88 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
plainly distinguishable from one another, and each appears to be com- 
posed ofa much convoluted and tahgled chromatin thread, which is 
bound into a more or less compact mass by linin fibrils. Although, 
it is impossible to count these masses accurately owing to the great 
difference in their size and to the fact that a cross section of all cannot 
be obtained in the same plane, their number is undoubtedly approxi- 
mately that of the autosomes of the spermatogonia. Moreover, the 
masses of chromatin have approximately the same orientation that 
the autosomes had during the preceding telophase of the last sperma- 
togonial division. For these reasons I believe we are justified in con- 
cluding that each of these masses represents a univalent autosome. 
Later, each mass becomes converted into a single chromatin thread 
by a sort of unraveling process. In Figures 43 and 44 (Plate 3) four 
of these masses are drawn in detail. ‘This is a slightly later stage than 
that shown in Figure 41, and the method of formation of the chroma- 
tin thread is well shown. If each of these masses represents, as I 
believe, a univalent chromosome, then it is evident that during this 
stage each chromosome becomes converted into a long thread com- 
posed of a single row of chromatin granules imbedded in a linin matrix. 
In the following stage (c) these threads become evenly distributed 
through the nudeus, giving the appearance of a very tortuous but 
continuous spireme, and they have been so interpreted by most investi- 
gators. ‘There is, however, no evidence that such is the case. In the 
next stage, which rapidly follows this, the spireme, Just as in Dissos- 
teira, is made up of loops the ends of which are attached to the nuclear 
membrane at the distal pole. Since the number of loops even at this 
early stage is, I believe, only one-half that of the autosomes in the 
spermatogonia, it is probable that during stage c, or earlier, the chro- 
matin threads unite in pairs end to end to form loops, while the oppo- 
site ends become attached to the nuclear membrane at its distal pole. 
However, it is impossible to say precisely when the conjugation of the . 
autosomes takes place. Possibly it may occur as early as stage 5), 
as I have in a few cases seen structures which might be interpreted 
as the result of the fusion of two chromatin masses end to end; but 
as such structures may be only accidental, I have not thought it best 
to attach any weight to them. 
2. Monosome. 
The history of the monosome during the growth period is practically 
the same as in Dissosteria. 
