FISHES FOR CONTROL OF MOSQUITOES. 33 



ary and March a close watch was kept for Gambusia, but none was 

 seen. Early in April boj's displaced the screen which had been placed 

 across a trench connecting this pond with a larger one and admitted 

 some simfishes. On discovering this the small pond was immediately 

 seined and the intruders removed. Only a solitary female Gambusia 

 was found on April 12. There is some uncertainty whether the others 

 were eaten by the sunfishes, but no traces of any were found in the 

 digestive tracts of the latter, and in view of Hildebrand's observation 

 of the successful association of the two this seems improbable. Small 

 plantings in two ponds at Media and single fish placed in rain barrels 

 had likewise disappeared, except that a dead fish was found in one 

 of the latter. A brood stock of several hundred kept in an aquarium 

 tank in the University of Pennsylvania vivarium and fed first on 

 CuUx pipieiis larvae and later on ground boiled liver and living 

 Daphnia not only survived but increased in number. 



On Maj^ 26, 1919, after repeated thorough seining which yielded 

 no additional Gambusia or sunfishes, and the temperature of the 

 water being 64° F., 40 female and 12 male top minnows from the 

 reserve stock were introduced. At this time the water was thick 

 with entomostracans. chiefly Daphnia, upon which the Gambusia 

 immediately began feeding eagerly. While these lasted the Gam- 

 busia partially neglected the mosquitoes, and moderate breeding 

 continued at the rate of 3 to 4 per dipper in easily accessible places 

 and at nearly 10 times that rate in protected places. By June 5 

 the swarm of entomostracans had nearly disappeared and the Gam- 

 busia were searching among the plants and close to shore for food. 

 Several small groups of young were already seen. Mosquito larvae 

 were present in exposed places at the rate of 2 to every 3 samples. 

 In the fine screen check pen the numbers ran 8 to 15 Culex and 1 to 

 6 Anopheles per sample. On June 18 Gambusia were much in 

 evidence all over the pond and mosquito breeding had fallen to only 

 4 Culex larvae in 30 dippers taken all round the border of the pond, 

 including thick vegetation, while in the check pen the average was 

 4.8 per dipper, about 20 per cent being Anopheles. 



On June 30 conditions were the same, except that the rate of 

 breeding in the check pen had risen to an average of 14.3. Dur- 

 ing July the Gambusia increased in number astonishingly. On 

 July 23 they swarmed everywhere in both the small and larger 

 ponds, into which latter they had escaped through flooding of the 

 ponds. A remarkable succession of heavy rains during midsum- 

 mer caused the ponds to overflow several times. The surplus water 

 poured down a walk and over the grounds, leaving short-lived pools 

 here and there when the rain ceased. From a single one of these 

 301 Gambusia were recovered and in addition many dead were seen 

 on the ground. When it is considered that these wore only a small 

 part of those that escaped and that the fishes still swarmed in both 

 of the ponds it can readily be seen that the 52 fishes introduced 

 on May 26 must have increased to several thousand in the course 

 of only three months. This is quite consistent with what is known 

 of the number and size of the broods and early maturity of this 

 fish. With the exception of a few dragonfly and Dytiscus larvae 

 which were abundantly supplied with other food they had no enemies 

 whatever in the small pond. They continued to increase through- 



