Andrews, Karyokinesis in Magnolia and Liriodendron. 157 



lipon tliemselves directly and these give rise to U's. Not infre- 

 quently, hoM'ever, tliis U-shaped appearance occurs before any 

 Splitting is seen and the division then results directly in tvvo 

 U-shaped chromosomes. In other instances the chromosomes begin 

 to split at tbe center, while the ends remain attached to one 

 another and the central parts separate until the appearance of 

 closed rings is produced. Again open rings are formed in the 

 way just described, by the ends of the newly formed chromosomes 

 separating slightly from one another. The same idea would be 

 given if two of the U-shaped chromosomes or crescents should be 

 placed with the openings together. 



The number of chromosomes formed in the wav above men- 

 tioned is unusually large^ for I counted as many as forty-eight 

 (48) in a single polar view. In Figure 2 about twenty are shown, 

 but this represents only one section of the nucleus. The nuclei 

 were all well fixed and thoroughly infiltrated so that the chromo- 

 somes remained in position. Thus the entire number may be ob- 

 tained by counting those in each section. The chromosomes which 

 are at first small lumps (Fig. 1) become eventually short thick 

 U shaped bodies (Fig. 2). At almost regulär intervals along their 

 length are conspicuous enlargements having a spiny appearance. 

 The regularity of their enlargements is striking in that they are 

 generally three in number, one terminating each end of the chro- 

 mosome and one at the bottom of the shallow U. The arrange- 

 ment of these enlargements usually increase the difticulty of 

 correct Observation when the chromosome is seen from the open 

 end of the U; for the terminals are often so wide as to almost or 

 completely close the entrance to the depression in the U. It is 

 then only from a lateral view that the true form of these chromo- 

 somes in such cases can be ascertained. In addition to the 

 U-shaped forms the chromosomes may form open or closed rings 

 (Fig. 2). Both forms occur side by side in the nucleus, and pre- 

 sent the same appearance when arranged in the nuclear plate. 



Formation of the spindle and orientation of the 

 chromosomes in the äquatorial plate. 



The nuclear membrane disappeared (Fig. 3) in the earlier 

 stages and with it the circular weft of fibres which give rise at 

 least, to the greater part of the spindle. The spindle arises as a 

 multipolar structure which is gradually transformed into the bi- 

 polar type. The number of poles in the young spindles varies 

 somewhat, but usually is three or four and these are always drawn 

 out into very long sharp points. The poles observed no regularity 

 so far as direction is concerned but were found projecting at any 

 angle. In not a single instance of the many hundreds of spindles ob- 

 served and studied was a single spindle bipolar from the beginnig. 

 Finally by a withdrawing and fusion into two groups of the seve- 

 ral poles, the above mentioned bipolar form always resulted. The 

 two poles of the mature spindle are in nearly every instance very 

 broad, (Figs. 3, 4, 5) but later become more or less pointed. 



