26 U. S. BUEEAU OF FISHERIES. 



with killies, and near low water multitudes were seen in the pools 

 or Avriggling across the mud. Except in a narrow fringe aroimd the 

 landward side of this marsh, beyond free tidal action and the reach 

 of the killies, no mosquito breeding was detected here. 



On September 13 a pool about <0 by 40 feet, cut off by the exten- 

 sion of a new railroad embankment across a large ditch, was exam- 

 ined. In it were numerous landlocked killifishes (Table 3, No. 

 18913g, p. 27), and not a single mosquito larva could be detected any- 

 where. Across an old railroad embanlnnent bounding this pool on 

 the other side, and separated from the latter by the width of this 

 embanlonent only, was a similar but smaller and much older pool. 

 This was alive with all stages of the young of Gulex pipiens. No 

 killifishes could be seen or found with a minnow seine. One hundred 

 of the fishes were transferred from the first to the second pool and 

 immediately began to feed activel}^ upon the larvae and pupse. The 

 author was unable to visit this pool again to determine the final re- 

 sult. A number of pens and inclosures placed along another section 

 of Darby Creek to test the mosquito-eating capacities of the killies 

 were destroyed or tampered with by meddlers, but the experiments 

 of Viereck (m Smith, 1904), Chidester (1916), and others on the 

 salt marshes furnish all needed evidence on this point. 



Contamination of water may serve as a bar to the killifishes. A 

 number of examples of this were met with, and the reports of the 

 New Jersey Mosquito Extermination Association record several. 

 One such case affected an extensive area along Darby Creek. In 

 1917 a large manufacturing plant emptied great vats of strongly 

 alkaline water into the creek. Probably as a result of the conse- 

 quent change in the normal acidity of the water the killifishes aban- 

 doned a stretch of the creek exceeding a mile in length for a period of 

 several weeks. The result, which appeared to be directly connected 

 with the absence of the fishes, was that a great brood of mosquitoes 

 matured on this area where but few were produced before and after. 



Little has been published relating to the food of this species, the 

 most important being the paper by Chidester (1916), who writes 

 (p. 11) : 



Examinations of the stomachs of adult Funduli showed that they eat larvae, 

 pupse, and adults of all the salt-marsh mosquitoes. They also eat Dytiscus, 

 Notonectii, and many Daphnids. In the winter small quantities of algal mat- 

 ter and a few small shrimps constitute the most of the food of the active indi- 

 viduals. In tlie early fall the chief food besides mosquitoes seems to be insect 

 and snail eggs and occasionally a few fish eggs. 



The author has examined 42 stomachs of this killy, mostly taken 

 from the fresh waters below Philadelphia (Table 3, p. 27). In only 

 four cases were mosquito larvae found. One was in a fish one-half inch 

 long taken on the tidal flats of Darby Creek on July 26 (No. 18726k). 

 The other food was ostracods and copepods. Two others were from 

 the pumping station pond (No. 18816m). The stomachs of adults 

 contained large quantities of organic ooze with plant debris, minute 

 animal and plant life, oligochsetes, mollusks, entomostracans, and 

 insect larvae. In some cases the sole stomach content was Spirogyra 

 and other Confervas. 



May it not fairly be claimed that the use of Fundulus on the salt 

 meadows is the classical example of the employment of a natural 

 agent in mosquito control and that no other species of fish, not even 



