54 U. S. BUEEAU OF FISHERIES. 



waters result cliiefly from this influence. Indeed, it mav safely be 

 said that were these fishes suddenly to be wiped out mosquitoes would 

 immediately and aenerally become intolerable nuisances. That the 

 latter statement is not fanciful is sufficiently shown by the very 

 numerous cases where precisely this result has followed such local 

 disturbances of the balance. Many instances could be cited (for sev- 

 eral see p. 2.5-2G) where through human interference waters have 

 become inaccessible or unsuited to mosquito-eating fishes chiefly 

 through industrial or domestic pollution and thereupon have produced 

 vast numbers of mosquitoes where previously there were few or none. 

 Not a few of the worst mosquito pests from which the more populous 

 regions of the northeastern States suffer are in whole or in part man 

 made. 



Nevertheless, vast numbers of mosquitoes are produced under un- 

 disturbed natural conditions. Witness the enormous swarms that 

 occur at certain seasons even in many regions nearly or quite un- 

 inhabited by man. For the most part these arise from temporary or 

 other fishless waters. But even in waters in which fishes abound 

 mosquito control is seldom or never perfect. Almost always some 

 breeding may be detected, and it is evident that great differences in 

 completeness of control exist in different ponds, in different parts 

 of the same pond, or even in the same parts at different seasons. Fre- 

 quently from ponds abundantly stocked with mosquito-eating fishes 

 great numbers of mosquitoes will emerge at a certain part or season 

 while otherwise they are effectually checked. To what may this va- 

 riation be attributed? Evidently the initial rate of mosquito breed- 

 ing is not the primary factor, for there is abundant evidence to show 

 that with increase in the number of eggs laid and hatched within a 

 given area there is a corresponding increase in the number eaten, so 

 that this relation is largely self-regulating. There are other con- 

 ditions limiting the degree to which aquatic stages of mosquitoes 

 enter into the diet of fishes. 



If fishes are to be employed to the best advantage in mosquito- 

 control work, these limiting conditions must be clearly defined. While 

 in their entirety they are complex and variable with species and con- 

 ditions, some of them are brought out by the studies described in the 

 section on " Methods and Eesults of Investigations on Fishes" (p. 7). 

 In the first place, the farther typical pond conditions are departed 

 from and the more closely typical still-water swamp conditions are 

 approached the less effective fish control becomes until a point is 

 reached at which it appears to cease altogether. Both direct obser- 

 vation and experiment show that shallow water in itself does not 

 deter the fishes, for young sunfishes and others will enter and feed 

 in the very shallowest water clear to the actual shore line. If the 

 water be physically fit, nothing but actual mechanical barriers will 

 prevent the entrance of the more aggressive species, as was shown ex- 

 perimentally. It is the broken and obstructive character of the waters 

 and perhaps in addition the absence of deep-water retreats that de- 

 bars them. 



By far the most generally prevalent and effective of such barriers 

 is the vegetation growing in the shallows and at the shore. If these 

 growths be dense, they will almost always afford shelters in which 



