140 Botanisches Centralblatt. — Beiheft 2. 



of the daugther nucleus from the surrounding cytoplasm. The 

 nuclear nct-work which is now quite evident reveals the sarae 

 mesli likc appearance as in the resting nucleus. During these 

 processes the nuclcoli often two or more make their appearance 

 and the nuclear cavity increases as before (Fig. 6). 



Shortly aftcr the daugliter nucleus has atained the resting 

 stage (Fig. 6) a very uneven and interrupted spirem is developed. 

 It is often so narrow and incomplete that without the most care- 

 ful Observation it would not be noticed. This thread or band is 

 finally segmented into chromosomes. At first the chromosomes 

 are recognizable as very irregulär lumpy bodies, but soon lose 

 their irregularity and become somewhat longer. Presently by 

 bcnding the usual U-shaped chromosomes arise. (Figs. 1, 8.) 



Second division. 



Soon after the chromosomes are formed they become attached 

 to the spindle-fibres at the point of bending (Figs. 7 and 8). Their 

 orientation in this division is such that the equal drys of the U's stand 

 perpendicular to the long axis of the spindle. In this position it 

 is very evident that each consists of two segments resulting from 

 a longitudinal division. The chromosomes are now mostly in the 

 form of shallow U's though a few assume the shape of V's. They 

 are quite small and the regularity in size and form is striking. 

 Atkinson^) states that in Trillium the chromosomes of the pro- 

 phase of the second mitosis have the same general form as those 

 of the anaphase of the first division and that for this reason they 

 very likely preserve their identity through the short intervening 

 period. A transverve division, he says, further occurs at the apex 

 of the V's or U's giving rise to rod-shaped chromosomes as clairaed 

 by Belajeff in Ins paper already referred to. In Magnolia no 

 such behavior on the part of the chromosomes occurs, but as 

 above stated, their dissolution is complete so that in the resulting 

 resting daughter nuclei no trace of their identity is present. Con- 

 sequently a positive proof of a reduction division is out of the 

 question. 



While the daughter nucleus is still in the resting stage pro- 

 toplasmic radiations in great numbers are visible, ranning from 

 the nuclear membrane out into the surrounding cytoplasm and ex- 

 tending nearly to the plasma — membrane coll-wall (Fig. 6). 

 These radiating fibrcs do not seem to entcr the nuclear cavity at 

 this time. Finally they appear to mass themselves together in 

 several cone-like points ahout the nuclear membrane, which now 

 dissappears. These conical projections are soon connected by 

 many of the fibrcs that penetrate the nuclear cavity an so give 

 rise to the multipolar form of spindle. The number of points in 

 the multipolar spindle varied somewhat; but was always more 

 than two — three being the usual number. Often both of the 



*) Studies on reduction in phinte. (Botan. Gaz. Vol. XVTTI. p. 17. 

 July 18'J9.) 



