194 S. W. GEISER. 



fish reaches a length of 13 mm., or at any later stage until it 

 attains its maximum normal growth of about 25 mm." 



Mast ('21) as noted above, has given further evidence of the 

 tardy development of the gonopod in the " sterile males " of 

 Gambusia Jwlbrooki. He found in these sterile males that the 

 testes had already differentiated and were in a condition of ad- 

 vanced spermatogenesis. Even ripe spermatozoa were present 

 in the form of sperm-balls. Yet with all this evidence of func- 

 tional maturity, the anal fin gave no evidence of the sex of the 

 individual. 



The writer's studies show that the gonopod develops rapidly 

 during the summer, so that by the time the season is over prac-" 

 tically all the gonopods are developed and recognizable as such. 

 The differentiation of the terminal portion of the gonopod into 

 the characteristic hooks of, especially, the third, and the posterior 

 branch of the fourth fin-rays does not, however, take place often 

 until very late in the Fall or during the following Spring. 



It is clear that this is hardly a criterion of sex that can be 

 serviceable with younger fish of this species. 



The Position of the Anal Fin as a Criterion of Sc.v. In 

 Gambusia the anal fin migrates forward in the male during the 

 period of sex-differentiation with the result that when adulthood 

 is reached the relative positions of the fin in the two sexes is 

 quite dissimilar. The exact value of this " index," i.e., the re- 

 lation between the total length of the fish and the distance from 

 the tip of the nose to the anterior border of the anal fin, is easily 

 calculated. Measurements made of a large number of Poeciliid 

 fish, embracing several genera, show that the adult index is 

 always higher in the male, i.e., the anal fin is placed relatively 

 farther forward in the male. The mean index-value varies with 

 the different species. 



Graph I. shows the index values of two lots of Gambusia ob- 

 tained from Beaufort, N. C., in 1920/21. While there is con- 

 siderable variation, the indices for both sexes have very definite 

 norms. As will be seen from the graph, in the males the class 

 2.7-2.799 is the mode; in the females, the class 2.1-2.199 is the 

 mode. A comparison of the data shows that there is very little 

 overlapping. Thus, in this population, only 3.6 per cent, of the 



