332 



BULLETIN OF THE BUREAU OF FISHERIES 



their time of spawning by a period as great or greater. Inasmuch, then, as spawning 

 time is so variable, the deviation in this particular of the reighardi and nigripinnis 

 forms loses significance as a specific character. In spite of marked changes in a few 

 systematic characters, the varieties still closely resemble the typical members of their 

 species group; in fact, the resemblance is far closer than that between the whitefish 

 artificially reared in the New York Aquarium and their parents. In the case of 

 reighardi, individuals in northern Lake Michigan even show a tendency to vary from 

 the typical form (which occurs in the southern part of that lake) in precisely the 

 same direction that has produced the dymondi type. 



The systematic scheme here outlined presupposes the presence of 10 distinct 

 species in the Great Basin before the close of the glacial period. The facts of the 

 geological history of the lakes do not contradict the assumption. Two species of 

 coregonids, artedi and clupeqformis, are distributed in all six of the Great Lakes. 

 Lake Erie is so shallow and warm over most of its extent that probably it is unsuited 

 for any but the most adaptable coregonids; in fact, the bulk of its population is made 

 up of species that thrive best in the bays of other lakes. Three more, nigripinnis, 

 hoyi, and quadrilaterale, occur in each of the other five lakes; reighardi and Tciyi are 

 found in four of the lakes, johannx and alpenx occur in only two and nipigon in only 

 one. 



Most of the facts of the distribution of the species can be explained by assuming 

 the presence in the basin of an original stock of 10 coregonids. Only 2 of the 10 sur- 

 vived in all the lakes. None of the rest survived in the warmest lake, but three 

 species found suitable conditions in the other five lakes. One {reighardi) survived in 

 all the lakes but Huron, and one (Tciyi) survived in all but Nipigon. Why their range 

 was thus limited is not clear from any known geological facts. One (zenithicus) is 

 absent from Lake Ontario; it may have been unable to return after the late marine 

 inundation of old Admiralty Lake. Two (johannas and alpenx) occur only in Lakes 

 Huron and Michigan; they may have originated in one or other of the lakes subse- 

 quent to the Lake Algonquin stage, as these lakes are rather intimately connected, or 

 they simply may have perished in the other lakes from competition or from failure 

 to find suitable conditions. 



The restriction of nipigon to Lake Nipigon (though it is known to occur also in 

 Lake Winnipeg) seems to support the view that Lake Agassiz was not contemporane- 

 ous with Lake Algonquin, which joined intimately Nipigon with Superior and the 

 other lakes but came later. In that case the elevation of the basin of Lake Nipigon 

 with the resultant falls that now prevent exchange between the fish faunas of the 

 two lakes has operated to prevent the spread of nipigon into the other lakes. 



It seems, thus, that the Great Lakes coregonids were differentiated specifically 

 before the Great Lakes attained their present form. The present distribution of 

 the species, considering Lake Erie to be unfit for most coregonids, can be explained 

 by assuming the extinction of three forms, one in each of the three lakes, by the sur- 

 vival of two forms only in two of the lakes or by their late origin in one or the other of 

 the lakes, and by the assumption that Lake Agassiz came later than the Algonquin 

 stage of the Great Lakes. As we know that the last period of glaciation fell within 

 relatively recent time, geologically speaking, it is likely that many of the present racial 

 distinctions originated during the 20,000 years that geologists estimate have elapsed 

 since the formation of something like the present lakes. 



