478 ANNUAL KEPOET SMITHSONIAN INSTITUTION, 1910. 



sibly be introduced by accident, since the species is so abundant. The 

 writer at one time supposed that there were two kinds of eggs, one 

 requiring fertilization and yielding females, the other developing 

 into males without being fertilized. But more recent experiment'^ 

 (made with the help of Mr. A. C. Tunstall, and not yet sufficiently 

 extensive) do not support this idea. Out of two broods — one of 58 

 eggs, the other of 102 — 53 and 67 male pupse or adult larvae were 

 reared, which clearly indicates that the absence of females was not 

 caused by their dying off; for among the eggs of fertilized females at 

 least 50 per cent commonly yield females. It was not possible in these 

 experiments to hatch out the flies, but the size of the adult larvae or 

 pupaj is an almost unfailing criterion of the sex. It seems probable 

 therefore that Nemotus ribesn must be placed in the same category 

 with the bee. The gallflies (Cynipidffi) offer another anomalous in- 

 stance, for in them there are two generations in the year, one of 

 which consists of both sexes, of which all the eggs are fertilized and 

 jdeld a parthenogenetic generation consisting wholly of females. The 

 eggs of the latter yield both males and females, and the writer has 

 shown that the male-producing eggs undergo maturation; the female- 

 producing do not.^ 



One more point must be mentioned here. In bees some hives pro- 

 duce a large proportion of gynandromorphic individuals, whicli are 

 irregular mixtures or mosaics of male and female characters. Von 

 Siebold- described such a case, and found that all the " zwitter- 

 bienen " were in worker cells, the drones being all pure. The hive was 

 hybrid from Italian stock crossed with black ; the drones were of the 

 Italian type, the workers mixed. Two possible explanations may be 

 hazarded: First, that the egg which develops into a g}mandromorph 

 has begTin to segment, and that the male pronucleus conjugates Avith 

 one of the segmenting nuclei ; or, secondly, that the male pronucleus 

 conjugates with one (probably the second) of the polar nuclei, and 

 that both the zygote nucleus so produced and the female pronucleus 

 take part in the development. Of these, the second is perhaps the 

 more likely hypothesis. 



It has been mentioned above that Beard was one of the first to 

 suggest that the germ cells bear the determinant for one or the other 

 sex ; it now remains to discuss the evidence which has since accumu- 

 lated in favor of that hypothesis. It has received support on several 

 very distinct grounds. We may take, first, the cytological results 

 with which the names of several American investigators are chiefly 

 associated, although much similar work has been done in Germany, 

 France, and elsewhere. To give an at all adequate account of the 



1 [Proc. Roy. Soc. B. 82, 1910. p. 88, and second part of the same paper to be published 

 in Proc. Roy. Soc. shortly.] 

 - Zeit. Wiss. Zoo., vol. 14, p. 73. 



