60 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
trace of an enclosing membrane. The central portion of this mass 
is more compact and the nuclei much less distinct. Indeed the nuclei 
in the central portion seem more or less disorganized, appearing more 
like a granular central portion with nuclei at the periphery. 
Into this mass are inserted the heads of a bundle of fully formed 
spermatozoa. I am unable to explain this except on either of two 
suppositions. It seems to be a case where only a few of the spermatids 
have developed into spermatozoa, the remainder being gradually 
absorbed as food. Possibly this may be due to a failure of the cyst 
cells to develop properly, which is cpiite evident. The central solid 
portion may be the head nurse-cell, which has become imbedded in 
the mass of undeveloped spermatids. 
I have thought that possibly also a number of mature spermatozoa, 
having been discharged from the cyst, but not finding their way into 
the vas deferens, have taken possession of an undeveloped gonocyst or 
cytocyst. 
Fate oj the spermatocyst .— When the spermatozoa are mature, the 
spermatocyst is discharged into the vas deferens. But just before 
this,, the cyst cells seem to dissolve, allowing, the spermatozoa to escape 
either singly (pi. 1.4, fig. 51), or collectively (pi. 14, fig. 53). Rem¬ 
nants of the cysts are found near the entrance to the vas deferens often 
crowded into masses of debris between the stalked cells (pi. 12, fig. 5). 
Traces of cyst cells are also found in the vas deferens, but usually 
only parts of the head nurse-cell to which the heads of the spermatozoa 
are still attached (pi. .14, fig. 55). I have, never found a complete 
spermatocyst in the vas deferens, though the spermatozoa frequently 
occur adhering in bundles (pi. 14^ fig. 55), as they are also found 
within the testis after being discharged from the cyst (pi. 14, fig. 53; ' 
pi. 12, fig. 5). 
History of Individual Cells. 
That part of the vas deferens which penetrates the testis, and 
which in Grapta silenus divides into four branches, one for each 
follicle, is lined with a modified epithelium differing in some respects 
from that lining the vas deferens proper. The cells are more spherical 
and the nuclei are unusually large, spherical, and very rich in chromatin 
(pi. 12, figs. 5, 8b, v. d.\. 
Stalked cells .—Apparently connected with the epithelial cells de- 
