NATURAL ENEMIES OF MOSQUITOS 55 



monly found in water too shallow for larger fish. 

 They chiefly feed on the eggs, larvcC and pups of 

 mosquitos. They multiply very rapidly, are oviparous 

 in captivity ; in water tanks, reservoirs, fountains, 

 and lily-tubs they thrive and multiply, and through- 

 out the island they are largely used for these receptacles 

 both in town and on estates to reduce the annoyance 

 from mosquitos. I have examined some sixty re- 

 ceptacles containing " millions," and I can testify to 

 their absolute use in killing the larva? of the Stegomyia. 

 In one house I examined twenty-seven lily-tubs con- 

 taining millions, and in no instance were larva? present. 

 In another case I examined twenty-five barrels 

 containing aquatic plants, and in no case were larvae 

 present ; but they were present in a dirty stream in 

 the vicinity. JNIy experience has been similar with 

 lily ponds, where I have failed to detect larvae. These 

 observations have been fully proved, and the " millions " 

 shown to be the natural enemy of the mosquito in 

 Barbados, by a series of experiments undertaken by 

 Dr. H. A. Ballou, on the staff of the Imperial 

 Department of Agriculture in Barbados, who, by 

 keeping these minute fish under observation in the 

 laboratory, has proved that they eat the eggs, larva?, 

 and pupa? of both Culex and Stegomyia. So successful 

 have these experiments been that the Imperial Depart- 

 ment of Agriculture is introducing them into other 

 mosquito plagued islands. But JNIr. Gibbons's ex- 

 planation was criticised because, in spite of the fact 

 of no " anopheles," nevertheless Culex and Stegomyia 

 were abundant. If the fish theory were correct, then 



