174 Edmund B. Wilson 



appearance which I at first thought must be due to an artificial 

 clumping together of the chromosomes through defective fixation; 

 but the study of very many of these figures convinced me that such 

 is not the case. As in the preceding types, ten of the chromosomes, 

 including the /7z-chromosomes, have the form of symmetrical 

 dumb-bell shaped bodies w^hich are equally halved in the ensuing 

 division. The remaining chromosomes are usually aggregated to 

 form a compound element (Fig. lo, h-l, Photos 22, 23) in which 

 may be very clearly distinguished the same components as those 

 that appear in the chromosome-nucleoli of the growth-period; 

 and the size-relations make it evident that one of them is the large 

 idiochromosome, one the small, while four are the supernumer- 

 aries. In other words, these six chromosomes, which divide as 

 separate univalents in the first division, have now again conju- 

 gated to form a hexad group. This compound element almost 

 always lies near the center of the group. Polar views of this divi- 

 sion accordingly show typically 1 1 chromosomes, of which the 

 central one is compound (Figs. 10, a~g, Photo 13). Not infre- 

 quently, however, one or more of the supernumeraries may be sep- 

 arate from the others (Fig. 10, /, g), the apparent number in polar 

 view varying accordingly. 



In side views the grouping of the components of the hexad 

 element is seen to vary considerably though the large idiochro- 

 mosome is more frequently at one end of the group. In the ensu- 

 ing division the other ten chromosomes divide equally, while the 

 hexad element breaks apart into two groups that pass to opposite 

 poles (Fig. 10, l-p). The distribution of the various elements 

 is difficult to determine exactly, since they always lag behind the 

 others in the anaphases and are scattered along the spindle in such 

 a way as often to give confusing pictures. The study of many such 

 anaphases leads me to conclude, however, that at least one of the 

 smaller components always passes to the opposite pole from the 

 larger one, while the other four undergo a variable distribution. 

 In Fig. 10, /, the group is just separating into three toward each 

 pole; in 10 m, it is quite clear that three of the small ones are pass- 

 ing to one pole, while the large one and two small ones are passing 

 to the other, and Fig. 10, «, is probably a similar case. In these 



