836 BULLETIN OF THE BUREAU OF FISHERIES. 



afford them shelter from the predaceous fishes. They are also to be found in 

 shallow ditches and surface drains where the water is not foul, even where it is 

 but the fraction of an inch deep. In fact, if any fishes will find their way to the 

 remotest possible breeding places of the mosquito it will be Gambusia and 

 Heterandria. And they are the only ones, so far as the writer's observation 

 goes, that can be considered at all useful as destroyers of Anopheles larvse. 



To what extent they could be acclimated in northern waters has yet to 

 be determined. They are to be found in the Ohio Valley as far north as south- 

 ern Illinois, hundreds of miles above tide water, where the climate must be 

 quite severe. In 1905, at the earnest request of Prof. John B. Smith, state 

 entomologist of New Jersey, the writer planted about 10,000 Gambusia and 

 Heterandria in New Jersey waters. Some 8,000 were planted in one locality 

 which was thought to afford very favorable conditions. In 1907 Mr. Henry 

 W. Fowler, ichthyologist of the Academy of Sciences of Philadelphia, and 

 author of " Fishes of New Jersey," found considerable numbers of Gambusia in 

 the vicinity of Cape May, some 90 miles from where the plant was made. This 

 opens up a very interesting question. Mr. Fowler contends that Gambusia 

 should be considered as indigenous to New Jersey. Very strong arguments 

 to the contrary can be advanced, but the question is not of importance 

 in connection with this paper, except that it either gives a farther northern 

 range to the species or that, on the other hand, it shows the possibility of 

 introducing them. 



The writer has come to the conclusion, after many experiments in small 

 ponds, that a combination of the goldfish, which is ornamental and useful in 

 the open water, the roach or shiner, which is a very active species, two small 

 species of sunfish, which live among plants, and the top minnow would probably 

 prove to be more effective in preventing mosquito breeding than any other 

 fishes. The goldfish is somewhat lethargic in habit, and is also omnivorous, 

 but there is no doubt that it will devour any mosquito larvae that may come 

 in its way or that may attract its attention. The one great objection is that 

 it grows too large and that it is cannibalistic, so that when a pond is once stocked 

 with large goldfish the number of young to survive will be small. 



The roach is probably the most widely distributed and abundant of all 

 the small fishes except the cyprinodonts. It is a very active fish, always ranging 

 about in search of food. 



The two small species of sunfish, of the genus Enneacanthus , are very widely 

 distributed. They live wholly among plants and feed upon larvae of all kinds. 



The top minnows are foragers always on the move in the search for food, 

 skimming over the tops of plants with restless energy. 



All of the above-mentioned species are among the most abundant wherever 

 found. If the range of the top minnows can be extended north it will prove to 



