The Spermatogenesis of the Opossum (Didelphys virginiana) etc. 
59 
determines the relative numerical abundance of spheres and vesicles. 
Only the spheres and granules appear to contribute to the formation 
of the spiral filament of the intermediate-piece of the spemiatozoon. The 
vesicles probably represent liquifying spheres and granules and disappear 
within the cytoplasm forming the flagellum. Some are cast out with 
a discarded portion of the cell. 
Düring the maturation mitoses, the mitochondria seem to be ap- 
proximately equaUy distributed to the daughter cells. There appears to 
be no coincident fission 7 ) of individual rods or granules such as Duesberg 
(1910) describes for chick and Faure-Fremiet (1910) for certain protozoa; 
nor is there ever an appearance of filaments (chondriomites) until the 
spiral filament is formed. One gets the impression of a slight augmentation 
in the number of mitochondria, but only during the resting stage of the 
secondary spermatocytes is any tangible clue given as to the source of the 
increase. Here a few basophile granules (morphologically and tinctorially 
similar to the mitochondria) seem to pass from nucleus to cytoplasm; 
likewise in the early spermatid stage. This Observation however, has 
significance only if we can establish the primary nuclear origin of mito- 
chondria (as cliromidia), a point which will be further discussed below. 
Both spennatogonia and Sertoü cells contain a very few granules 
very similar to the later mitochondria. Both of these types of cells contain 
considerable deutoplasm. It is very difficult here to distinguish between 
the fine spherules of fat and the mitochondria. In unstained preparations 
of FLEMMiNG-fixed material, however, fat is deep black ; the mitochondria, 
grayish (metallic) black. On the basis of this differentiation and on 
morphological grounds, I beheve that we are dealing with both a few 
mitochondria and much deutoplasm in the spennatogonia and the 
Sertoh cells. 
But my preparations leave no doubt respecting the absence of mito- 
chondria during the early growth period of the primary spermatocytes. 
For this generation of cells they first appear during the later growth period 
— and during a period coincident with a transitory achromatic reticular 
phase of the nucleus. This Observation is the more significant in view 
of the fact that both within and without the nuclear wall are similar 
darker-staining bodies. Subsequently such bodies (now deeply-staining, 
sharply-contoured spheres and dumb-bells) are aligned on the nuclear 
membrane externally. All the evidence here points to a nuclear origin 
of mitochondria i. e. they appear to be transformed chromidia, The 
bearing of this on the several theories regarding the origin and function 
of mitochondria will be pointed out below. 
