24 U. S. BUREAU OF FISHERIES. 



in places fully accessible to the lollies, and although vast numbers 

 of the larvae are killed by the rapid drying of shallow pools the meas- 

 ure of control is chiefly determined by the success with which this 

 principle has been applied. This is shown by numerous cases where 

 long continued rains have kept in a flooded state portions of the 

 meadows usually well drained but not open to the fishes, by the effect 

 of obstructed drainage ditches and of construction work which inter- 

 fered with the movements of and renewal of the supply of killies, and 

 by the results of the repellent influence on the fishes of contamination 

 of certain waters by manufacturing wastes. In many such cases the 

 areas affected, some of them extensive, have produced great swarms 

 of mosquitoes which quickly waned whenever the barrier conditions 

 were corrected and the killies permitted to do their work. A volume 

 might be filled with the description of such instances. Headlee 

 and Carrol (1919, p. 17) write: 



Two natural agencies limit or entirely eliminate the broods. If the weather 

 is briglit, the shallow sheet water covering large areas of marsh surface and 

 filling the shallower pools is quickly evaporated and the wrigglers die. The 

 deeper pools harbor small killifish, which promptly eat all wrigglers that may 

 appear. The result is that comparatively few mosquitoes get on the wing. If, 

 however, the weather is cloudy and the atmospheric moisture high, the sheet 

 water disappears very slowly and an enormous brood of mosquitoes may 

 escape. If the covering tide is very high — so high as to bury the meadows 

 deeply — killifish penetrate everywhere, and whether the weather is bright or 

 cloudy prevent the emergence of a large brood of mosquitoes by eating up the 

 wrigglers. 



Under these conditions it seems remarkable that the similar relation 

 which these fishes bear to the fresh-water tidal swamps should be so 

 little appreciated. For example, one sanitary officer having direction 

 of an extensive antimosquito project involving large areas of such 

 swamp told the writer that the use of fishes in such an area was not 

 a practical proposition. For this reason a general study of mosquito- 

 breeding conditions in the marshes bordering both sides of the Dela- 

 ware Eiver was made, especially favorable facilities being afforded 

 through the interest of William V. Becker, the engineer in charge of 

 the operations in the Hog Island region. A full description of the 

 many interesting and sometimes exasperating conditions in this rap- 

 idly developing industrial district would require too much space, 

 but a brief summary may be given. 



Broadly considered, three sets of conditions prevail. First, there 

 are the natural tidal swamps bordering the river and the lower por- 

 tions of its tributaries. These are intersected by a network of creeks 

 and channels and for the most part covered with a heavy growth of 

 reeds. At flood tide they are covered ordinarily with shallow water 

 which again drains off with the ebb but leaves numerous small pools 

 and puddles. Second, are the reclaimed tidal swamps, rich alluvial 

 lands originally taken up for agricultural purposes but now mostly 

 either neglected meadows or passing over to industrial uses. These 

 are inclosed by dykes and drained by systems of ditches emptying 

 into the river by sluiceways guarded by automatic tide gates. Some- 

 times the drainage ditches are supplemented by wider canals origi- 

 nally designed to facilitate transportation about the farms. Many of 

 these drainage systems are very old, dating back to the original settlers. 

 With the industrial development of the Delaware River front many 

 of them have become obstructed, disarranged, or neglected. The 



