AND OF HYBRID PIGEONS. 
39 
been lost in the preceding division. In two or three instances 
in the division of the primary spermatocytes, one of the large 
chromosomes was observed to be made up of four small rings 
or vesicles (Fig. 35). It is possible that this indicates a de¬ 
marcation into tetrads, visible in hybrid material because the 
firmer fusion which occurs normally at this stage is lacking. 
These irregular distributions of chromatin seem to affect 
^he size of the spermatozoa in the hybrid. There is apparently 
a greater variaion in the size of the spermatozoa than in normal 
pigeons, although the evidence is not as conclusive as is desir¬ 
able, because killing reagents do not affect all spermatozoa 
similarly. However, where the same methods have been used 
on the spermatozoa of hybrid and of normal birds, there is cer¬ 
tainly much more variation in length among the former. It 
was not uncommon to find giant forms three or four times the 
size of the ordinary ones, as well as unusually small ones. This 
extreme variation in the size of the spermatozoa was seen, how¬ 
ever, only in hybrids from very different species, and such birds 
are, for the most part, sterile. 
The objection arises, of course, that it is questionable if ir¬ 
regular mitoses result in the formation of spermatoza. When 
we take into account, however, the great frequency with which 
these irregularities occur in some hybrids, and the (fact that 
enormous numbers of spermatozoa may, nevertheless, be 
formed, it seems certain that many of them do mature into sper¬ 
matozoa. The irregularity in the formation of the spindle, 
moreover, does not indicate necessarily that the succeeding 
cells will be abnormal. It simply means, as we shall see, that in 
all probability there is an incompatibility between the chromo¬ 
somes from the two different species which, in some cases, pre¬ 
vents them from uniting for the pseudo-reduction that normally 
occurs at this time. In the other primary spermatocytes the 
same tendency must be present, and even if pairing of the 
chromosomes does occur, the succeeding division doubtless 
severs the, two plasma again and sets them apart in different 
cells, just as would be the case in many of the cells which, with 
their double spindles, visibly manifest the antagonism. From 
the division of a tripolar spindle three cells will result, of which 
the one containing the end of both spindles will probably re- 
