RELATIONSHIPS AMONG NORTH AMERICAN SALMONIDAE 



By George A. Rounsefell, Fishery Research Biologist 

 Bureau of Commercial Fisheries 



This paper is third in a series in which I am 

 attempt inji to compile and evaluate published in- 

 formation on North American Salmonidae. Defi- 

 nition of the relationship among species is ex- 

 tremely complex and although I would preferably 

 avoid the subject, it must necessarily Ije considered 

 in order to decide on the grouping of taxa for 

 evaluating the significance of various life-history 

 phases. In such a plast ic group as the Salmonidae 

 there are all shades of differentiation from the 

 species down almost to the individual. With our 

 present knowledge, probably the best we can hope 

 to do is to gain some appreciation of the relative 

 closene.ss of the relationships between taxa. 



Basically, we are not so much concerned with 

 whether two populations of any one species of 

 Salmonidae differ phenotypically as we are with 

 their response to similar habitats. Differences in 

 physiological reactions may be just as real as those 

 morphological differences which can be demon- 

 strated statistically. 



In our zeal to be objective and quantitative, we 

 must not overlook many of the nonmorphological 

 characteristics that, although perhaps more diffi- 

 cult to assess, nonetheless may show very real dif- 

 ferences. I am speaking of such things as color, 

 spawning habits, migratoiy tendency, growth rate, 

 age at maturity, attainable size, temperature toler- 

 ance, and doubtless other yet undefined character- 

 istics inherent in different sti-ains. 



The use of such new approaches as serological 

 techniques and paper chromotography may fur- 

 nish a clue to differences not readily discovered 

 by the classical morphological approach. Counts 

 of the chromosomes, while rendered difficult by the 

 large numbers involved, may be of great taxo- 

 nomic value, at least at the species levels. 



In discussing classification of the Salmonidae it 

 is instructive to commence by observing the rela- 

 tionships among the North American genera. 

 Following the basic work done by Vladykov 

 (1954) we chose tentatively to consider Crist l- 

 vomer as a separate genus, resulting in four North 



Approved for publication, August 18, 19G1, Kishory liullotin 

 209. 



American genera, Cristivomer, Salvelinu-s, Salmo, 

 and Oncorhynchun. 



Since all salmonids spawn in fresh water (pre- 

 sumably their ancestral home), the anadromous 

 habit may have evolved gradually from popula- 

 tion pressure and a higher survival of fish feeding 

 in the sea. 



In the genus Gristivomer this seagoing habit (if 

 ever present) is almost if not entirely lost. The 

 genus extends in lakes with sufficient cool oxygen- 

 ated water in summer (only deep, stratified eu- 

 trophic lakes toward the southern part of its 

 range) across North America from arctic Alaska 

 to eastern Quebec. Since it is lacustrine and 

 seldom enters streams, the fact that only one 

 species, C. TMmaycush, covers this entire area 

 might seem a little surprising; usually long- 

 isolated populations tend to develop distinguish- 

 able morphologic differences. This lack of differ- 

 ences over such an extended range might be cited 

 to postulate a theory of fairly recent origin for the 

 genus, which however is geologically untenable; 

 but there may be other reasons why differences 

 failed to develop. Differences between isolated 

 populations usually develop through environ- 

 mental selection. In stream-dwelling fish where 

 environmental differences between localities are 

 often large the selection may be rather severe, but 

 Cristivomcr inhabits a relatively stable lacustrine 

 habitat that differs little from lake to lake. 

 Furthermore, most geneticists support the postu- 

 late (National Research Council, 1956, p. 16) that 

 mutations are induced by naturally occurring radi- 

 ation : "To the best of our present knowledge, if we 

 increase the radiation by X7c, the gene mutations 

 caused by radiation will also be increased by X%." 



Folsom and Ilarley (1957), from data of Libby 

 (1955) and George (1952), have estimated that 

 radiation from cosmic rays at latitudes midway 

 between the geomagnetic equator and 55° N. (geo- 

 magnetic) decreases, because of the shielding effect 

 of the water, from 35 millirads per year at the 

 water surface to 10.1 millirads at 10 metei-s, -1.86 at 

 20 meters, 1.40 at 50 meters, and only 0.47 milli- 

 rads per year at 100 meters. Folsom and Harley 



235 



