740 THOS. H. MONTGOMERY, JR. 
complicating the drawing, is represented as a line simply to show 
its position and length. A comparison of fig. 48 with fig. 42 
demonstrates that just after the pachytene condition the bivalents 
have the same number and form as just before it, another indica- 
tion that their univalent components had not fused in the pachy- 
tene stage. In the earlier portion of the strepsinema all six 
bivalents may be readily distinguished and their boundaries 
accurately determined, provided one select nuclei where they are 
scattered and not massed and then draw their outlines patiently 
with the camera lucida. In most of the cells of this stage in order 
to show their details with the greatest distinctness only a few of 
the gemini have been drawn. 
Generally the two components of a geminus appear of equal 
thickness and length, but occasionally there is to be noted a 
difference of their diameters as in the case of the right hand 
geminus of fig. 49. 
Shortly after the time of disappearance of the sphere in the 
cytoplasm the autosomes become for a while grouped around the 
plasmosome, as exhibited in figs. 58, 60, 61-66, fig. 61 showing 
the most common appearance of autosomes converging towards 
the plasmosome. Most of the figures are incorrect in one respect, 
namely, in omitting to show how the gemini end upon the plasmo- 
some, a relation difficult to determine because at this stage the 
plasmosome stains deeply, but they do not appear to penetrate 
into the plasmosome substance. The meaning of this phenome- 
non, which is a constant one for this period, will be discussed 
under the subject of the plasmosome. 
The later history of the autosomes shows that the gemini 
widen out in such a manner that the two univalent components 
of each continue to be connected at one end while diverging at 
the other, the end result being, in the definitive gemini of the 
reduction division, every two univalents joined end to end. 
This process is effected gradually, it does not take place in all 
cells at the same time nor yet at the same time on all gemini of 
the same nucleus, and in its details shows considerable variation. 
Thus in figs. 50, 53, 56 the univalents are more widely divergent 
than in fig. 59, though the latter is a later stage of the cytoplasmic 
