BIOLOGY. 379 



The possibility that some g^^nandromorphs arise from an egg with 

 two nuclei has received confirmation from one conclusive case in D. 

 simulans. There are some further cases that make it even more 

 probable than formerly that the same explanation applies to some of 

 the mosaics of D. melonogaster. 



There is a peculiar phenomenon that has been observed from time 

 to time in our stocks, that has seemed of sufficient interest to study in 

 more detail. A few pure stocks that show wide variability sometimes 

 appear to keep near that end of their variation curve that is nearest 

 the normal firy. If this were the only observed relation it could be 

 explained on the ground that the flies that are farthest from the nor- 

 mal type are more likely to die off; so that they do not appear in the 

 mass culture. But that this is not the full explanation is shown by 

 the fact that the absence of the more aberrant flies is not due to their 

 elimination after hatching, as can be found by recording all the flies 

 that hatch. Either the aberrant tj^Des die as larvae or pupse, or else 

 something else has happened to such stocks. That the latter is the 

 case is shown by breeding out the flies to wild-type stock and extract- 

 ing the mutant type in question. It is then found that the original 

 type reappears, and if such flies are inbred the stock returns to its 

 original curve of variation. Three examples may be given. There 

 is a tumor stock called I7, in which all the male larvse that carry the 

 tumor die. The tumors can be seen as black spots in these larvse. 

 This stock has changed in the sense that these tumor larvse no longer 

 appear. But that the lethal is still present is shown by the output of 

 a lethal-bearing female. She gives two female to one male offspring, 

 showing that half her sons still die. The linkage relations between the 

 lethal and other sex-linked genes present in this stock show that the 

 missing sons are still those that would show the tumor if they developed. 

 Evidently they now die before the tumor appears. 



The stock called "bent" — a recessive IV-chromosome mutant — has 

 always contained some flies with straight wings that are almost or en- 

 tirely indistinguishable from normal flies. But the stock now consists 

 almost entirely of normal-appearing flies. If, however, any of these 

 are crossed to wild-type (or to any other stock, in fact), and the F2's 

 are inbred, extreme individuals with markedly bent wings reappear — 

 individuals that are not found in the parent stock to-day. If these 

 "extreme bents" are inbred, one can quickly obtain a stock, all the 

 individuals of which are extremely bent. The result shows that the 

 original factor for bent is still present. It has not changed, but some- 

 thing else has come in that has modified the bent, so that it appears 

 under the normal form. WTiat has happened can be studied through 

 linkage relations, for which the Drosophila stock is peculiarly 

 adapted. This work is being carried out. It appears that modifjdng 

 factors have appeared in the stock of such a kind that they change the 

 bent character back towards normal type without, of course, affecting 

 the gene for bent itself. 



