738 THOS. H. MONTGOMERY, JR. 
lishes, together with the other evidence that is to follow, that we 
are dealing here with parallel conjugation and not with longitu- 
dinal splitting. For there being in fig. 34 just ten threads, two 
of which are double, and the spermatogonial number of auto- 
somes being twelve, it necessarily follows that each of the double 
threads could have been produced only by juxtaposition of two 
single threads. In fig. 38 the pairing of certain autosomes is 
seen clearly. In fig. 42 where the entire nuclear content is shown, 
there are seen exactly six chromatin loops (excluding the pair of 
idiochromosomes, D, d), and each of these is double; these are 
then six bivalent elements, each produced by the lateral apposi- 
tion of two univalents. In fig. 41 five entire double loops are 
drawn, the sixth being omitted so as not to obscure the picture. 
Thus it results that the six bivalent autosomes or gemini of figs. 
41 and 42, and consequently of later stages, are produced by 
lateral conjugation of univalent autosomes; and the clear line 
that one sees along the axis of each such double thread cannot be 
a longitudinal split of a single univalent. 
The stages delineated in figs. 34 to 42 have been arranged 
according to the stage of cytoplasmic differentiation, particularly 
the growth of the sphere. This cytoplasmic development does 
not keep exact step with the nuclear, for, e.g., the cell of fig. 39 
contains exactly eleven autosome loops, therefore cannot contain 
more than one bivalent, and with regard to its nuclear compo- 
nents should precede fig. 34 that contains two bivalents. 
The zygotene condition is closely followed by the pachytene 
(figs. 48 to 45), in none of which are all the gemini represented ; 
this stage is found just before the sphere (Sp) has reached its 
greatest size. Here the autosomal loops are in one-half the nor- 
mal number, and, for the most part, each of them appears solid 
and undivided. But in the greater number of cases there is to 
be seen on each geminus at least traces of a clear space which 
marks the original point of meeting of two univalents. Thus in 
Kuschistus there is no valid evidence of any actual fusion of the 
two members of any pair. 
It will be convenient to stop at this point to recount another 
process of the autosomes. In the earliest growth period (figs. 
