S. VV. GEISER. 



of the top-minnows or Poeciliida:. The results in reference to 

 sex obtained in these collections are presented in Table I. By 

 referring to this Table it will be seen that in practically all ratios 

 reported for Gambusia, and in 80 per cent, of the records of the 

 pceciliids generally, the females were greatly in excess. 



It has been suggested that the discrepancy in the proportions 

 of the sexes is due to the fact that in this group of fishes the 

 males are much smaller and more agile than the females, and 

 hence either more re'adily avoid, or pass through the meshes of, 

 the net. However, in those cases where the net used 'had a mesh 

 so small that loss of the smaller males was thereby excluded, 

 collections still showed the same inequalities of numbers of the 

 sexes. It therefore seems well-established that in the field there 

 is usually, if not always, a marked excess of females in these 

 fishes. The writer has had populations under observation in 

 which the proportion of males present was as low as 3.8 per cent. 

 By the method employed in catching them it was not possible 

 for a single individual, male or female, to escape. 



Moreover, Hildebrand ('i/) observed the sex-ratios of speci- 

 mens of Gambusia raised in the laboratory under conditions in 

 which every individual was examined, and found them to be 

 approximately the same as those observed in collections taken in 

 the field. He concludes ('17, p. 10) that "It seems entirely 

 probable that the normal ratio of males to females is about I 

 to 8 or 9." 



Mast ('21) found on dissection that, in some populations of 

 Gambusia that he studied, a considerable proportion of the in- 

 dividuals which looked like females and had been counted as 

 such were really males in which the secondary sex-characters 

 were not developed. He designated these " sterile males " and 

 suggested that the reports of excessive numbers of females 

 might be due to the fact that many such males might be counted 

 as females. However, in the course of the writer's studies, 

 when large numbers of " females " from populations showing 

 low percentages of males were dissected to ascertain whether 

 a considerable number of these " females " might not really be 

 sterile males it was found that the percentages of sterile males 

 in such populations was too small to account, even in a small 



