70 
H. E. Jordan 
point. What proof have we that the granules of the spermatogoiiia 
stained with Benda’s stain are real mitochondria ? (ef. H. Heidenhain, 
1911, p. 1080 and 1088). There raay be still otlier cytoplasmic granules 
besides yolk and mitochondria. Perhaps Benda’s stain does not differen- 
tiate between certaiu food granules and mitochondria, just as iron-haema- 
toxvlin does not, or only poorly, between yolk and mitochondria. It will 
be noted, however, that my identification of mitochondria rests as much 
upon morpliological evidence as upon the iron-haematoxylin staining 
reaction. The point liere is that Benda’s stain does not appear to be 
any more specific for typical mitochondria than iron-haematoxylin. 
Again, the loss of a considerable amount of mitochondrial substance 
in the cast-off portion of the original spermatid — a product analogous 
to the “cytoplasmic vesicle” described by Meves (1899) in the guinea-pig 
and illustrated also by Duesberg (1910) — during the process of meta- 
morpliosis, militates against the interpretation of mitochondi'ial conti- 
nuity and a hereditary röle of this substance. Tliis loss is probably what 
Retzius refers to when he speaks of the »bedeutende Verminderung ihrer 
Substance, als sie in die Bildung der Spiralfaser eingehen« — a Statement 
which Duesberg says he is unable to comprehend, consequently does 
not consider in the light of his tlieory. 
The above detailed evidence strongly indicates a nuclear origin for 
the mitochondria in the opossum, and furnishes a number of facts which 
do not seem to conform with the attractive theory of mitochondrial con- 
tinuity. 
Spermatozoon. 
Summary. The mature Spermatozoon consists of a liead, body 
with middle-piece (text fig. 1), and flagellum. The head is horse-shoe- 
shaped and represents the spermatid nucleus very compact, chromatic 
and split from before backwards, possibly by inpushing of the V ebenkern. 
The head forms the rim of the funnel-shaped body (open anteriorly, the 
opening representing the cephalic end of the V-shaped ventral space in the 
“funnel”). The latter is thicker clorsallv and contains the intermediate- 
or middle-piece. The body represents a portion of the original cyto- 
plasm of the spermatid (i. e. that portion included within the lateral 
limits of the caudally migrating Nebenkern — figs. 71 and 72), a consi- 
derable portion with mitochondria bring discarded, and another con- 
siderable portion going to form the flagellum or tail. The origin of the 
middle-piece in toto is obscure. The Nebenkern seems to persist in part 
in the anterior portion (growing progressively smaller — forming possibly 
