330 



BULLETIN OF THE BUREAU OF FISHERIES 



deeper bodied and fewer scaled than their relatives of colder waters. (Nothing is 

 known about the characteristics of the bay races of whitefish in Lake Huron.) 



The artedi of Nipigon are nearest to the albus type, even though Lake Nipigon 

 lies in the highest latitude of the chain; but Lake Nipigon also is much shallower 

 than any of the lakes except Erie, and its annual heat budget is relatively high 

 (Clemens, 1923). 



It should be repeated that facts do not warrant the assignment of temperature 

 as a direct factor in occasioning the variations discussed. The cases cited may be 

 coincidences or temperature may act indirectly in numerous ways. The segrega- 

 tion of these variants presumably is a result of physiological differences, differences 

 that have enabled certain individuals to meet the conditions arising from increased 

 warmth or, in the case of the deep-water variant of hoyi in Huron, from increased 

 depth. The segregated variants thus are subjected to unlike physical conditions. 

 They differ in certain structural characters. If we assume that the structural 

 differences result from isolation, they may be, in part, the direct or indirect effect 

 of environment (somatic) and in part the result of germinal changes. 



Somatic variations might be the direct effect of the activities of the fish in its 

 relation to the degree of mass movement of the water, the abundance and character 

 of food, or of other factors. These should affect the form and proportions of the 

 body through the degree of induced development of muscles or fat. Differences of 

 this sort are well known between individuals of certain species of fresh-water fish 

 taken from different environments, as in the case of the yellow perch. Such somatic 

 variations may be "adaptive," as in the case of alteration of form or proportions 

 due to the degree of development of body muscles. It is also conceivable that 

 differences in physical conditions affect directly the early-growth stages of fish in 

 different enviromnents in such a way as to give rise to somatic variations that are 

 nonadaptive, indifferent, or even harmful. Such variation may appear in "passive" 

 structures such as the skeleton (Jordan, 1892). The monstrosities that often arise 

 from ova developed in hatcheries probably are, in part, an extreme instance of this 

 type of "variation," as are the monsters produced under experimental conditions. 



At the same time isolation presumably is accompanied by germinal changes 

 that become manifest in heritable somatic alterations. As in the case of mutations, 

 the adaptiveness of these is wholly contingent. They may or may not prove to be 

 useful. It is possible that the variation in number and form of the gill rakers is of 

 this type. A detailed study of the food of the Lake Huron forms described in this 

 paper indicates that within the genus Leucichthys the relation between the number 

 and form of the gill rakers and the character of the food is very loose. All the deep- 

 water forms of the genus have long, slender rakers, but these differ in number and 

 length in such a way as to be characteristic of species and varieties and thus afford 

 one of the most valuable diagnostic characters. Yet there appears to be very little 

 difference in the food of these forms, which consists chiefly of the schizopod crustacean 

 Mysis relicta. Living with the deep-water coregonids is the lawyer, Lota maculosa, 

 virtually devoid of gill rakers but often found with its stomach filled with Mysis. 

 The little knowledge that we possess thus suggests that the mean differences in gill 

 rakers characteristic of the coregonid forms are of germinal origin and not primarily 

 of individually adaptive nature. In that case such relation as they now bear to the 



