The Spermatogenesis of the Opossum (Didelphys virginiana) etc. 69 
verv young spermatocytes, as they seem to be in the opossum and in 
many other cases reported; and their origin in Sertoli cells and sperma- 
togonia may be the same (i. e. nnclear) as in the primary spermatocytes. 
I beheve that the fact of their apparent absence in the yourg spermatocytes 
of the opossum (and possibly other forms) is one of the strongest argu- 
ments against the Benda-Meves-Duesberg theory of theh continuity 
and hereditary significance. If they are simply undifferentiated integral 
elements of the cytoplasm carried unchanged into the spermatozoa why 
should they have disappeared at a particular stage (or even decreased 
in number, or changed chemicallv) represented by a cell treated similarly 
to cells which clearly show them? 
It would seem that even a poor technique that, however, rcveals 
them clearly and typically at a certain stage and subsequently should 
reveal them at every stage (when present) represented by cells in the 
same tissue. Their absence at a particular phase and their sudden appea- 
rance in great quantity at a definite stage must be explained on other 
grounds than poor fixation, or inappropiate staining methods. I must 
emphasize the fact here that the bodies described are true mitochondria, 
for they can easily be traced into the spiral filament of the middle-piece 
even in unstained preparations. The mitochondria of the later spermatids 
are for the most part (some have become vesicular — abundant transition 
stages leave no doubt regarding the identity and origin of the latter) exactly 
similar to those of the prophase of the spermatocyte, and meanwhile 
they have greatly increased in number in spite of the enlargement of the 
spermatid. I cannot satisfy myself as to the mode of increase but there 
is some evidence that chromidia are again formed (though only in small 
numbers) in the resting stage of the secondary spermatocytes and early 
spermatids. Of course I cannot be certain that the chromidia metamor- 
phose into mitochromidia. But the evidence points strongly that way, 
and if the identity is admitted I see no escape from the conclusion that 
the mitochondria arise in the nucleus as urged by Hertwig and Gold- 
schmidt and others. It is of course possible that we are dealing here 
with two different substances. The evidence of Scott and of Dolly, as 
above stated, indicating a nnclear (chromidia!) origin for tigroid substance, 
and that of Meves and Duesberg indicating a mitochondrial origin of 
neurofibrillae, suggests such a probability 9 ). Nor can the objection to 
iron-haematoxylin (in preference to Benda’s supposed specific mito- 
chondrial stain) in this connection be accorded much weight. For whv 
in this instance should one put more reliance in one method that another? 
Iron-haematoxylin reveals true mitochondria at all stages after a certain 
