34 U. S. BUREAU OF FISHERIES. 



out Aiimist and September and nearly eliminated mosquito breed- 

 ing in both ponds except in the fine wire check pen. In October 

 a reserve stock for the winter was secured and kept as before. The 

 Avinter of 1919-20 was long-continued and severe. These ponds 

 were frozen over about Thanksgiving Day and remained locked in 

 ice, which reached a thickness of 18 inches, until March 23, 1920. 



During the last week of March and during April repeated searches 

 failed to disclose any Gambusia, and as only five sunfishes were seined 

 in the small pond their complete disappearance can be attributed 

 only to their inability to survive the winter. On April 22, 1920, the 

 water temperature being 60° F., 23 females and 13 males were intro- 

 duced into the small pond. By the middle of May young began to 

 appear and rapidly increased in number. On August 2 they had 

 become very numerous and young, one-half to three- fourths inch 

 long, were everywhere and very active. Mosquito larvce which were 

 widely distributed in this pond in moderate numbers early in the 

 season had now totally disappeared. As no other fishes were present 

 in the pond their disappearance may be safely attributed to the 

 Gambusia, though stomach examinations always gave negative evi- 

 dence. On October 29 the winter's brood stock of 300 was removed 

 to the vivarium. The winter of 1920-21 was one of the mildest ever 

 recorded at Philadelphia. Only three times did ice form on the 

 pond, and at no time did it remain as long as eight days and only 

 then exceeded three-fourths inch in thickness. During most of the 

 winter the w^ater was open. Nevertheless not a single Gambusia 

 could be found in the spring of 1921. 



Similar experiments were tried at Broomall's dam near Media. 

 Here several small ponds about 12 to 15 feet square, separated by 

 narrow earth embankments and cut off from the main pond by 

 similar embankments, were arranged. These were filled partly 

 by seepage from the pond and partly by a small spring and were 

 planted with aquatics and densely grown with vegetation around the 

 margins, especially saw grass, spearmint, and tear thumb, and at 

 times are the most prolific source of Anopheles of which the writer 

 knows in this neighborhood. Both the shallows of the main pond 

 and certain of these small ones were planted with Gambusia with 

 the same result, so far as rapid increase and winter killing is con- 

 cerned, as described above. In one experiment eight females and 

 three males were placed in a pond al^out 12 feet square and 2 feet 

 deep on June 9. At that time there was light breeding of Culex and 

 no Anopheles. On August 27 Gambusia was plentiful, and 10 

 samples of water yielded 1 Culex pipiens larva and 1 pupa and 2 

 Anopheles larvfB, while an exactly similar contiguous pool without 

 Gambusia yielded 22 Culex and 12 Anopheles. On October 8 the first 

 pool yielded in 20 samples 1 Anopheles and no Culex, and the second, 

 20 larvae and 1 pupa of Culex ancl 10 larvse and 2 pupse of Anopheles. 

 Gambusia was then present in abundance in tlie first. 



A single Gambusia placed in a rain-water barrel was found to 

 keep it clear of (hilex pipiens and Aedps trlseHatus^ which also lireeds 

 in rain barrels at the author's home, but they did no better than small 

 goldfish or sunfish and had the disadvantage that owing to their 

 amall size and to])-swimming habits they were washed out in heavy 

 rains or if the barrels were covered Avere injured on the top. This 



