FISHES FOR (!ONTR0L OF MOSQUITOES. 47 



with the range of the others. Moreover they were the only species 

 that was actually seen to eat the larvaa and the only one in whose 

 stomachs larv£e were actually found. 



On the plant of the Hercules Powder Co., at Kenvil, N. J., is a 

 natural pond of several acres area. This was visited in September, 

 1918, and thoroughly examined on August 13 and 14, 1920, a time 

 when mosquito breeding was generally at a rather low ebb. Most 

 of the shore of this pond was clean, and while vegetation was often 

 plentiful it was seldom sufficiently dense to form an effective barrier 

 to prying fishes. There were many excellent places for the nesting 

 of sunfishes, and that they had been utilized Avas shown by the large 

 number of young seen everjMvhere patrolling the shores to the water's 

 edge. Young roach also were plentiful but showed less of the ex- 

 ploring instinct. Nowhere round these shores, even among the 

 vegetation, were any mosquito larvae found except at one point, where 

 a heavy log had lod^^ed across the mouth of a small bight and being 

 both deep sunken in the bottom and reaching above the surface 

 formed a complete dam. Behind this in the cut-off pool larvae of 

 both Culex pipiens and Anopheles were found, and this was the 

 only place in the pond that Avas not open to the fishes and the only 

 one where breeding was detected. 



On the other hand, in a near-by swamp where there were no fishes 

 there was abundant breeding of Culex pipmns and Aedes sylvestris, 

 and immediately across a low narrow railroad embankment was a 

 large poo], except for size, exactly similar to Duck Pond and probably 

 originally part of it and now fed by seepage through the embank- 

 ment. In this pool there were no fishes of any kind and Culex pipiens 

 was breeding everywhere in moderate numbers (3.4 per dipper). 

 The stomach contents of the fishes taken in Duck Pond are shown in 

 Table 7 (No. 20813g, p. 44). The sunfishes contained 40 per cent of 

 chironomid and a few other larvae, the remainder being entomos- 

 tracans with a trace of alga?. In the 13 stomachs examined no 

 mosquito remains were found. More than a score of similar instances 

 might be given showing a correlation between the distribution of the 

 common sunfish in particular bodies of water and the absence or 

 paucity of mosquito larvse, together with their presence in parts of 

 the same waters to which fishes do not penetrate. 



The history of mosquito breeding (chiefly Culex pipiens) in the 

 larger pond in the University of Pennsylvania botanical gardens il- 

 lustrates some of the conditions limiting the effectiveness of the com- 

 mon sunfish in relation to mosquito control. This pond (figs. 12,^13, 

 14) is filled with water lilies and other ornamental plants, but in few 

 places is the growth dense enough to prevent the small sunfishes 

 which occur in great numbers from reaching all parts of the shore 

 line. Besides the common sunfish were some long-eared sunfish in 

 the ratio of about 3 to 1, many large goldfish, and a very few roach. 

 Bullfrogs and tadpoles were very abundant, and there were some 

 predacious beetles and bugs. There were no young goldfishes, as the 

 sunfishes ate their eggs as fast as laid. The sunfishes, however, bred 

 successfully. Many nests were seen, and the young abounded dur- 

 ing the summer. 



During May and the first half of June, 1918, mosquitoes were 

 flying in considerable numbers (produced by some tubs of water in a 



87028°— 22 4 



