72 THE AMERICAN NATURALIST. (Vor. XXXIX. 
therefore say that in all cases in which either both sexes, or even 
one sex, invariably develop from fertilized eggs, every centro- 
some (even that of the egg, which is destined to disappear) is a 
descendant of a sperm-centrosome — a male centrosome, if I 
may so call it. There is apparently a difference between this 
and pure parthenogenesis, where every centrosome is a descend- 
ant of the egg-centrosome and could be called a female centro- 
some. Does such a difference in reality exist? If we believe 
with Wedekind (:02) that parthenogenesis is an original, pri- 
mary mode of development, we must admit that it does. But 
this would mean that animals which are closely related to each 
other differ in their phylogenetic origin. Is it not simpler to 
admit that even in pure parthenogenesis the ostensibly female 
centrosomes are descendants of a remote ancestral male centro- 
some? It seems, indeed, as though in animals where fertiliza- 
tion occurs only as a rare exception the males had gradually 
disappeared yielding to hermaphroditism and parthenogenesis, 
as Maupas (:00) has so beautifully shown for certain nematodes. 
Still more important biological problems are connected with 
the behavior of the chromosomes in parthenogenesis. In pure 
parthenogenesis only one maturation division takes place and 
the number of chromosomes remains normal; the individuality 
of the chromosomes is preserved and thus the questions of varia- 
tion and hereditary transmission are greatly simplified. Not so 
in the case of that form of parthenogenesis in which the number 
of chromosomes is reduced. I have elsewhere (Petrunkewitsch, 
:02) pointed out that the great difficulties here encountered by 
the theory of Weismann — which maintains not an essential, but a 
qualitative difference only between the chromosomes — grow out 
of the cessation of hereditary transmissible variability. There 
seem to be fewer difficulties in the way of the theory of Sutton 
‘ and Boveri,— who affirm that the chromosomes are essentially 
different, — since I have shown — and still maintain after a care- 
ful re-examination of my sections, in opposition to the objections 
of Doncaster (:04) — that the number of chromosomes in the 
first cleavage nucleus of the drone-egg again becomes normal, 
probably through longitudinal splitting without a corresponding 
division of the cytoplasm. I thought to find a support for 
