SEX-RATIOS IN GAMBUSIA HOLBROOKI. 199 



individuals were taken with a dipnet at random from the pond. 

 Of these 284 Gambiisia, 94 were males, 60 were females, and 

 130 were so small that the sex could not be ascertained by ex- 

 ternal observation. Both of these collections were from the 

 same parental stock as the females whose litters were studied. 

 One is impelled to conclude as a result of these observations 

 that the great excess of females in the parental stock (nearly 8 

 to I ) does not represent the proportions of the sexes at birth, but . 

 must have been due to a greater mortality of the males during 

 either the juvenile period or later. 



The only careful piece of work that has hitherto been done in 

 the attempt to ascertain the sex-ratio at birth in Gambusia is 

 that by Hildebrand, and his results are vitiated by a high death- 

 rate. He found that five months after birth, of an original 

 litter of 46 fish, none had developed a gonopod ;, and at the age 

 of approximately 13 months, only six of the surviving fish pos- 

 sessed a gonopod, thus giving a very low ratio, even if a possible 

 very high death-rate, which he mentions, is taken into considera- 

 tion. But, as has already been seen, the factors of food and 

 temperature are particularly potent in determining the rapidity 

 of sexual development and the production of young in these 

 viviparous fish. This was shown clearly in the work of Schmidt 

 ('19, J i9a), 8 as well as by the writer's studies on duration of 

 pregnancy and gonopod-development. 



In confirmation of "the correctness of the conclusion that the 

 normal secondary sex-ratio in at least some Poeciliids is approxi- 

 mately i : i, Henn's work ('16) may be brought forward. A 

 total of 2,070 individuals of Lebistes reticulatus, the millions fish, 

 was obtained in a single collection with a very fine-meshed net 

 in the Barbadoes under the direction of Professor C. H. Eigen- 

 mann. The collection gave an approximately i : i ratio. In the 

 lot there were 520 males and 630 females, besides 920 fish less 

 than 10 mm. long, and too small to permit ascertainment of 

 the sex by external examination. Henn says on the point of the 

 sex-ratio that " it is quite certain that this count of males in- 

 cludes only members of that sex while a few of the smaller speci- 



8 Duration of gravidity in Lebistes at 25 is about one month; at 18, more 

 than three months. (Schmidt, 'igb, p. 4.) 



