74 University of California PubUcafions in Zoology [Vol. 20 



chromosome will be found near or attached to the end of another, 

 though the opposite ends may be widely separated. This is shown 

 more clearly in the diagrammatic scheme in figure D, where one- 

 half only of the actual number of chromosomes has been drawn in 

 each nucleus. Each of these groups of two chromosomes is probably 

 formed by the splitting of a single thread in the earlier stage 

 (fig. D, 2). 



Following this, each chromosome becomes looped or doubled upon 

 itself (fig. D, 5). In this stage also it is found that there is some 

 suggestion of a grouping of the ehromcsomes in pairs (pi. 9, figs. 46, 

 49), though the ends may be more widely separated than in the pre- 

 vious stage. Following this the chromosomes condense into a tangled 

 mass of threads or contraction stage (fig. D, 6; pi. 9, figs. 42, 43, 

 45, 50), from which emerge twenty-six looped or V-shaped chromo- 

 somes (fig. D, 7; pi. 9, fig. 51a; pi. 10, figs. 52, 53). What happens 

 in this contracted condition can only be conjectured, since thus far 

 the details have escaped detection. 



When the chromosomes divide in the metaphase the point of 

 separation occurs at the apex of the V or loop. If division here is 

 longitudinal, and the evidences of the earlier stages (fig. D) confirm 

 this view, the line of separation must be at one end of the original 

 split found in the chromosomes when first formed. This would postu- 

 late the supposition that the two halves of the original chromosome 

 are reunited at the time the change in the number of units from 

 fifty-two to twenty-six occurs. To illustrate this, let us follow the 

 course of the chromosomes marked A in figure D. In figure 2 the 

 chromosome A is a single thread that has partially split. This is com- 

 pleted in figure 4 but the ends of the chromosomes are still connected. 

 In figure 3 this connection has apparently been lost, for the ends 

 are separated bj- a considerable space. The course of these two threads 

 is lost in the contraction stage in figure 6 but they emerge again in 

 figure 7 as a single split chromosome which parts "transversely" on 

 the spindle in the following figure, the two threads having apparently 

 become united end to end into one, in the stage represented by figure 6. 

 The seemingly transverse division is in reality the final step in longi- 

 tudinal splitting of the original chromosome. 



The onlj^ difficulty in this explanation lies in the lack of a visible 

 mechanism by means of which two threads which are more or less 

 widely separated from each other, are reunited, with their relative 

 positions those of the original split. This difficulty is, however, not 



