SPERMATOGENESIS OF EUSCHISTUS 741 
structures; fig. 59 exhibits five entire gemini, and by a heavy line 
the position of the sixth. The two univalents of each geminus, 
as they begin to unravel, appear at first spirally twisted around 
each other (figs. 48-57). Then they begin to diverge at one 
end. The point of continuing contact of the two univalents of a 
geminus, or point of persisting conjugation, is marked by « in 
figs. 58, 60, 62. 
No rest stage follows this condition, contrary to my first deserip- 
tion, but the gemini immediately approach the nuclear membrane 
and the prophases proper of the first maturation mitosis are 
established. Fig. 63 marks the beginning of the prophase, and 
later prophases are shown in figs. 64-86. The concurrent changes 
of the gemini are theoretically so important that they deserve 
to be described in detail. In fig. 68, which presents the greater 
portions of five gemini, the conjugation point of each geminus.is 
marked by the letter x. Figs. 64 and 64a show five entire gemini 
of one nucleus, the sixth being omitted because it lay in an oblique 
position; this shows various stages of gemini, the one of fig. 64a 
illustrating particularly clearly how the univalents remain 
attached at one point and separate at the other. The conjuga- 
tion points, marked by x, can be seen in some of the gemini of 
figs. 65, 66, the latter exhibiting all the gemini, but in both of these 
the elements lie so obliquely as to obscure the relation. As the 
prophase advances the process becomes clearer, for the autosomes 
gradually become shorter and more regularinform. Thus in fig. 
67 two whole gemini are drawn, each of which clearly possesses a 
V-form; the same condition is shown somewhat less clearly in figs. 
69 and 70; in figs. 71-84 also, the letter 2 denotes the conjugation 
point. Figs. 71 and 72 are from one cyst, the latter showing two 
gemini. Fig. 73 illustrates a part of a geminus on the right and a 
whole one on the left, the latter exhibiting clearly the mode of 
4 Stages like those of figs. 57-62 are in the large spermatocytes of follicles 1 and 
3 very much more like rest stages, for in them the autosomes become more diffuse 
and their boundaries more difficult to determine; it was from a study of those large 
cells that I was led, in 1898, to conclude the occurrence of arest stage at this period. 
It will be recalled that in the present paper the conditions in follicles 1 and 3 are 
disregarded. 
