454 McComb : Development of the Karyokinetic 



which converge toward the cell wall indicate the poles of the fu- 

 ture spindle, but they lie in the equatorial region. It is clear that 

 the spindle shown in Fig. 8 arose from such an anlage as Fig. 2. 

 In these first two figures mentioned, the chromatin thread had not 

 segmented into chromosomes, so that it may be asserted with all 

 certainty that the anlage of the spindle may appear in Allium be- 

 fore the segmentation into chromosomes. 



The fibers of this weft, as they assume more and more a longi- 

 tudinal direction, form cone-shaped masses of spindle fibers on op- 

 posite side of the nucleus. These cone-shaped caps may at first 

 enclose a space nearly free from kinoplasmic fibers ; but later fibers 

 penetrate the cone- or cylindrical-shaped spaces and become at- 

 tached to the nucleus membrane. 



Very frequently the fibers forming the periphery of the conical 

 cap are arranged so closely together as to give in optical section, 

 the appearance of a membrane defining a hyaline space on oppo- 

 site sides of the nucleus, similar to that described by Nemec. 



In some cases, however, it seems evident that the spindle an- 

 lage does not consist of two distinct halves, but certain fibers ex- 

 tend from one polar region to the other, a condition which must 

 necessarily follow in case the fibers form a weft in the beginning 

 (Figs. 2, 3 and 6). 



It may be reasonably questioned whether such stages as rep- 

 resented in Figs. 3, 4 and 5 ever passed through that of Figs. 2 or 

 1. There was nothing in the appearance of the nucleus which 

 could justify the unqualified assertion that Fig. 3 or Fig. 4 was a 

 later stage than Fig. 2, yet much may depend upon whether the 

 division follows a period of longer rest or whether it was one of a 

 more rapid succession of divisions. We shall return to this phase 

 of the subject later on. 



It is evident from Figs. 3, 4 and 5 that we have to do with 

 mono-axial spindle anlages which are not necessarily strictly bi- 

 polar, although this is sometimes true. Frequently fibers of the 

 anlage may converge into one point at one end and into more than 

 one at the other (Figs. 3 and 4), but the developing spindle is not 

 so pronouncedly multipolar as in reproductive cells (Mottier, '97). 



It is a matter of frequent observation that the fibers on one 

 side of the nucleus are much longer than those at the other 



