The Spermatogenesis of the Opossum (Didelphys virginiana) etc. 
43 
latter disappears prior to mitosis, leaving a plastin remnant, by process 
of extrusion of chromatic buds which disperse throughout the nuclear 
reticulum. There is here another instance favoring the position that the 
nucleolus functions in part as a store-house of chromatin, which contri- 
butes at mitosis to the formation of chromosomes. However, I can 
identify no structure at this early stage as the future accessory chromo- 
some. 
Metaphase plates of dividing spermatogonia contain 17 chromosomes. 
These vary somewhat in size, are rod-shaped and variously curved (figs. 1, 
3, 4, 5 and 6). One usually appears somewhat larger than the rest, and 
may represent the monosome or future accessory chromosome. I am 
unable, though, to establish a direct continuity between a spermato- 
gonial chromosome and the chromatin nucleolus of the primary sperma- 
tocyte, such as Wilson has so clearly demonstrated for Lygaeus (1905 b) 
and Pyrrhocoris (1909 a), Davis for Dissosteira carolina (1908) and as I 
was able to show in Aplopus mayeri (1907). In metakinesis the behavior 
of all the spermatogonia! chromosomes appears similar. 
IV. Interstitial Cells. 
Equatorial plates of dividing yonng interstitial cells (from the adult 
testis) also yield a chromosome count of 17 (fig. 7). Further description 
of these interesting cells, which are here very abundant and very large 
and contain mitochondria, will be reserved for a separate paper. 
V. Primary Spermatocyte (Auxocyte). Growth Period. 
a) Resting phase (interkinesis). The chromatin (chromosome) 
nucleolus. 
The resting primary spermatocyte (meiotic cell — Moore and Walker) 
is considerably smaller than the flat elongate spermatogonium. The 
nucleus contains several nucleoli (usually three, figs. 8, 9 and 10). With 
iron-haematoxylin all stain intensely. One, however is usually consi- 
derably larger and has a sharper contour (fig. 9). Moreover, this parti- 
cular nucleolus is frequently located close to the nuclear wall at a point 
corresponding to the position of the centrosphere (idiozome, Meves, 
fig. 10). However, this is not an invariable condition as is conspicuously 
the case in later stages. Possibly not sufficient time had elapsed in all 
cases for the attractive influence of the centrosphere to have prodnced 
the final effect in non-conforming cells. The inference seems justified 
