136 
Thos. H. Montgomery, jr. 
iiumerary (x, fig. 55), agaiii a total of 8. 3. The normal condition "svith 
the addition of 2 siipernumeraries, a total of 9 (fig. 54). In the second 
s])ermatocytes the smaller diplosome can always be recognized by its 
smallcst size; and the supernnnieraries niay or may not be attached to 
others. 
In the anaphases of sonie but not all second spermatocytes the con- 
dition of fig. 56 is foimd, onc or tvvo of the chromosomes of each danghter 
plate being drawn out into a long thread parallel with the connective 
fibres; perhaps these represent chromosomes that had behaved abnor- 
mally in preceding stiiges. They niove more slowly towards the spindle 
poles tlian the others do, but are never left at the equator. Such a con- 
dition I have never seen in other testes of this species. 
In the few instances where the chromosomes could be counted in 
the spermatids, 7 were found in 6 cases, and 8 in 3 cases^). 
IV. Conclusions on the Chromosomal Variation. 
1. In the first kind of chromosomal Variation deterinined thenumber 
of the univalent elements is constant and the same in the spermatogonia 
and first spermatocytes. In 41 spermatogonia the number found was 
exactly 14; a single case exhibited only 13, which is certainly due to the 
smaller diplosome being hidden and not to any Variation in number. 
This Variation consists in abnormal Separation of the univalent compo- 
nents of a normally bivalent chromosome, or in much rarer cases of the 
abnormal Separation of the univalent components of two normally bi- 
valent chromosomes. Whether this Variation affects only a particular 
autosome (ordinary chromosome) was not deterinined. Such separated 
univalent components, the s-chromosomes, divide equationally in the fh'st 
maturation mitosis, and in the following mitosis separate reductionally 
usually after conjugation in the plane of the equator of the spindle, only 
in a few cases do they fail to conjugate at that time. As a result of this 
kind of Variation all the spermatids receive the same number of chromo- 
somes. This is then a Variation in the number of the bivalent chromo- 
somes of the first spermatocytes, and though there may be 8 (the norm) 
or 9 or 10 separate chromosome masses in the fh’st maturation spindle, 
tlie total number of univalent elements there is always 14 as in the sper- 
matogonia. This Variation does not follow any law of change, it is in no 
1) In 22 spermatogonia of E. irisligmus exactly 14 cliromosomes were found, 
and in 1 only 13. 
