14 DUBLIN 



are very nearly in the form in which they appear in the ensuing 

 division (Figs, 28-31). The cells are perfectly spherical with 

 the cytoplasm reduced to a very thin ring around the much 

 enlarged nucleus. In the cytoplasm, I have often observed a 

 deep staining granule, throughout the growth period, which 

 very probably is the centrosome. This rapidly divides, the 

 nucleus elongates and the prophase of the first maturation 

 division is attained. The types of chromosomes, viz., the rings 

 of various form and the double elongated bars are at this stage 

 found lying irregularly over the whole spindle (Figs. 32-34). 

 From these figures, two important conditions are clearly pre- 

 sented. In the first place, the longitudinal split shows distinctly 

 in many of the bars running throughout their length especially 

 in Fig. 34 and second, the several chromosomes, both bars and 

 rings, present remarkable size differences, some being as much 

 as three or four times the bulk of others. 



The chromosomes now move uniformly into the equatorial 

 plate, givmg in the metaphase pictures of striking clearness 

 (Figs. 35-40). There are at this point fewer rings in propor- 

 tion to the bars and these that still persist are much more 

 slender, showing that they are being converted, through elon- 

 gation, into the chromosomes of the other type. Finally, in a 

 considerable number of spindles, obviously of a later metaphase, 

 the bars alone exist and the time for division is at hand (Fig. 

 41). It is hardly necessary, at this point, to discuss the nature 

 of this division process. From the evidence presented, it has 

 been fully demonstrated, first, that the bar and ring-figures are 

 structurally tetrads, i. e., composed of four portions, a, b, a, h, 

 each of which is a quadrant (Text-fig. Ij/and^), and second, 

 that the split running in the long axis of the figures is a longitudi- 

 nal one separating the sister bivalents. The extremities of the 

 figures are then the longitudinal ends, while those intermediate 

 between these, are the points of synapsis. It is obvious that 

 the long axis of the bars lies in the long axis of the spindle, 

 and that the points of synapsis are in the plane of the division. 

 This division is, therefore, transverse or reducing. On this 

 point there can be no room for doubt. 



