132 
Tlios. H. ilontgomer}’, jr. 
was probably due to obscuration of tlie smallcst eleinent, the smaller diplo- 
some ; there is liardly a cliauce of this small eleinent being actually lacking 
from this cell, the probability is that itismerclyhiddenbyanotherchromo- 
sonie, for in all the other polar views of spermatogonia {d, figs. 8, 16, 33) 
it was recognizable as well as in all fhst and second sperniatocytes. 
But while 89 first sperniatocytes showed each the normal number 
of 8 elements (figs. 2, 10, 11, 30, 34), 12 of these cells exhibited each 9 se- 
parated elements (figs, 1, 3, 18), IVhen one compares the latter cases 
with the former it is found that the higher number 9 is due to the Separa- 
tion of the two components of one of the normally bivalent autosomes. 
F or in the normal cases, with 8 elements, no two of the chroniosomes are 
ever exactly equal in volume but the scries of theni jiresents a ränge of 
sizes. IVhen there are 9 elements, two of them always resemble each other 
in size (s, figs. 1, 3). This Variation is always just 1 niore than the normal 
number, never 1 less, and is always associated vith the phenomenon of 
2 of equal size. In these abnormal cases the separated univalent chromo- 
sonies divide separately and equationally, just as the diplosomes do, the 
jiroof of which is as follows. The second maturation mitosis follows inime- 
diately upon the fh'st without any change in consistency of the chronio- 
somes, consequently there is no })ossibility of any recombination of par- 
ticles from different chroniosomes. Polar views of each daughter plate of 
the aiiaphase of the first maturation division show normally 8 separate 
elements (fig. 21), which are 6 constricted (longitudinally split) autosomes, 
and 2 still sejiarated but non-constricted diplosomes {d, e) of very different 
volunies. All of these 8 are clearly the daughter of the 8 of the preceding 
equatorial plate, and the mode of division of them was detailed in my 
earlier papers. But occasionally 9 separate elements are seen in a daughter 
])late of the aiiaphase of the first maturation (fig. 20, 19; in the latter 
figure only the lower half should be considered, for in the upper cell the 
chroniosomes were so densely inassed that only five of them could be 
seen distinctly). Aow when we compare figs. 19 and 20 with fig. 21 we 
find in all a larger and a smaller non-constricted diplosome {d, e), but 
while fig. 21 shows 6 constricted autosomes, figs. 19 and 20 exhibit each 
ö constricted autosomes and 2 non-constricted ones (s, s) of equal volunies. 
These 2 non-constricted autosomes of equal volunies of figs. 19 and 20 
are then clearly the daughters of the 2 univalent autosomes of equal vo- 
lunies of the first sperniatocyte (s, figs. 1, 3). These elements of the ab- 
mosomes and bcliove that my previous supposition was correct, that this number is 
due to precocious division of two of them. 
