556 Memoir Sears Foundation for Marine Research 



reported for Vigo in northern Spain (5: 132); the Arctic coasts of Eurasia, Alaska, 

 and Canada eastward as far as the estuary of the Mackenzie River; the western 

 Pacific southward to northern Japan and northern China; and the eastern Pacific to 

 Yakutat Bay, Alaska. They are landlocked in many lakes of Europe, including the 

 Volga River system and the basin of the Caspian Sea (^^y. 241, 308, 315, 323, 346; 

 4g: 102; 95: 649), eastern North America, eastern Asia, and the White Sea-Kola 

 Peninsula region. We should perhaps emphasize that, while distributed widely, 

 smelts are unknown along the Arctic coast of Canada east of the Mackenzie, 

 in Hudson Bay, Greenland, Iceland, Spitsbergen, and along the coast of western 

 Norway. 



Species. The members of the genus Osmerus as limited above were originally 

 described under two names: (i) Sa/mo eperlanus Linnaeus 1758 for the smelts of 

 Europe, with which spirinchus Pallas 181 1 (7J: 387), from the lakes and rivers of 

 Russia,* and vulgaris Gaimard 1838 {2g: pi. 18, fig. 2), from Iceland, clearly are syno- 

 nyms; and (2) Atherina mordax Mitchill 18 15 for western Atlantic smelts. S. eperlanus 

 and A. mordax (as viridescens LeSueur 1 8 1 8) were retained as separate species, first by 

 Cuvier and Valenciennes, chiefly, it seems, on the assumption that LeSueur would have 

 recognized at first sight a fish with which he had doubtless been familiar since child- 

 hood {ig: 388), and then by Gunther (j^: 166, 167). The list was soon enlarged with 

 sergeanti, proposed by Norris in 1868 for the smelts of the Schuylkill River (Delaware), 

 of Brandywine Creek (Pennsylvania), and of the tidal part of the Delaware River {yo : 

 93); dentex, proposed in 1870 by Steindachner and Kner for those of northern China 

 (loy. 429); abbotti and spectrum in 1871 by Cope for those of W^ilton Pond and 

 "Cobessicontic" Lake, Maine (j6: 490); and dvinensis in 1882 by Smitt for those of 

 the White Sea {102: 34; 104: 867, ftn. k). 



So far as we can learn, Rice's comparison of the number of transverse rows of 

 scales on smelts from various American localities with others from Liverpool, England 

 {86: 91), was the earliest comparison of the sort that has been made; he concluded 

 that the American smelts are identical to the European eperlanus. The next landmark 

 in smelt taxonomy was Smitt's (JOJ: 168-188) very detailed comparison of the pro- 

 portional dimensions for the smelts of Scandinavia (as eperlanus)., the White Sea (as 

 spirinchus), and the North Pacific (as dentex)\ from this he concluded that the three 

 represented three distinct species. In 1895, however, he followed Rice in uniting the 

 smelts on the two sides of the Atlantic under the older name eperlanus {104: 869). 

 Since that time opinions have differed widely as to how many recognizably distinct 

 units (species or subspecies) the eperlanus-dentex-mordax group actually includes. 



One extreme among students concerned primarily with the smelts of North America 

 may be represented by Jordan, et al., who recognized four distinct species {^2: 67): 

 eperlanus (tacitly), mordax (with spectrum and abbotti as subspecies), dentex, and sergeanti. 

 The other extreme may be illustrated by Fowler's union of mordax, including sergeanti, 



8. The Smelt from Kamchatka which Pallas reported as spirinchus belong to the form that has subsequently been na.Tied 

 dentex. 



