196 S. W. GEISER. 



* 



developed. The indices have rather sharply-defined limits, and 

 these are nearly mutually exclusive. Applying these facts to the 

 examination of suspected male Gambusia, we find that the 

 number of sterile males is inconsiderable and quite inadequate 

 to explain the occurrence of unusual sex-ratios. 



With young, immature Gambusia, however, the adequacy of 

 this criterion of sex is extremely doubtful. In the case of young, 

 if the gonopod has not commenced to differentiate, the only re- 

 liable method of ascertaining the sex of a given individual is by 

 cytological examination. 



After this digression on criteria of sex, we may resume dis- 

 cussion of methods. 



Methods and Results. During the spring of 1921 it was found 

 possible in large indoor aquaria to raise young Gambusia and to 

 keep them, with very little mortality until the males could be 

 distinguished from the females. Four litters were raised as 

 follows : 



On May 15, 1921 four females, each approximately 45 mm. 

 long, and well advanced in pregnancy were taken at random 

 from a lot of gravid Gambusia which had been collected at 

 Beaufort, N. C., March 22, and shipped to Baltimore. These 

 were isolated in special breeding aquaria which consisted of 

 ordinary 3. 5-liter battery jars in which were hung cages made of 

 wire netting of 3.5 mm. square mesh. The cages were coated 

 with beeswax to prevent rusting. The gravid females were put 

 into these wire baskets, one in each, and the aquaria stocked 

 with plants as usual. When the young were born they darted 

 out through the meshes of the mother's cage and were thus 

 able to escape her cannibalism. Gambusia females very fre- 

 quently eat their new-born young a habit that is shared with 

 some other viviparous cyprinodont fishes. 



The females were fed a variety of foods: boiled white of egg, 

 bread soaked in beaten egg and dried, and finely chopped snails. 

 They devoured eagerly microcrustacea, enchytraeid worms, and 

 mosquito larvea, but these food materials were not always avail- 

 able. 



The four females designated a, b, c, and d had young on the 

 following days: May 19, May 19, May 21, and May 21, re- 



