564 Memoir Sears Foundation for Marine Research 



may be plentiful close in to the tideline, few of them run up into salt marsh creeks 

 unless these creeks receive some inflow of fresh water; we have never heard of them 

 being trapped on the flats by the outgoing tide, as happens now and then to herring. 

 Although they are not Arctic fish, we have never heard of their being killed by winter 

 chilling, as in the case of some other fishes from time to time. 



Their actions, if these are watched from some raft or float, suggest that they travel 

 but little during their inshore sojourn, except as they may be carried to and fro by 

 the tide, as just remarked. Anglers have long known that American Smelt do not dart 

 at bait, but approach it slowly and appear to suck it into the mouth rather than seize 

 it, as most of the predaceous fishes do. 



Food. This species is carnivorous and predaceous. In salt and brackish water, shrimps 

 (decapod and mysid) probably are their chief support on the Massachusetts coast; sim- 

 ilarly, the stomach contents of those in the Gulf of St. Lawrence have consisted chiefly 

 of copepods, amphipods, and mysids, with algal debris probably taken incidentally. 

 In some localities small fish rank next. We have found them packed full of young 

 Atlantic herring, on the coast of Maine, and a wide variety of fishes has been recorded 

 as occurring in their stomachs, at Woods Hole. They also take small shellfish, small 

 squid, annelid worms, and small crabs as occasion offers. But they cease to feed during 

 the spawning season, as many other fishes do." 



Enemies^ Disease^ and Mass Fatalities. Doubtless the marine population is preyed 

 upon by a wide variety of the larger predaceous fishes, but we have no definite informa- 

 tion in this regard for the American coast (see 6j: 236 for the European smelt). In 

 American lakes they are perhaps the favorite prey of landlocked salmon and lake trout 

 and are commonly used as bait for these. 



In European harbors, mass destruction of them, from some unknown cause, takes 

 place from time to time (6j: 39-42). 



A mortality of smelt, affecting all sizes and both sexes, occurred in Lake Huron 

 and in Lake Michigan during the fall and winter of 1942— 1943 (72: 310—337), with 

 the first indication of a recovery being apparent in 1945. It was concluded that this 

 mortality, which all but wiped out the populations in these two lakes, could be ex- 

 plained only as resulting from a communicable disease (bacteria or virus). We have 

 found no record of any happening of this sort in brackish or salt water along the 

 American coast. 



Variations. A separate specific name, sergeanti, was given years ago to American 

 Smelt of the streams tributary to the head of Delaware Bay because of their smaller size 

 and a slight difference in color (Norris, 6g: 59; yo: 93). Hubbs recognized sergeanti 

 as at least a "race" because of the fewer gill rakers on the first gill arch and because of 

 a slightly higher anal fin than on the more northerly fish {43: 52). Similarly, Rice 

 long ago reported small differences in color between those of the Raritan River, New 

 Jersey, and those of more northern localities {86: 79), though he did not consider 

 those differences wide enough for recognition in nomenclature. 



II. For their diet in fresh water, with extensive tables for many different lakes, see Kendall, 56: 281-287. 



