MUNSON: SPERMATOGENESIS OF PAPILIO. 
97 
cell becomes more and more elongated and compressed, it assumes 
an oval form (pi. 17, figs. 142, 143). 
Finally it leaves the nucleus and moves down along the axial fila¬ 
ment (pi. 17, figs. 140, 141). It is now a mass of granules suspended 
in a fluid, which collects at points along the axial filament giving rise 
to swellings or oval vacuoles at regular intervals (pi. 17, figs. 140, 141, 
143, 144, 145). In fresh living material of teased testis, these spherical 
vacuoles can be seen distributed at regular intervals along the lumen 
of the tail of the spermatozoon. In this mature state, they lbok like 
clear vacuoles. I take these to be temporary accumulations of cyto- 
lymph in the lumen of the tail surrounding the axial filament. The 
granules of the disintegrating nebenkern are evidently dissolved finally 
in this cytolymph and carried along with it through the cytocoel of 
the tail of the spermatozoon. 
In this respect, the nebenkern behaves very much like the yolk 
nuclei described by me in the egg of Clemmys. It shows many of 
the chemical and physical properties of those bodies. I came to the 
conclusion that the yolk nuclei in the oocyte of Clemmys represent 
a kind of metaplasm arising in the neighborhood of the nucleus, and 
by the influence of the nucleus and cytoplasm combined; that after 
flowing throughout the cytoplasm chiefly in the cytocoel, it is gradually 
absorbed by the living substance of the egg as food. The nebenkern 
originates essentially in the same way, but indirectly through the spin¬ 
dle, being partly of nuclear, partly of cytoplasmic origin. Its fate, too, 
is essentially the same, namely, absorption as food. 
In the larger vacuoles formed by the disintegrated nebenkern and 
combined wdth cytolymph, the separation of the axial filament into its 
constituent fibrils can sometimes be seen (pi. 17, fig. 141). I take 
this to be another proof of my assumption that the axial filament is 
not a mere fiber but a greatly compressed and attenuated reticulum 
belonging to the central portion of the original spermatid, but obscured 
in that stage of the cell by the substance of the nebenkern. That 
cytolymph may accumulate within the area corresponding to the 
cytocoel, is suggested by the appearances represented in plate 17, 
figures 125, 136, 137, 138, and 143, where a clear space is seen around 
the nebenkern. I do not believe that these are artefacts due to shrink¬ 
age, but that they are essentially similar to the vacuoles seen in the 
living material, along the axial filament of the spermatozoon. 
The centrosome .— During the earlier stages of transformation of 
