FISHES FOR CONTROL OF MOSQUITOES. 61 



fishes, and young" roach) were abundant on the clear shores and 

 among the more open vegetation. 



On the afternoon of July 11 this area was cleared of obstructions, 

 some grass was cut and removed, and some detached pools were 

 opened up. Some of the latter and several small portions of the 

 former were allowed to remain unchanged as checks. On July 12, 

 a. m,, a collection of fishes (Table T, No. 23o5, p. -l-l) was made close 

 to shore in the opened waters. Six of 10 sunfishes examined contained 

 Culex larvEe; 1, a pupa. On July 15 mud minnows and sunfishes 

 in large numbers had penetrated to all parts of the opened area and 

 the breeding rate had fallen to an average of 3.6 per sample, while 

 it aA-^eraged 25, including a few Anopheles, in unopened parts. 



On July 24: swarms of young sunfishes averaging about three- 

 fourths of an inch long were feeding here close to shore. There 

 were also a few yearling sunfishes, mud minnows, and a few blue- 

 spotted sunfishes. Collections were made at three points — on the 

 originally clean shore where mosquito breeding had not been ob- 

 served, on the shore at the head of the opened area, and within sev- 

 eral of the opened pools. Sample stomach contents of these are 

 shown in Table 7. Of the first lot, five stomachs (No. 2356b) con- 

 tained no moscjuito larvae; of the second lot (No. 2356d). 6 out of 

 10 of the stomachs showed a total of at least 13 larvse and 2 pupae 

 together with several egg boats; of the third lot, every stomach ex- 

 amined contained larvse, those (No. 235Tb) shown in the table being 

 all that were taken in one pool. The 4 contained at least 14 

 larvae and 1 pupa. The remaining stomach contents were a few 

 chironomid larvce and entomostracans. In spots where two weeks 

 before the breeding rate exceeded an average of 40 the niunber had 

 fallen to 3.6 where the fishes had entered, whereas in pools and small 

 areas from which they had remained debarred the average now ex- 

 ceeded 50 per sample. 



On July 30 the average population of larvse in opened-up pools 

 from which mud minnows and sunfishes were collected was nearly 

 4 (mostly newly hatched) per dipper, while exactly similar check 

 pools containing no fishes so far as could be ascertained ran from 

 45 to 80 or more, of which about 20 per cent were nearly full-grown 

 larvse, 3 or 4 per cent pupse, and the rest mostly young stages. 

 Both contained egg boats in about ecjual numbers. Large numbers of 

 young sunfishes (Table 7, No. 2362b) about an inch long were taken 

 alongshore close to the pools. The stomachs of 3 out of 10 of these 

 contained mosquito larvse, the bulk of the contents being chironomid 

 larvse and entomostracans. Five stomachs of fishes taken directly 

 from the opened pools all contained mosquito larvae, the remaining 

 contents being similar to the last. 



Mosquito production in the opened waters continued at a low 

 point throughout August. On August 23 breeding in the opened 

 waters was at the rate of 1| small larvae per dipper and in the 

 checks 12-| per dipper, including 8 per cent of last stage larvse 

 and 2 per cent of pupae. In the shallow water fishes were much less 

 in evidence on this date, but some small sunfishes, roach, and horned 

 pout were seined close to shore. Of 6 of the former examined 

 (Table 7, No. 2378a and 2378b) one 32 mm. long contained some 

 eggs and larval bristles of Gtdex pipiens, and one yearling 3 full- 



