FISHES rOR CONTROL OF MOSQUITOES. 49 



erages were, respectively, 0.25, 3.6, 14, and 5 and on June 30, 2.5, 3, 

 about 20, and 13, respectively. On July 21 they were nearly 3, 4, 

 and 16, there being no record for the gutter, as the water had flowed 

 around the screens. 



The next thorough examination was made on August 15. During 

 the interval heavy rains had flooded the ponds and escaped Gambusia 

 as well as young sunfish had penetrated everywhere except the line 

 screen check pen. In correspondence with this there was practically 

 no breeding anywhere except in this pen. In 60 samples taken in all 

 parts of the two ponds no larvae were found in the small pond, 2 

 small Culex in the gutter, and 3 in the large pond. How much of 

 this reduction was due to the sunfish and how much to the top 

 minnow could not be determined. Stomach contents of sunfishes 

 (Table 7, p. 43, Nos. 19T21a, d; 19806b; and 19815a, e) taken 

 during this period gave results similar to those obtained in 1918, 

 namely, during the period of great abundance of chironomid larvae 

 and other food in July and early August mosquito larvae were neg- 

 lected but reappeared in the food on August 15 as part of a much 

 more varied diet. On later dates, owing probably to the presence 

 of Gambusia and greatly diminished breeding, none were found in 

 the stomachs. 



These and similar experiments made on ponds in Delaware County 

 and nmnerous observations made at other places make it clear that 

 the common sunfish is very effective in keeping culicine breeding in 

 check, except for a period from about the middle of June to early 

 August, during which the great abundance of attractive and easily 

 secured food leads to the neglect of the mosquito larvae, at that time 

 well protected by the increased vegetation. The fact that this is 

 their breeding season probably also affects this result. Later in the 

 summer they again turn to mosquitoes, as is evidenced both by the 

 diminution in the numbers of the latter and by their presence in the 

 stomachs of the fishes. There is also a parallel effect upon anopheline 

 breeding, but this is far less evident and certain and is not supported 

 by the results of stomach examinations. The very efficient control 

 that was exercised during August and September, in spite of the 

 greater density of the plant growths, is due mainly to the large num- 

 bers of young sunfislies which penetrated the vegetation and entered 

 the very shallow waters. 



Some of the experiments in Palisades Interstate Park during July 

 and August, 1919, were made on a larger scale and show the positive 

 role of this species in mosquito control even more clearly. Of these, 

 two typical experiments in which the common sunfish was the chief 

 actor will be described. The first (D 1) relates to a little sheltered 

 cove (figs. 6. 7. 8) described on page 21. On the open shore of the 

 lake outside of this area there was no breeding except in detached 

 pools, etc. No fishes were found in the barred area, but outside 

 of it they vv^ere numerous. Mud minnovvs, young roach, long-eared 

 sunfishes, blue-spotted sunfishes, stone catfislies, and especially com- 

 mon sunfishes were seined. Adults of the latter and of the long- 

 eared sunfish were breeding in large numbers along a sandy and 

 gravelly shore here, and 100 nests were counted without exhausting 

 them. No mosquitoes were detected in the stomachs of sunfishes 

 taken on this date (July 5), 



