■184 



SCIENCE 



[X. S. Vol. XXXII. No. 814 



for the first set of broods, from parents one or 

 both of which were submitted to very high 

 ■temperatures at some period during the larval 

 stage, or during the early adult stage before 

 mating. A similar result was, however, ob- 

 tained from a control series, as shown in 

 Table II., indicating that the high tempera- 

 ture used was not the cause of the unisexual 

 ibroods. 



TABLE 1. (51 PAIBS) 



A. Unisexual Broods (frorn 9 Pairs) 



Male Female 



1 135 



2 108 



3 104 



4 73 



5 63 



6 45 



7 43 



8 33 



9 31 

 Total T35 5W 



B. Bisexual Broods {from ^2 Pairs) 



Botli 

 Male Female Sexts 



Xargest brood 99 + 95 = 194 



Smallest brood 12 -)- 11 z= 23 



Total number of flies 1994 -|- 1992 =3986 



Average brood 47.47 -|- 47.42=: 94.9 



Deviations from normal sex-ratio of 50 per cent.: 



Maximum 15.00^ (d' 35.00^ + 5 65.00 ^) 



Minimum 0.00;^ (equality) 



Average 4.53 ^ 



TABLE II. (21 PAIKS) 



A. Unisexual Broods [from S Pairs) 



Male Female 



1 68 



2 1 52 



3 30 

 Total T 150 



B. Bisexual Broods {from IS Pairs) 



Both 

 Male Female Sexes 



Largest brood ISl -|- 177 = 358 



Smallest brood 36 -|- 23 = 59 



Total number of flies 1428 -|- 1383 =2811 



Average brood 79.33 -|- 76.83= 156.1 



Deviations from normal sex-ratio of 50 per cent.: 



Maximum 11.01^ (c? 61.01 ^ -f- ? 38.99 ^) 



Minimum 0.35^ (c? 50.35 ;*-+-$ 49.65 ;f) 



Average 3.11 ^ 



The total number of flies making up the 

 bisexual broods consists of males and females 

 in almost equal proportions. The sexes of 

 these individual broods also ran fairly even ex- 

 cept in a single ease (Table II., A, ISTo. 2) 

 where the sex ratio is 52 :1, and which is 

 placed with the unisexual broods on this ac- 

 count, as well as for another reason which 

 will appear below. 



The flies were taken from a stock originally 

 collected at Woods Hole by Professor T. H. 

 Morgan, and bred in large numbers in several 

 vessels. The offspring of these secondary 

 stocks were isolated at various times in the 

 pupal stage, and the virgin flies thus secured 

 were paired in separate vials; the families of 

 these pairs constitute the broods referred to 

 in the present note. It is hardly possible that 

 the parents of all the unisexual broods of 

 Tables I. and II. can have sprung from a 

 single pair of flies, and it therefore seems 

 probable that the twelve pairs were separately 

 acted upon by some unknown external factor 

 which so strongly influenced the process of 

 sex-determination that only one or the other 

 sex was produced. This is, of course, not 

 definitely proved, since no record was kept of 

 sex-mortality, but it will be observed that the 

 number of individuals in some of the uni- 

 sexual broods is high enough, as compared 

 with that of the bisexual broods, to suggest 

 that the effect was not due to elimination of 

 one sex but to substitution by the other. 



In a second set of controls the 27 pairs all 

 produced normal broods. The attempt to se- 

 cure further results was then continued with 

 the same stock at the Marine Biological Lab- 

 oratory at Woods Hole through the summer 

 of 1909, and again for more than half the 

 season of 1909-10 at Columbia University, 

 but without success, though over TOO pairs 

 were bred and experimented with during that 

 period. 



The second fact connected with these uni- 

 sexual broods became apparent after making 

 many unsuccessful efforts to breed from them; 

 the flies were all sterile, including the single 

 male and all the females in the second " uni- 



