Skip booklet index and go to page content

Sockeye Salmon (Oncorhynchus Nerka)

COSEWIC Status Report
on the
Sockeye Salmon
Oncorhynchus nerka
Sakinaw population
in Canada

Species Information

Name and Classification

Sockeye salmon, Oncorhynchus nerka Walbaum 1792, is in the order Salmoniformes, family Salmonidae, and is one of seven Canadian species in the genus Oncorhynchus, of which five are Pacific salmon and two are trout (Smith and Stearley 1989, Stearley and Smith 1993). Common names include “blueback salmon”, “red salmon”, and “saumon rouge” or “saumon nerka” when anadromous, and “kokanee” or “little redfish”, among others, when non-anadromous (Scott and Crossman 1973).


Oncorhynchus species are distinct from other salmonids in having 12 or more rays in their anal fin. Sockeye salmon are unique from other Oncorhychus salmonids by having 28 to 40 long, slender, closely spaced gill rakers on the first arch, relatively few (45-115) pyloric caeca, and fine black speckling on their back (Hart 1973, Mecklenburg et al 2002). At sea, both sexes are metallic dark blue to greenish blue on the head and back, with silver sides fading to white below. At spawning, they become red with olive green heads. Males are more brilliantly coloured and develop elongate hooked jaws and humped backs at maturity (Figure 1). Sockeye salmon can reach a total length of 84 cm and weigh up to 7 kg, but their spawning size varies with age of maturity; both age of maturity and size at age vary widely among populations (Foerster 1968). Precocious males (“jacks”), which spend only one winter at sea, are common in some populations (Burgner 1991). Kokanee typically mature at a smaller size and may lack brilliant colouration because they feed on small freshwater zooplankton throughout their life.

Figure 1: Mature Sockeye Salmon (female above, male below)

Figure 1: Mature sockeye salmon (female above, male below) (reprinted from Scott and Crossman 1973).

Reprinted from Scott and Crossman 1973.

Nationally Significant Populations

Like most salmon, sockeye salmon exist as reproductively isolated populations; however, sockeye salmon populations are discrete at a much smaller geographical scale than most other salmon (Wood 1995). This is because juvenile sockeye salmon typically rear in nursery lakes, which by their nature are discontinuous and geographically isolated, and often very different in physical and biotic characteristics (e.g., temperature and water flow regimes, nutrients, light penetration and primary productivity, competitors and predators, parasites and diseases, and factors that challenge anadromous migration). Reproductive isolation among sockeye salmon populations inhabiting different lake environments promotes the evolution of unique adaptations to the local freshwater environment. Consequently, sockeye populations can differ considerably in life history traits and phenotypic characters (reviewed by Foerster 1968, Burgner 1991). The special significance of fine scale population structure and local adaptation in sockeye salmon is reflected in decisions by the US National Marine Fisheries Service (NMFS) that individual nursery lakes may be Evolutionarily Significant Units (ESUs): e.g., Redfish Lake (Snake River), Osoyoos Lake (Okanogan River), Lake Wenatchee, Quinault Lake, Ozette Lake, Baker Lake (Baker River), and Lake Pleasant (Gustafson et al. 1997). Two of these, Redfish Lake and Ozette Lake, have been listed under the US Endangered Species Act as Endangered and Threatened, respectively.

The sockeye population in Sakinaw Lake (henceforth Sakinaw sockeye) warrants similar designation as an ESU (or Nationally Significant Population). An ESU is defined as a population or group of populations that (1) is substantially reproductively isolated from other conspecific population units, and (2) represents an important component of the evolutionary legacy of the species (Waples 1991). As described by Gustafson et al. (1997), designation of an ESU follows a two-part test: reproduction isolation, and local adaptation (evolutionary legacy). Sakinaw Lake sockeye qualify under both parts of this test.

Evidence for Sakinaw sockeye reproductive isolation -- Several surveys of genetic variation in allozymes (Wood et al. 1994), microsatellite DNA (msatDNA, Nelson et al. 2003) and mitochondrial DNA (mtDNA, Murray and Wood 2002, Wood unpubl. data) demonstrate significant reproductive isolation between Sakinaw sockeye and other anadromous sockeye populations in the region (Figure 2). Pairwise-FST statistics1 based on comparisons of allele frequencies at 10 msatDNA loci between Sakinaw Lake sockeye and the nearest other sockeye populations range from 0.06 (Koeye Lake, Area 9) to 0.13 (Heydon Lake, Area 13 and Nimpkish River (Woss Lake) in Area 12) (Table 1, above diagonal; some of these lakes are shown in Figure 7). These values (0.06 – 0.13) are large relative to those observed in other species over comparable distances and suggest that successful reproduction following immigration into Sakinaw Lake from other populations has been very rare; estimates of historical gene flow are less than 4 and 2 successful migrants per generation, respectively, under the usual assumptions for calculating gene flow based on allele frequencies at equilibrium between genetic drift and migration (Wright 1951).

Figure 2: Principal Components Analysis of Cavalli-Sforza and Edwards' Genetic Distance Between Central Coast Sockeye Populations Based on Differentiation at 10 Microsatellite DNA Loci (from Nelson et al. 2003).

Figure 2: Principal components analysis of Cavalli-Sforza and Edwards' genetic distance between central coast sockeye populations based on differentiation at 10 microsatellite DNA loci (from Nelson et al. 2003).

Pie diagrams indicate relative frequencies of mitochondrial DNA haplotypes (haplotype #1 is shown as white, haplotype #5 as grey, all others as black). Fraser River populations are included for comparison because they were the source of attempted transplants to Sakinaw Lake (from Murray and Wood 2002).

With one exception, pairwise-FST statistics based on comparisons of mtDNA haplotype frequencies (Table 1, below diagonal) range from 0.33 (Atnarko river system, Area 8) to 0.60 (Heydon Lake), indicating even lower rates of gene flow (0.5 to 0.2 female migrants per generation) than those based on allele frequencies in nuclear DNA. The exception is Kimsquit Lake which is indistinguishable using mtDNA; however, a very large difference in allele frequency (16% versus 66%) at the PGM-1 locus, and smaller differences at two other allozyme loci (Wood et al. 1994), together with the msatDNA differences in Table 1 (FST =0.09), confirm that this is a coincidental result of random genetic drift rather than continuing gene flow between Kimsquit and Sakinaw lakes.

Evidence for local adaptation -- Sakinaw sockeye are distinct from other sockeye populations in the Pacific Northwest (data summarized by Gustafson et al. 1997) in terms of their early and protracted river-entry timing, extended lake residence prior to spawning, small body size and low fecundity at spawning, large size at smolting and unusual incidence of age 2+ smolts despite large size at age 1+. These characteristics are described further in the ‘Biology’ section.

Table 1 (populations 1 to 10): Pairwise Fst Statistics for Mitochondrial DNA (below diagonal, from Murray and Wood 2002) and Microsatellite DNA (above diagonal, from Nelson et al. 2003)

PopulationSample sizePopulation number
1Upper Fraser158--0------------------
3Birkenhead River25--0.360.390--------------
4Weaver Creek23--
5Harrison Rapids25--
7Pitt (Widgeon)13--0.530.400.840.200.430.520------
14Atnarko River79520.260.090.440.


Table 1 (populations 11 to 21): Pairwise Fst Statistics for Mitochondrial DNA (below diagonal, from Murray and Wood 2002) and Microsatellite DNA (above diagonal, from Nelson et al. 2003)

PopulationSample sizePopulation number
1Upper Fraser158------------------------
3Birkenhead River25------------------------
4Weaver Creek23------------------------
5Harrison Rapids25------------------------
7Pitt (Widgeon)13------------------------
14Atnarko River79520.260.040.0000.


Local adaptation accounts for the widespread failure of attempts to transplant sockeye salmon runs from one lake system to another (Withler 1982, Wood 1995) or of restoring wild salmon populations in modified habitat (Williams 1987). Mitochondrial DNA data reported by Murray and Wood (2002, Table 1) provide compelling evidence that all five attempts (each year from 1902-1906) to transplant sockeye fry to Sakinaw Lake from various locations in the lower Fraser River and from Shuswap Lake have failed. Only two mtDNA haplotypes (distinct maternal lineages) were found in adult sockeye spawning in Sakinaw Lake in 1988, 2000, and 2001. These are designated haplotype#5 and haplotype#1. Haplotype#5 was predominant in Sakinaw Lake sockeye at a frequency of 88% (±12% 19 times out of 20). But haplotype#5 was absent in samples from the Fraser River, including samples from all of the original donor lake systems. Except for haplotype#1, none of the haplotypes observed in the donor lake systems (#1, 2, 3, 4 and 6) were observed in Sakinaw Lake. Haplotype#1 is almost ubiquitous throughout the whole Asian and North American range.

To save the hypothesis that transplanted sockeye may have survived in Sakinaw Lake, it would be necessary to assert that the mtDNA samples are not representative, for reasons not yet understood, and that more extensive sampling would change these conclusions; or that haplotype composition has changed such that the Fraser River donor populations once had a very high proportion of fish carrying haplotype#5 and that these have died out; or that only a minority of transplanted fish (those carrying haplotype#1) survived in Sakinaw Lake. Because the haplotypes differ only in a few redundant nucleotides (third base pairs), they are almost certainly not expressed phenotypically and are considered "invisible" (neutral) to natural selection. Thus, such postulated changes in haplotype composition could only occur by chance (genetic drift) and would be extremely unlikely given the sample size of introduced fish (380 000 fry over five years).

In conclusion, Sakinaw sockeye warrant designation as an Evolutionarily Significant Unit or COSEWIC Nationally Significant Population based on the two-part test developed to define salmonid “species” under the US Endangered Species Act. Protein electrophoresis and molecular DNA analyses indicate that Sakinaw sockeye are substantially reproductively isolated from other populations. Their distinctive life history characteristics (early river-entry timing, protracted adult run timing, extended lake residence prior to spawning, small body size, low fecundity and large smolt size) suggest that they are also evolutionarily distinct from other sockeye populations in North America. The evidence for very restricted gene flow between Sakinaw and other populations, and the distance to the nearest extant sockeye population both confirm that there is virtually no possibility of natural rescue from neighbouring sockeye populations. All previous attempts to transplant sockeye to Sakinaw Lake have almost certainly failed. Consequently, we cannot be optimistic about prospects for re-establishing a sockeye run to Sakinaw Lake if the native population were to become extinct.


1 FST statistics are a measure of genetic differentiation among populations commonly used to infer gene flow.