12 DUBLIN 



the growth period. In cells of this small size, it is quite impos- 

 sible to trace anything corresponding to constant size differences 

 in the chromosomes. Yet it is clear both from the study of 

 the various spermatogonia and the spermatocytes that many 

 gradations actually exist. 



A longitudinal splitting of the arms of the V's now makes 

 itself apparent (Fig. 21). Appearing first at the lower end of 

 one of these larger bivalents, the split makes its way upward to 

 the apex, w^ien the other arm becomes involved. Soon most 

 of the chromosomes show the same condition, and it appears 

 very much as though the V's had each split longitudinally along 

 its whole length, the two sister portions remaining united at the 

 apices (Fig. 22). The process is, however, a very gradual one, 

 and all degrees of variation in the time of its occurrence are 

 present. The chromosomes at this point stain much less in- 

 tensely and are granular. Indeed, the cells in this stage can be 

 most readily distinguished without any closer study by their 

 lightly staining nuclei. 



At the completion of the longitudinal splitting (Fig. 22) the 

 chromosomes are at their maximum length. The nucleus is 

 comparatively small and, as a consequence, the individual 

 chromosomes are much crowded and cross and recross each 

 other. This coupled with their granular appearance, and their 

 slight staining capacity, makes it extremely difficult to follow 

 the processes in which they are involved. I have succeeded, 

 however, in finding a considerable number of nuclei in which 

 this crowding is not so marked, and where, in consequence, the 

 more careful study of the individuals was possible. From such 

 cells (Fig. 23), it is quite clear that the chromosomes, now 

 longitudinally split along their entire length, are passing through 

 marked changes. The acute angle of the V's gradually opens 

 out, until at a later period of development, the chromosomes 

 entirely lose their original form, and become pairs of sinuous or 

 nearly straight parallel rods. Fig. 23 shows this process in 

 all its stages (cf. Text-fig. I, p. 1 3). To the extreme left, a large 

 chromosome shows but a slight opening of the angle, while in 

 the center the very end of the process is already attained, i. e., 



