
the adjoining figures it will be seen that the origin of the spindle 
is multipolar, and, with the possible exception of the peculiar be- 
havior above described, corresponds to the descriptions of Oster- 
hout * and others. The definitive spindle of the first division is 
strictly bipolar, and differs in the three plants in one or two details 
of form only. 
In Crucianella gilanica the spindle is of very regular, broad, 
oval form, with sharply-pointed ends ; in C._macrostachya it has a 
looser, more slender form, with attenuated ends, while in Asperula 
this character is still more accentuated. Such differences may in’ 
the main be regarded as inconstant and of no general importance, 
and may in many cases also be appearances referable to the plane 
in which the material happens to be cut. Insufficient observation 
of this point may very easily lead to false impressions, as Stras- 
burger t has hinted, as with reference to the character of 
the spindle poles. 
Generally the ends of the spindle poles lie freely in the cyto- 
plasm, inasmuch as the spindle axis usually lies parallel to the 
longo axis of ms cell, which, as oe shown, is er per 
* Ueber E 
"s Pg TETE don d a Spindel bei سر‎ Jah wiss. 
Ueber Red 
am e Spindelbildung, Centrosomen und Cilienbildner im 
* 
