18 U. S. BUREAU OF FISHERIES. 



(13.6 per cent). Forbes and Richardson (1908) found that WolfBa and unicel- 

 lular algoe formed three-fourths of the food of this species, insects and crusta- 

 ceans making up the rest. Hankinson (1908) reported entomostracans, algse, 

 niites, midge larvae, snails, and insects. 



The detailed records show a much greater variety than this list 

 would indicate, including many protophytes and protozoans, rotifers, 

 oligochaates, leeches and nematodes, amphipods and isopods, spiders, 

 and considerable quantities of silt and mud. It must be noted that 

 of all the considerable number of individuals covered by these authors 

 not a single one is reported to have eaten mosquitoes, though it is 

 quite certain that some of them must have been taken from waters in 

 which immature stages of mosquitoes were present. This may be 

 taken as indicating the ease with which they may be overlooked 

 rather than their total absence. 



Most of the stomachs reported upon by the author (Table 2, p. 

 16-17) were collected in the course of experiments in Palisades Park 

 from specimens 17 to 73 mm. long. The food is quite similar in 

 general character to that of Vinbra, Imii^ but the percentages differ. 

 Taking all of the stomach contents of the 50 specimens together, the 

 approximate percentages of the principal items are : Insects of all 

 kinds, 70 per cent or over; crustaceans (chiefly entomostracans), 12 

 per cent; mollusks, 6 per cent; plants (exclusive of protophytes), 

 4 per cent; organic mud and silt, containing large quantities of 

 diatoms, desmids, protozoans, and other minute organic remains and 

 debris, 5 or 6 per cent. Stomachs of individuals taken in the after- 

 noon were generally empty or nearly so, while the intestines were 

 full. Those taken in the early morning usually had the stomachs 

 distended with food, much of which was little digested. The infer- 

 ence from this condition is that the mud minnow feeds habitually 

 by night and fasts during the day while lying concealed in the mud 

 and weeds. Many specimens taken in the ditches along the Dela- 

 ware and Schuylldll Rivers in Pennsylvania and New Jersey had 

 eaten much larger proportions of ooze and plants, but the diet was of 

 the same varied character. 



A comparison of stomach contents with conditions of collection 

 and the progress of experiments clearly brings out the relation of 

 the mud minnow to mosquito breeding. Of the 50 stomachs reported 

 22, or 44 per cent, were found to contain mosquito larvae and 

 pupse, which formed about 4 per cent of the total contents, or about 

 6^ per cent of the insect contents. These remains represent 40 larvae 

 and 12 pupse of Culex pipiens and Aedes sylvestris actually separated 

 and counted. Probably others were present but not isolated. 



The results of three experiments at Car Pond involving this spe- 

 cies are now described. Station J 6 (figs. 2, 3) is a nearly circular 

 hole about 7 feet in diameter and 20 inches deep situated in a little 

 delta in the midst of a dense growth of tussock ferns at the mouth of 

 Stahahe Brook which opens into the head of a shallow bay. Evi- 

 dently the earth dug from the hole had been built up all around to 

 form a solid dam, shutting off completely the deeper water of the 

 brook on one side and the shallow water of the bay and marsh on 

 the others. The pool was therefore completely isolated. The bottom 

 was of soft ooze supporting a strong growth of Potamogeton and 

 Utricularia, and a rank growth of grasses on the banks formed a 

 pendant fringe. The water was foul with decaying vegetation and 



