6 MIT^NOWS GAMBUSIA AFFIXIS AXD CYPRINODON VARIEGATUS. 



While this fish in captivity will readily eat dead food, such as 

 minced fish, oysters, clams, corn bread, the yolk of hard-boiled egg, 

 etc., it shows a preference for living food. Mosquito larvae were 

 killed and introduced along with live ones, and in each instance no 

 attention was given to the dead larvae until the live ones had been 

 consumed. It is probable that this fish has a preference for insects 

 as food, but it is evident that it is by no means dependent upon these 

 for subsistence. Apparently it devours nearly anything of suitable 

 size, whether animal or plant. It is well known that in the aquarium 

 Gamhusia will eat its own young, but this cannibaHstic habit is cer- 

 tainly not restricted to aquarium life, since the writer has captured 

 specimens in nature which contained in the stomachs fish of their 

 own kind. 



OBSERVATIONS ON BREEDING. 



In the Beaufort region this fish delivers its first young of the 

 season some time during May, or in some years possibly as early as 

 the latter part of April, depending largely upon the temperatures 

 which prevail durmg the early spring. The spring of 1915 was some- 

 what cooler than the spring of 1914, and the breeding season, there- 

 fore, began at least two weeks later. It continues to breed through- 

 out the summer and as late as October, 



Copulation, although carefully looked for, was not satisfactorily 

 observed. Apparently it is a very quick process °' and is accom- 

 pUshed during what appear to be frequent fights in which the oppo- 

 site sexes engage. That some of these fights are quite real was evi- 

 denced by the fact that a female which was confined in a smaU rec- 

 tangular jar kiUed and partly devoured three males that were from 

 time to time introduced for breeding pui'poses. In order to protect 

 the male from this ferocious female it became necessary to place in 

 the jar a partition of wire netting, with mesh large enough to permit 

 the male to pass through, yet small enough to keep the female back. 

 The male continued to venture out from his compartment quite 

 frequently, and notwithstanding that he was obliged to make many 

 hasty retreats he survived and successfully fertilized the eggs for the 

 future broods. 



That a single female may produce as many as six broods of young 

 during a single season was demonstrated through aquarium experi- 

 ments. In one instance a medium-sized female, about 40 mm. in 

 length, was placed in a small rectangular jar early in the spring of 

 1914. She gave birth to young as follows: First brood, May 20; 



a The act of copulation in Gamhusia holbrookii and Heterandria formosa was observed and described by 

 Seal (1911). Gamhusia Tiolbrookii is now considered a synon3Tn of G. affinis. This process was also ob- 

 served and described by Philippi (1908) in Glaridichthys januariusand G.decemmaculatus. (The first of 

 these fishes according to Henn (1916) was Phalloceros caudomaculatus (Hensel) and the other is placed in 

 the genus Cnesterodon Garmaa by the same author following Eigenmaon.) 



