U.S. Dept Commerce/NOAA/NMFS/NWFSC/Publications

NOAA-NMFS-NWFSC TM-33: Sockeye Salmon Status Review (cont)
Back to Previous Section

Life History of Oncorhynchus nerka

With the exception of certain river-type and sea-type populations, the vast majority of sockeye salmon spawn in or near lakes, where the juveniles rear for 1 to 3 years prior to migrating to sea. For this reason, the major distribution and abundance of large sockeye salmon stocks is closely related to the location of rivers that have accessible lakes in their watersheds for juvenile rearing (Burgner 1991). Although there are no commercially exploited sockeye salmon populations north of the Kuskokwim River in Alaska, small populations occur in the Yukon River and rivers flowing into Norton Sound (L. Buklis4), and perhaps in the Noatak River in Kotzebue Sound (Atkinson et al. 1967). In North America, the two dominant areas of sockeye salmon production occur in areas with extensive lake-rearing habitat: the Bristol Bay watershed in Alaska (Kvichak, Naknek, Ugashik, Egegik, Wood, and Nushagak Rivers) and the Fraser River in British Columbia. Other watersheds with major sockeye salmon stocks include the Chignik, Karluk, and Copper Rivers, and rivers draining into Cook Inlet in Alaska; and the Skeena, Nass, and Somass Rivers, and Rivers and Smith Inlets of British Columbia (Ricker 1966, Aro and Shepard 1967, Atkinson et al. 1967, Poe and Mathisen 1981, Burgner 1991). In Asia, the major sockeye salmon producing systems are on the Kamchatka Peninsula: the Kamchatka River, draining east through central Kamchatka; the Paratunka River in south-eastern Kamchatka; and the Ozernaya and Bolshaya Rivers in southwestern Kamchatka (Hanamura 1967, Burgner 1991). Kuril Lake in the Ozernaya River Basin on the Kamchatka Peninsula produces nearly 90% of all Asian sockeye salmon. Approximately 8% of sockeye salmon production in Asia comes from the Kamchatka River, while all other systems account for only about 2% (N. V. Varnavskaya5).

Sockeye salmon exhibit a greater variety of life history patterns than either chum, coho, chinook, or pink salmon. The vast majority of sockeye salmon spawn in either inlet or outlet streams of lakes or in lakes themselves. The offspring of these "lake-type" sockeye salmon utilize the lake environment for juvenile rearing for 1, 2, or 3 years and then migrate to sea, returning to the natal lake system to spawn after spending 1, 2, 3, or 4 years in the ocean. However, some populations of sockeye salmon spawn in rivers without juvenile lake rearing habitat. The offspring of these riverine spawners utilize the lower slow-velocity sections of rivers as the juvenile rearing environment for 1 or 2 years ("river-type" sockeye salmon), or migrate to sea as underyearlings after spending only a few months in the natal river and therefore rear primarily in saltwater ("sea-type" sockeye salmon) (Gilbert 1918, Foerster 1968, Wood 1995). In common with lake-type sockeye salmon, river/sea-type sockeye salmon return to the natal spawning habitat following 1 to 4 years in the ocean.

Certain populations of O. nerka that become resident in the lake environment over long periods of time are called kokanee, silver trout, or little redfish in North America and himemasu in Japan (Burgner 1991). Occasionally, a proportion of the juveniles in an anadromous sockeye salmon population will remain in the rearing lake environment throughout life and will be observed on the spawning grounds together with their anadromous siblings. Ricker (1938) defined the terms "residual sockeye" and "residuals" to identify these resident, non-migratory progeny of anadromous sockeye salmon parents. Kokanee and residual or resident sockeye salmon are further discussed in the "Nonanadromous forms" section below.

Sockeye salmon exhibit the greatest diversity in selection of spawning habitat among the Pacific salmon. Sockeye salmon also exhibit great variation in river entry timing and the duration of holding in lakes prior to spawning. Although sockeye salmon typically spawn in inlet or outlet tributaries of a nursery lake, they may also spawn in 1) suitable habitat between lakes, 2) along the nursery lakeshore on outwash fans of tributaries or where upwelling occurs along submerged beaches, 3) along beaches where the gravel or rocky substrate is free of fine sediment and the eggs can be oxygenated by wind-driven circulation, or 4) in mainstem rivers without juvenile lake-rearing habitat (Foerster 1968, Burgner 1991).

Adaptation to a greater degree of utilization of lacustrine environments for both adult spawning and juvenile rearing has resulted in the evolution of complex timing for incubation, fry emergence, spawning, and adult lake entry that often involves intricate patterns of adult and juvenile migration and orientation not seen in other Oncorhynchus species (Burgner 1991). Adult sockeye salmon home precisely to the natal stream or lake habitat (Hanamura 1966, Quinn 1985, Quinn et al. 1987). Stream fidelity in sockeye salmon is thought to be adaptive, since this ensures that juveniles will encounter a suitable nursery lake. Wood (1995) inferred from protein electrophoresis data that river/sea-type sockeye salmon have higher straying rates within river systems than lake-type sockeye salmon.

Velsen (1980) reported that, at a constant temperature of 10oC, sockeye salmon had the longest incubation period to 50% hatch of five salmon species tested. Benefits of intergravel incubation include protection from predation, freezing, fluctuating flows, and desiccation. Survival during incubation is influenced by environmental conditions, the degree of crowding during spawning (Foerster 1968, Burgner 1991), the type of gravel in which eggs are laid, and the gravel's permeability to water (Foerster 1968). Little is known about predation during incubation. Desiccation, freezing (in northern latitudes), low oxygen due to siltation, and dislodgement by later spawning fish are important mortality factors (Burgner 1991). In addition, severe water-flow changes (floods or drought) can lead to heavy losses during incubation (Foerster 1968).

Upon emerging from the substrate, sockeye salmon alevins exhibit varied behavior: 1) Inlet spawners may proceed downstream to the nursery lake or remain in the stream and show substantial growth before migrating downstream; 2) lakeshore beach spawners take up residence directly in the lake; 3) outlet spawners may require a period of growth before migrating upstream to the nursery lake; and 4) riverine spawners without lake access travel downstream to backwater sections of the lower river, where they may rear for a short period before going to sea as underyearlings (sea-type sockeye salmon) or they may rear for longer periods prior to going to sea in their second or third year of life (river-type sockeye salmon). Predation on migrating sockeye salmon fry varies considerably with spawning location (lakeshore beach, creek, river, or spring area). Sockeye salmon fry mortality, due to predation by other fish species and birds, can be extensive during downstream and upstream migration to nursery lake habitat and is only partially reduced by the nocturnal migratory movement of some fry populations (Burgner 1991). Predation losses during fry migration to Lakelse Lake, British Columbia down Scully Creek were estimated at 63-84% over 4 years (Foerster 1968). In Karymaisky Spring (Bolshaya River, Kamchatka) predation losses ranged from 13% to 91% over 8 years (Semko 1954). In the Cedar River, predation losses of sockeye salmon fry migrating to Lake Washington were 25% and 69% in two separate tests (Stober and Hämäläinen 1980), while 15% of the migrating sockeye salmon fry in the Cedar River in 1985 were eaten by wild steelhead smolts (Beauchamp 1995).

Juvenile sockeye salmon in lakes are visual predators, feeding on zooplankton and insect larvae (Foerster 1968, Burgner 1991). In certain lakes (Wood River lakes in Alaska, Lake Dalnee in Kamchatka, and Babine Lake in British Columbia), sockeye salmon fry feed initially in the littoral zone and subsequently migrate offshore to the limnetic zone (various references in Burgner 1991). In other lakes (Cultus Lake beach spawners and fry migrating from the Cedar River to Lake Washington), fry move directly into the limnetic zone upon reaching the nursery lake (Brannon 1972b, Woodey 1972, Dawson 1972). Although previous studies have not found sockeye salmon fry in the littoral zone of Lake Washington or other lakes in Washington (WDFW 1996), Martz et al. (1996) reported that

The majority of the sockeye fry were found in the limnetic zone; however, a smaller but significant number of sockeye fry also utilize the littoral zone for up to one month after emigrating from the Cedar River.

Juvenile sockeye salmon in lakes commonly undergo diel vertical migrations such that they are present in deeper water by day than by night (Levy 1987). In Lake Washington, juvenile sockeye salmon were reported to undergo diel vertical migrations at all times of year and to occupy certain depths in direct relation to water temperature (Woodey 1972). The surface area and productivity of a nursery lake limit population size of sockeye salmon, and offspring of large return years may show reduced growth due to intraspecific competition. Increased growth in freshwater may lead to higher marine survival and decreased ocean age at return.

Smolt migration typically occurs between sunset and sunrise, beginning in late April and extending through early July, with southern stocks migrating earliest. Some sockeye salmon smolts undergo a complicated migration to reach the lake system outlet (Johnson and Groot 1963). Once in the ocean, sockeye salmon feed on copepods, euphausiids, amphipods, crustacean larvae, fish larvae, squid, and pteropods. Increase in length is typically greatest in the first year of ocean life, whereas increase in weight is greater during the second year. Northward migration of juveniles to the Gulf of Alaska occurs in a band relatively close to shore, and offshore movement of juveniles occurs in late autumn or winter. Sockeye salmon prefer cooler ocean conditions than other Pacific salmon (Burgner 1991).

Lake-Type Sockeye Salmon

The vast majority of sockeye salmon typically spawn in inlet or outlet tributaries of lakes or along the shoreline of lakes where upwelling of oxygenated water through gravel or sand occurs. Growth influences the duration of stay in the nursery lake and is influenced by intra- and interspecific competition, food supply, water temperature, thermal stratification, migratory movements to avoid predation, lake turbidity, and length of the growing season. Lake residence time is usually greater the farther north a nursery lake is located. In Washington and British Columbia, lake residence is normally 1 or 2 years, whereas in Alaska some fish may remain 3, or rarely 4 years in the nursery lake, prior to smoltification (Burgner 1991, Halupka et al. 1993).

While in the lacustrine environment, fry and yearlings feed as visual predators, primarily on copepods (Cyclops, Epischura, and Diaptomus), cladocerans (Bosmia, Daphnia, and Diaphanosoma), and insect larvae. In some lakes, sockeye salmon fry initially feed near the lake shoreline in the littoral zone, subsequently shifting to the deeper waters of the limnetic zone. In other lakes, sockeye salmon fry enter the limnetic zone directly. In many lakes, juveniles feed in the limnetic zone at dusk and dawn to avoid day-time visual predators, and this feeding pattern may be tied to diel vertical migrations (Eggers 1978, Pauley et al. 1989, Burgner 1991). In summer and fall 1972, juvenile Lake Washington sockeye salmon fed intensively during the afternoon through dusk. In winter 1972-1973, a high percentage of the population did not feed on a daily basis. No feeding occurred at night during any season of the year (Doble and Eggers 1978).

Competitors for common food of sockeye salmon during lake residence may include threespine and ninespine sticklebacks (Gasterosteus aculeatus and Pungitius pungitius), red sided shiner (Richardsonius balteatus), pond smelt (Hypomesus olidus), pygmy whitefish (Prosopium coulteri), lake whitefish (Coregonus clupeaformis), northern squawfish (Ptychocheilus oregonensis), yellow perch (Perca flavescens), peamouth (Mylocheilus caurinus), and kokanee (Foerster 1968, Dlugokenski et al. 1981, Burgner 1991). Longfin smelt (Spirinchus thaleichthyes) are reportedly an important competitor with sockeye salmon in Lake Washington (WDFW 1996). Nursery lake area and productivity coupled with inter- and intra-specific competition can exert a limiting effect on smolt size, and ultimately population size, of sockeye salmon (Kyle et al. 1988).

Potential predators on lake resident sockeye salmon fry throughout their North American range include: lake trout (Salvelinus namaycush), rainbow trout (O. mykiss), Dolly Varden charr (Salvelinus malma), Arctic charr (Salvelinus alpinus), cutthroat trout (O. clarki), juvenile coho (O. kisutch) and chinook salmon (O. tshawytscha), lake whitefish (Prosopium clupeaformis), mountain whitefish (Prosopium williamsoni), northern squawfish (Ptychocheilus oregonensis), burbot (Lota lota), northern pike (Esox lucius), prickly sculpin (Cottus asper), and bird predators (Foerster 1968, Hartman and Burgner 1972, Burgner 1991, Emmett et al. 1991, Beauchamp et al. 1995). Principle bird predators include: common loon (Gavia immer), red-necked grebe (Podiceps grisegna), common merganser (Mergus merganser), belted kingfisher (Megaceryle alcyon), osprey (Pandion haliaetus), bald eagle (Haliaeetus leucocephalus), terns, and gulls (Emmett et al. 1991).

River-Type and Sea-Type Sockeye Salmon

In areas where lake-rearing habitat is unavailable or inaccessible, sockeye salmon may utilize river and estuarine habitat for rearing or may forgo an extended freshwater rearing period (Birtwell et al. 1987; Wood et al. 1987a, Heifitz et al. 1989; Murphy et al. 1988, 1989, 1991; Lorenz and Eiler 1989; Eiler et al. 1992; Levings et al. 1995; Wood 1995). Riverine spawners that rear in rivers for 1 or 2 years are termed "river-type" sockeye salmon. Riverine spawners that migrate as fry to sea or to lower river estuaries in the same year emergence occurs, following a brief freshwater rearing period of only a few months, are referred to as "sea-type" sockeye salmon.

River-type and sea-type sockeye salmon are common in northern areas and may predominate over lake-type sockeye salmon in some river systems (Wood et al. 1987a, Eiler et al. 1988, Halupka et al. 1993, Wood 1995) (see Table 2). River/sea-type sockeye salmon have been rarely reported in rivers south of the Stikine River, although those that spawn in the Harrison River rapids and rear in the lower Fraser River system of southern British Columbia are the exception (Gilbert 1918, 1919; Schaefer 1950, 1951; Birtwell et al. 1987; Levings et al. 1995). Halupka et al. (1993) suggested that the lack of reported river/sea-type sockeye salmon stocks south of the Stikine River, with the exception of the Fraser River population, may be due to any of three factors: 1) the lack of sufficient colonists with the genetic capacity for developing this life-history pattern, 2) the lack of suitable habitat for development of a river/sea-type life history pattern, or 3) small river/sea-type stocks exist in southern rivers but their presence has been overlooked. Eiler et al. (1992) indicated that riverine spawning has been reported (if only sometimes anecdotically) throughout the range of sockeye salmon. Known self-sustaining populations of river/sea-type sockeye salmon throughout the Pacific Rim are listed in Table 2.

Many populations of river/sea-type sockeye salmon spawn and rear in close proximity to glaciers or in glacially influenced drainages. Milner and Bailey (1989) observed that sockeye salmon were one of the first salmonids to colonize clearwater streams in Glacier Bay National Park, Alaska following glacial retreat. Lorenz and Eiler (1989) noted that river/sea-type sockeye salmon in the glacial Taku River, Alaska preferred main channel or off-channel areas for spawning, where upwelling groundwater occurs. Fish spawning in these areas had, on average, two times more fine sediment in their redds than had been previously measured in other sockeye salmon redds.

Several studies have indicated that sea-type sockeye salmon possess heritable physiological adaptations for successful migration to sea as underyearlings (Rice et al. 1994, Wood 1995). Underyearling sea-type sockeye salmon from the East Alsek River, Alaska (Rice et al. 1994) and river/sea-type sockeye salmon from the Scud River in the Stikine River basin (Wood 1995) showed superior seawater adaptability over lake-type fry when exposed to similar seawater challenge. When reared in 30 ppt seawater, sea-type sockeye salmon fry from the East Alsek River grew significantly faster than river-type sockeye salmon, which in turn grew faster than lake-type sockeye salmon (Rice et al. 1994). Underyearling sockeye salmon in the Situk River, Alaska and Fraser River estuaries grow unusually fast in nature, obtaining a size similar to that of age 1+ lake-type smolts by the middle to end of their first summer (Birtwell et al. 1987, Rice et al. 1994). Juvenile sea-type sockeye salmon in the Situk River estuary in southeast Alaska rear in the estuary for 3-4 months in 0-30 ppt salinity until they are large enough to tolerate full-strength ocean salinity as underyearlings greater than 50 mm in length (Heifitz et al. 1989). Both Craig (1985) and Wood (1995) have reported that egg size in river/sea-type sockeye salmon is significantly larger than in lake-type sockeye salmon within the Stikine River Basin, and this may provide a size advantage to river/sea-type sockeye salmon fry over lake-type fry. At present, it is unknown whether these egg size differences are heritable (Wood 1995).

In studies of genetic variation of sockeye salmon in the Stikine River drainage, widely separated spawning populations of river/sea-type sockeye salmon showed much less genetic differentiation than did widely separated spawning populations of lake-type sockeye salmon (Wood 1995). This apparent lack of reproductive isolation among spawning populations of river/sea-type sockeye salmon in the Stikine River (some populations separated by over 180 km) was interpreted by Wood (1995) to indicate that precise homing may be relaxed in river/sea type sockeye salmon in systems where river/sea-type fry from throughout the watershed rear together in the lower river or estuary.

Wood (1995) speculated further that the combined traits of living in glacially influenced drainages and having higher straying rates than lake-type sockeye salmon give river/sea-type sockeye salmon the role of primary colonists of new habitat following glacial retreat.

Nonanadromous Forms

"Kokanee," for the purposes of this review, are defined as the self-perpetuating, nonanadromous form of O. nerka that occurs in balanced sex-ratio populations and whose parents, for several generations back, have spent their whole lives in fresh water. Commonly, kokanee occur in land-locked lakes, where access from the ocean has become difficult or impossible (such as Lake Whatcom, Washington). Kokanee and sockeye salmon also co-occur in many interior lakes of the Skeena, Fraser, and Columbia River Basins, where access from the sea is possible, although energetically costly. Kokanee are rarely found in easily accessible coastal lakes that contain sockeye populations and where the energetic costs of migration are minimal.

The terms "residual sockeye" and "residuals" have been used to identify resident, non-migratory progeny of anadromous sockeye salmon (Ricker 1938). Ricker (1938) was of the opinion that it would be unusual if residual sockeye salmon were not found in most lakes which have an anadromous sockeye population, although Burgner (1991) stated that residual sockeye salmon are rare or absent in most northern sockeye salmon lakes. For the purposes of this review, we have defined the term "resident sockeye salmon" to indicate those fish that are the progeny of anadromous parents, yet spend their adult life in freshwater and are observed together with their anadromous siblings on the spawning grounds. The degree to which resident sockeye salmon produce anadromous offspring is generally unknown.

Both kokanee and resident sockeye salmon are normally smaller at maturity than anadromous sockeye salmon, primarily because of productivity differences between their respective freshwater and oceanic post-juvenile rearing environments. According to Ricker (1938, 1940, 1959), Burgner (1991), and Wood (1995), "residual" or resident sockeye salmon 1) mature earlier (males earlier than females) and at a smaller size than anadromous sockeye salmon, 2) have a sex ratio biased toward males, 3) spawn in the vicinity of anadromous individuals, and 4) develop a dull olive-green spawning coloration, although this later character may not be expressed in all resident sockeye salmon populations (Ricker 1959). According to Ricker (1938, 1940, 1959), Burgner (1991), and Wood (1995), kokanee have a balanced sex ratio, spawn earlier in the year, mature at a smaller size, have more gill rakers, and absorb their scale margins to a greater degree upon maturity than anadromous sockeye salmon. These same authors stated that kokanee typically display a bright red body coloration at spawning, although this trait is not expressed in all populations of kokanee. On the other hand, Brannon (1996) argued that exceptions to the above differences in spawn timing, size at maturity, and spawning coloration between kokanee and resident sockeye salmon invalidate these criteria as characters that can be used to define the types.

All three forms (sockeye salmon, resident sockeye salmon, and kokanee) typically spawn in the vicinity of a nursery lake, die after spawning a single time, and as juveniles rear in the pelagic zone of a nursery lake. Kokanee and resident sockeye salmon remain in fresh water for their entire life cycle, whereas sockeye salmon migrate to sea following 1 to 4 years in freshwater, grow to maturity in the ocean, and return to the natal freshwater habitat to spawn following an additional 1 to 4 years at sea.

Genetic differentiation among sockeye salmon and kokanee populations indicates that kokanee are polyphyletic, having arisen from sockeye salmon on multiple independent occasions, and that kokanee may occur sympatrically or allopatrically in relation to sockeye salmon (Foote et al. 1989, Wood and Foote 1990, Foote et al. 1992, Taylor et al. 1996, Wood and Foote 1996, Winans et al. 1996). In some cases, both forms may spawn at the same time and place (Hanson and Smith 1967, McCart 1970, Foote and Larkin 1988, Foote et al. 1994), although typically kokanee spawn earlier than sockeye salmon. According to Brannon (1996), early spawning is not a universal kokanee trait, as some populations spawn at the same time or later than sympatric sockeye salmon. In the locations that have been studied where sockeye salmon and kokanee remain sympatric and spawn in the same place and time, there is a high degree of size-based assortative mating (Foote and Larkin 1988). Assortative mating by body size usually leads to assortative mating by type; kokanee with kokanee and sockeye salmon with sockeye salmon. Even where sneak-spawning by small satellite kokanee males occurs, and results in successful fertilization of sockeye salmon eggs, substantial post-zygotic isolating mechanisms between kokanee and sockeye salmon may reduce gene flow (Wood and Foote 1996).

In relation to co-occurring sockeye salmon and kokanee-sized O. nerka, McCart (1970) asked the question, "Do they constitute distinct, non-interbreeding populations or are they simply alternative life-history types arising within single populations?" McCart (1970) showed that spawning sockeye salmon and kokanee in shallow streams tributary to Babine Lake, British Columbia, overlap almost completely in their spawning season and in their distribution on the spawning grounds. He suggested that hybridization between the forms probably occurs under natural conditions. Foote et al. (1989) studied genetic relatedness of sockeye salmon and kokanee from these same streams in Babine Lake and stated that "there were significant differences between sockeye and kokanee in all systems where they spawn sympatrically." In relation to the Babine Lake tributaries, Foote et al. (1989) stated that "Despite apparent interbreeding, there is an effective restriction in gene flow between sockeye and kokanee that indicates that they do not constitute a single panmictic population."

Foote et al. (1989) further showed that sympatric kokanee and sockeye salmon in each of three different lake systems in British Columbia were genetically distinct from each other, but were more similar to each other within a lake system than either was to the same morph in another lake system. Likewise, Taylor et al. (1996) showed that, based on allelic variation in mitochondrial DNA Bgl II endonuclease restriction sites and two minisatellite nuclear DNA repeat loci, genetic affinities among sockeye salmon and kokanee throughout their range in the North Pacific were organized more by geographic proximity than by life-history type. Within Takla Lake, British Columbia, Wood and Foote (1996) showed that genetic differences between kokanee and sockeye salmon in the same stream were much greater than within morph differences among either sockeye salmon or kokanee spawning in different streams in the same lake system.

However, there are exceptions to the above pattern of sockeye salmon and kokanee genetic relatedness. Winans et al. (1996) investigated the genetic similarity of sympatric populations of kokanee and sockeye salmon in the Lower Shuswap River (in the Fraser River drainage of British Columbia) and Ozette Lake on the Olympic Peninsula and found significant genetic differences between sympatric morphs in both cases. In addition, kokanee from the Lower Shuswap River were genetically more similar to kokanee from Okanagan Lake in the Columbia River drainage than they were to sympatric sockeye salmon from the Lower Shuswap River. Robison (1995) also found genetic similarity between Lower Shuswap River kokanee and Okanagan Lake kokanee, which he interpreted as suggestive of transplantation of Shuswap Lake kokanee into Okanagan Lake. Winans et al. (1996) showed that kokanee from Ozette Lake were divergent from sympatric sockeye salmon, as well as from all other contiguous U.S. stocks of O. nerka investigated.

Craig (1995) has shown that although both sockeye salmon and kokanee from Takla Lake, British Columbia exhibit similar red spawning coloration, they are genetically divergent in the ability to utilize the carotenoid pigments in the diet that, when mobilized from the muscle tissue and deposited in the skin, produce the red coloration. Carotenoids are more abundant in the marine diet of sockeye salmon than in the freshwater diet of kokanee, and apparently, kokanee in the Takla Lake population are able to compensate for this difference by being more efficient at extracting carotenoids. Craig (1995) demonstrated that when reared under identical conditions in the hatchery, Takla Lake kokanee turned red at maturity, Takla Lake sockeye salmon were olive-green, and hybrid forms were intermediate in coloration. In addition, Craig (1995, p. 25) stated that "residuals do not turn red at maturity, presumably because they lack the genetic adaptations for increased carotenoid absorption needed to turn red in freshwater."

Taylor and Foote (1991) compared sustained swimming performance and morphology of sockeye salmon, kokanee, and hybrid juveniles obtained from sympatrically spawning populations in Babine Lake, British Columbia and showed that juvenile sockeye salmon are stronger swimmers than kokanee or sockeye salmon x kokanee hybrids. Similar comparisons of developmental rate (Wood and Foote 1990), ontogeny of seawater adaptability (Foote et al. 1992), and growth and onset of maturity (Wood and Foote 1996) between juvenile sockeye salmon, kokanee, and sockeye salmon x kokanee, indicated that progeny of hybrid crosses may be less successful than progeny of pure crosses of either type in their respective environments.

Danner (1994) studied behavioral, physiological, and genetic differences among sockeye salmon, sockeye salmon x kokanee hybrids, and kokanee. Danner (1994) showed that sockeye salmon were able to adapt to a 24-hour saltwater challenge 3 to 6 weeks before either sockeye x kokanee hybrids or pure kokanee, although he suggested that this earlier onset of saltwater adaptability may be a function of the larger size of sockeye versus sockeye x kokanee hybrids or pure kokanee of the same age. Growth rates of the three types exposed to identical conditions were greatest for sockeye salmon, intermediate for the hybrids, and slowest for kokanee (Danner 1994), although survival was not significantly different among types. Pure sockeye salmon also had significantly higher interlamellar chloride cell density than sockeye x kokanee hybrids or pure kokanee. Danner (1994) also measured migratory tendency of sockeye salmon, kokanee, and their hybrids by their ability to exit rearing tanks through a modified central standpipe. Migration tendency was similar for all three forms. According to Danner (1994), mixed DNA fingerprints of seven O. nerka populations differentiated stocks by origin but failed to reveal a marker separating migratory and non-migratory O. nerka stocks. Danner (1994) concluded that kokanee stocks with "smoltification characteristics similar to anadromous stocks have not likely separated far from anadromous ancestors" and that "characteristics necessary to become anadromous are maintained" in kokanee populations.

Robison (1995) examined mtDNA genetic divergence between sockeye salmon and freshwater resident O. nerka, in four systems where these life history forms spawn sympatrically: 1) Pierre Creek in the Babine Lake Basin, British Columbia, 2) Eagle River, British Columbia, 3) the Middle Shuswap River, British Columbia, and 4) a beach spawning site in Redfish Lake, Idaho. Freshwater residents and sockeye salmon were genetically indistinguishable in the Eagle River and on beach sites in Redfish Lake, but sympatric populations of the two forms were divergent in Pierre Creek and the Middle Shuswap River (Robison 1995). Robison (1995) interpreted these data to indicate that in the Eagle River and Redfish Lake populations, either sockeye salmon or the freshwater residents have been established recently from their counterpart form.

The above studies indicate that both sockeye salmon and kokanee exhibit a suite of heritable differences in morphology, rate of early development, seawater adaptability, growth, and maturation that appear to be divergent adaptations that have arisen from different selective regimes associated with anadromous vs. non-anadromous life histories. Although these heritable differences are strongly expressed in many populations, both indirect and direct evidence exists showing that kokanee are capable of producing anadromous offspring that return from the ocean with the sockeye salmon morphology (Chapman 1941, Foerster 1947, Rounsefell 1958a, Fulton and Pearson 1981, Chapman et al. 1995) and that sockeye salmon are capable of producing freshwater resident offspring (Ward 1932; Ricker 1938, 1959; Scott 1984; Graynoth 1995).

Based on indirect evidence, Chapman (1941) and Rounsefell (1958a) postulated that sockeye salmon observed at the base of Enloe Falls on the Similkameen River, and those below falls that are a natural barrier to fish passage downstream from Lake Chelan, may have been derived from downstream passage of kokanee from up-river lakes. Similarly, sockeye salmon that have been observed at the Whatcom Creek Hatchery of the Bellingham Technical School (Bellingham, Washington) (from 6 to 8 in most years, although none were observed in 1994 and only 2 in 1995), and below natural upstream passage barriers in Whatcom Creek itself, are presumed to be derived from returns of outmigrating Lake Whatcom kokanee (E. Steele6).

Several other researchers have provided more direct evidence that kokanee may at times go to sea, survive ocean life, and then return to spawn in freshwater. Foerster (1947) released almost 64,000 marked Kootenay Lake yearling kokanee in the outlet stream of Cultus Lake in 1934 and observed 5-year-old adult sockeye salmon with these markings that returned in 1937, with a calculated survival rate of 0.14% (Chapman et al. 1995). For comparison, survival rates for sockeye smolts returning as sockeye salmon to Cultus Lake were in the range of 1.9-2.6% during this time period (Foerster 1947). According to Foerster (1947) and Ricker (1972), since Kootenay Lake kokanee were known to mature at age-3, with a few at age-2 and age-4, surviving anadromous kokanee were expected to return at age-3 or age-4, not as age-5 sockeye salmon in 1937. Marked Kootenay Lake kokanee returning to Cultus Lake exhibited spawn-timing concurrent with Cultus Lake sockeye salmon (October-November) (Foerster 1947) rather than with the spawn-timing of Kootenay Lake kokanee (August-September) (Vernon 1957). The results of this study "indicated that, when liberated in a stream below a lake and barred from ascending into the lake, some of the kokanee proceeded to sea and returned" (Foerster 1947).

Similar kokanee-marking experiments were conducted in the Columbia River Basin in the 1940s (Fulton and Pearson 1981). Fin-clipped yearling Lake Chelan kokanee released in the Entiat River and Lake Wenatchee kokanee released in Lake Wenatchee and Icicle Creek returned as adults at rates of 0.004%, 0.50%, and 0.27%, respectively. Kokanee in Lake Chelan had been introduced from Lake Whatcom kokanee transplants. In the case of Lake Wenatchee kokanee, Fulton and Pearson (1981) stated that "there was a question as to whether [these] fish were far enough removed from seaward migratory behavior to be classified as kokanee." Also in reference to Lake Wenatchee, Ricker (1972) stated that "there may still be a very incomplete separation of kokanee from sockeye at this lake" and that Lake Wenatchee kokanee may have diverged from sockeye salmon only within the past 90 years, as difficulties in migrating up the Wenatchee River have increased due to water diversions, dams, and high water temperatures. Fulton and Pearson (1981) concluded that in these experiments "adult sockeye salmon that had kokanee parents were slightly smaller than adult sockeye salmon that had anadromous parents."

Kaeriyama et al. (1992) documented returns of sockeye salmon derived from marked kokanee plants in Lake Toro, Japan. Of the 60,000 smolt-sized O. nerka released in 1988-1989 in Lake Toro, a total of 20 adult sockeye returned (0.03% of those released). According to Kaeriyama et al. (1992), the parent kokanee population from Lake Shikotsu, Japan that was used in this experiment had been derived from sockeye salmon in Lake Urumbetsu on Iturup Island that were introduced into Lake Shikotsu between 1925 and 1940 and were subsequently landlocked for 15 generations. As such they may have retained more capability for anadromy than is typical of kokanee in general. Kaeriyama et al. (1992) concluded that "both anadromous and nonanadromous types of Oncorhynchus nerka can be produced from both sockeye and kokanee salmon."

Not only may kokanee occasionally give rise to anadromous individuals, but in several documented instances sockeye salmon stocked in lakes without ocean access have developed into self-sustaining resident "kokanee" or "residual sockeye salmon" populations (Scott 1984, Kaeriyama et al. 1992). For the sake of completeness, it should also be noted that in at least two instances (Ozette Lake, Washington and Lake Cowichan, Vancouver Island), large viable kokanee populations with no documented anadromous members exist in lake basins where access to and from the sea is relatively easy (Dlugokenski et al. 1981, Rutherford et al. 1988).

Historical Distribution

Most modern sockeye salmon populations in Alaska, Canada, and northern Washington arose within the last 10,000 years following retreat of the Cordilleran ice sheet at the close of the last ice age (Wood 1995). Sockeye salmon are thought to have survived the ice ages in refugia in the Bering Sea region of Alaska, south of the ice sheet in the Columbia River, on coastal islands in British Columbia, in Kamchatka, and perhaps on Kodiak Island in Alaska (Wood 1995).

Spawning populations of sockeye salmon do not presently occur in California, and it is uncertain whether they existed there historically. Jordan and Evermann (1896) stated that sockeye salmon occurred in the Klamath River, and Scofield (1916) relates the unsubstantiated claim that 20 sockeye salmon were taken in the commercial fishery in the Klamath River in the summer of 1916. Klamath Lake was accessible to migrating salmon prior to the construction of Copco Dam on the Klamath River in 1917. However, early reports of sockeye salmon in the Klamath River may be explained by Wilcox's (1898) statement that "silver salmon are locally known as blueback" and that "blueback" is the common name for sockeye salmon on the Columbia and Quinault Rivers. Taft (1937) reported the taking of a single sockeye salmon in the Klamath River in August 1936.

Cobb (1911, p. 8) reported that "small runs [of sockeye salmon] are said to occur in Mad and Eel Rivers" in Humboldt County, whereas Jordan and Gilbert (1881a) indicated that sockeye salmon were unknown in the Eel and Sacramento Rivers. Jordan and Gilbert (1881b) did not observe sockeye salmon in the Sacramento River. Rutter (1904) reported the occurrence of a single sockeye salmon in the Sacramento River in 1899. Hallock and Fry (1967) described the recovery of 22 sockeye salmon from the Sacramento River between 1949 and 1958 and speculated as to whether these fish may have been strays, part of a remnant run, or partially derived from kokanee planted in Shasta Lake. Currently, there are no recognized runs of sockeye salmon in California, although introduced kokanee populations have been established in numerous reservoirs and lakes.

Jordan and Gilbert (1881a) indicated that sockeye salmon were unknown in the Rogue River, Oregon, whereas Jordan and Evermann (1896) stated that sockeye salmon occurred in the Rogue River. Oakley and Kruse (1963) reported the occurrence of a stray female sockeye salmon in 1961 in the Kilchis River, a tributary of Tillamook Bay. Currently, there are no recognized populations of sockeye salmon in coastal Oregon streams.

Only about 5% of the pre-1900 nursery lake habitat in the Columbia River drainage remains accessible today to sockeye salmon (Mullan 1986). Historically, two Oregon lakes within the Columbia River Basin supported populations of sockeye salmon: Suttle Lake in the Deschutes River Basin (Nielson 1950, Nehlsen 1995), and Wallowa Lake in the Snake/Grande Ronde River Basin (Mullan 1986).

Suttle Lake, at the head of the Metolius River, has a surface area of 0.1 km2 (250 acres) and probably never supported a large population of sockeye salmon. A small dam and screen installed at the outlet of Suttle Lake in 1930 (Fulton 1970, Nehlsen 1995) blocked fish passage both into and out of the lake, while a swimming pool dam, built sometime between 1925 and 1938 at Lake Creek Lodge, impeded fish passage in Lake Creek below Suttle Lake (Nehlsen 1995). Three further dams were subsequently constructed downstream of Suttle Lake, on the Deschutes River: Pelton Dam and Pelton Re-regulating Dam were constructed in 1958, and Round Butte Dam was constructed in 1964. A small number of sockeye salmon are currently observed at the base of the Pelton Re-regulating Dam each summer; however, neither their origin nor whether they spawn below the dams, is known (ODFW 1995a).

Wallowa Lake, near the head of the Wallowa River in northeastern Oregon, once supported a substantial sockeye salmon population. Bartlett (1967) indicated that following the forced removal of members of the Nez Perce Tribe in 1877 and elimination of their ceremonial and subsistence fishery based on sockeye salmon from the Wallowa Valley, seining by horse and rowboat at the head of Wallowa Lake became a small industry that produced an annual catch of about 27,216 kg (60,000 pounds) of sockeye salmon by 1881. Bendire (1881) reported the taking of several O. nerka specimens in Wallowa Lake on 31 August and 1 September 1880. A rough dam to supply water to a small shingle mill was built across the lake outlet in 1884, and a more substantial dam and irrigation ditch were constructed in 1890 (Bartlett 1967). Bartlett (1967) indicated that this latter dam blocked the migration corridor for the Wallowa Lake population and resulted in land-locked sockeye salmon, locally termed "yanks," that spawned in tributary creeks in the fall. Evermann and Meek (1898) reported that both "large and small redfish" occurred in Wallowa Lake and spawned together. The small redfish or "yanks" or "grayling" were likely residual sockeye salmon, as they were overwhelmingly males and more silvery in color than larger fish (Evermann and Meek 1898).

Several authors (CBFWA 1990, ODFW 1995a) have reported that prior to increasing the height of the dam at Wallowa Lake in 1916, sockeye salmon continued to return and spawn above the lake. The last reported sockeye salmon were apparently observed in Wallowa Lake in 1916 (Toner 1960) or 1917 (CBFWA 1990). However, Cramer (1990) stated that "sockeye were extinct from the lake by 1904." ODFW (1995a) reported that sockeye salmon were observed until the early 1930s in the Wallowa River below the lake, while Parkhurst (1950a) and Fulton (1970) reported that construction of a 12-m high concrete dam at the lake outlet in 1929 finished off the population. Cramer (1990) stated that a 4-m tall dam that existed between 1906 and 1924 at the Wallowa River Hatchery, 43 miles below Wallowa Lake, completely blocked upstream fish passage. Considerable numbers of non-native sockeye salmon were stocked in the Wallowa River below Wallowa Lake in the 1920s and 1930s (Cramer 1990; his Appendix 8), perhaps contributing to reports of sockeye salmon returning to the Wallowa River up until the early 1930s (see Appendix Table D-2).

Historically, sockeye salmon spawned and reared in the Snake River in several high mountain lakes in Idaho. In the Salmon River Basin, sockeye salmon occurred in Alturas, Redfish, Pettit, and Stanley Lakes (Evermann 1895, 1896; Evermann and Scovell 1896; Evermann and Meek 1898), and perhaps in Yellowbelly Lake (Bjornn et al. 1968, Mullan 1986, Chapman et al. 1990). Mullan (1986) stated that the presence of kokanee indicated that sockeye salmon once used Little Redfish Lake and Hell Roaring Lake in the Stanley Basin and Warm Lake on the South Fork Salmon River. In the Payette River Basin, sockeye salmon reportedly occurred in Big Payette, Upper Payette (Evermann 1895, 1896; Evermann and Scovell 1896; Fulton 1970), and Little Payette Lakes (Mullan 1986). Currently, a genetically distinct sockeye salmon population exists in Redfish Lake, Idaho and is listed as endangered under the federal Endangered Species Act. Other anadromous O. nerka populations in Idaho are thought to be extinct, although O. nerka in Alturas Lake have been known to produce outmigrating individuals. In addition, kokanee stocks in Redfish, Alturas, and Stanley Lakes are genetically more similar to sockeye salmon from Redfish Lake than they are to other O. nerka stocks investigated from outside the Stanley Basin (Winans et al. 1996, Waples et al. in press).

The history of the decline of sockeye salmon in the Stanley Basin lakes on the Salmon River was reviewed in Bjornn et al. (1968), Chapman et al. (1990), and Waples et al. (1991). Evermann (1895) observed sockeye salmon spawning in the inlet to Big Payette Lake and related that local residents informed him that between 1870 and 1880 two fisheries operated on this lake and in some years took up to 75,000 fish (or 13,600 to 18,140 kg) and that a few sockeye salmon may have gone as far as Upper Payette Lake. A diversion dam built about 1914 near Horseshoe Bend on the Payette River, and Black Canyon Dam constructed in 1923, blocked access by sockeye salmon to the Payette Lakes (Parkhurst 1950b, Fulton 1970, Mullan 1986). Current accessible lake-rearing habitat for sockeye salmon in Idaho, based on lake area, represents about 25% of historically available habitat (Hassemer et al. 1996).

Within the Columbia River Basin in Washington, historical populations of sockeye salmon existed in the Yakima, Wenatchee, and Okanogan Rivers. Sockeye salmon populations reportedly existed in two small lakes at the head of the Yakima River on the present site of Lake Keechelus, as well as in Cle Elum Lake in the Yakima/Cle Elum River Basin, in Kachess Lake in the Yakima/Kachess River Basin, and in Bumping Lake in the Yakima/Naches/Bumping River Basin (Davidson 1953, Fulton 1970, Mullan 1986). The historical total run size of Yakima River sockeye salmon has been estimated at either 100,000 (Davidson 1953) or 200,000 (CBFWA 1990). Construction of crib dams without fish passage facilities at Lakes Keechelus and Kachess in 1904 and at Lake Cle Elum in 1905 eliminated sockeye salmon populations in these lakes (Bryant and Parkhurst 1950, Davidson 1953, Fulton 1970, Mullan 1986). Construction of an impassable storage dam at Bumping Lake in 1910 likewise eliminated a sockeye salmon population in that lake, with an estimated annual run of 1,000 fish (Davidson 1953, Fulton 1970).

The native population of sockeye salmon in Lake Wenatchee was severely depleted during the early 1900s (Bryant and Parkhurst 1950, Davidson 1966, Fulton 1970), with returns counted over Tumwater Dam on the Wenatchee River in 1935, 1936, and 1937 amounting to 889, 29, and 65 fish, respectively (WDF et al. 1938). Small dams and unscreened irrigation diversions on the Wenatchee River contributed to the decline of this population (Bryant and Parkhurst 1950).

Historically, sockeye salmon are thought to have utilized Lakes Okanagan, Skaha, and Osoyoos in the Okanogan River Basin for juvenile rearing (Bryant and Parkhurst 1950, Fulton 1970, Mullan 1986). Sockeye salmon access to Lakes Okanagan and Skaha in British Columbia was blocked by dams in 1915 and 1921, respectively. Access to Lake Osoyoos remained open, but the population was severely depleted in the early 1900s (Davidson 1966, Fulton 1970), with returns to the Okanogan River in 1935, 1936, and 1937, amounting to 264, 895, and 2,162 sockeye salmon, respectively (WDF et al. 1938).

In order to preserve a portion of the sockeye salmon stocks denied access to the Upper Columbia River in 1939 by Grand Coulee Dam, the Grand Coulee Fish-Maintenance Project (GCFMP) trapped all sockeye salmon at Rock Island Dam between 1939 and 1943 and relocated them to Lakes Wenatchee or Osoyoos or to one of three national fish hatcheries (Leavenworth, Entiat, and Winthrop) for artificial propagation. Numerous descendants of artificially propagated sockeye salmon trapped at Rock Island and Bonneville Dams, together with progeny of Quinault Lake sockeye salmon, were stocked into Lakes Wenatchee and Osoyoos between 1940 and 1968 (Mullan 1986, see Appendix Table D-2). Consequently, the current populations of sockeye salmon that return to Lake Wenatchee and the Okanogan River may consist of some mixture of native and non-native fish (see "Artificial Propagation" section below).

Fulton (1970) listed Palmer Lake on the Similkameen River, a tributary of the Okanogan River, as originally supporting native sockeye salmon, although some authors (Craig and Suomela 1941) suggested that salmon could not have ascended Enloe Falls. The current Enloe Dam blocks access to all but the lower six miles of the Similkameen River. Sockeye salmon have been observed on numerous occasions since 1936 in the Similkameen River below the dam, during the Okanogan River sockeye salmon migration (Chapman 1941, Bryant and Parkhurst 1950, Chapman et al. 1995). Sockeye salmon (see Appendix Table D-2) and kokanee (see Appendix Table D-5) have been released above Enloe Dam at various times.

In reference to sockeye salmon, WDF et al. (1938) stated that "it is certain that none go into the Entiat, and none have ever been seen in the Methow." During operation of the GCFMP, sockeye salmon fry and fingerlings were released in the Methow and Entiat Rivers (Mullan, 1986, see Appendix Table D-2), and currently small numbers of sockeye salmon are consistently seen each year in these rivers (Langness 1991, Chapman et al. 1995; see section below on "Information Specific to Sockeye Salmon Populations Under Review").

Historically, it is likely that Upper Arrow, Lower Arrow, Whatshan, and Slocan Lakes in the Upper Columbia River drainage in British Columbia were utilized by sockeye salmon, as nursery lake habitat (Mullan 1986). In addition, Fulton (1970) stated that sockeye salmon probably ascended to Kinbasket, Windermere, and Columbia Lakes in the Canadian portion of the Columbia River, and Mullan (1986) suggested that the presence of kokanee indicated the past use of these lakes by sockeye salmon. WDF et al. (1938) and Chapman (1943) reported observations of sockeye salmon at Kettle Falls on the Columbia River and at Upper and Lower Arrow Lakes on the Upper Columbia River in British Columbia prior to Grand Coulee Dam construction. Comparison of sockeye salmon counts at Rock Island Dam, Tumwater Dam on the Wenatchee River, and Zosel Dam at Oroville on the Okanogan River between 1935 and 1937 indicated that more than 85% of the sockeye salmon passing Rock Island Dam were bound for spawning areas above the Grand Coulee Dam site (WDF et al. 1938). Chapman et al. (1995) indicated that this value was likely overestimated by the amount of pre-spawning mortality that might have occurred between Rock Island Dam and both Tumwater Dam on the Wenatchee River and the mill dam at Oroville on the Okanogan River. Recent escapement data on Okanogan River sockeye salmon indicate that pre-spawning mortality between Wells Dam and spawning grounds on the Okanogan River is about 30%. If similar pre-spawning mortality occurred upstream of the Grand Coulee Dam site in the years 1935 through 1937, the percentage of sockeye salmon passing Rock Island Dam that spawned upstream of Grand Coulee Dam would have been 60-64% (Chapman et al. 1995).

Historically, sockeye salmon were known to occur in Puget Sound at Baker Lake on the Baker River, a tributary of the Skagit River, and probably in Mason Lake at the base of the Kitsap Peninsula (Nehlsen et al. 1991). It is uncertain whether sockeye salmon were present historically in the Skokomish River, which drains into Hood Canal, or in the Lake Washington/Lake Sammamish Basin, which drains into Puget Sound (see following discussion). A sockeye salmon population may have spawned in Mason Lake but was reportedly eliminated in 1852 when a dam was placed on Sherwood Creek, the outlet creek of Mason Lake (Nehlsen et al. 1991). It should be noted, however, that the information cited in Nehlsen et al. (1991) concerning Mason Lake was attributed to a Twana Indian born 13 years after this stock reportedly went extinct, and as such is second-hand information at best. Baker River sockeye salmon continue to return to the lower Baker River, where they are trapped and transported above one or both dams on the Baker River to spawn in artificial beaches provided with gravel substrate and upwelling water.

Historical information indicates that sockeye salmon may once have ascended the North Fork of the Skokomish River, located at the southern end of Hood Canal (Wampler 1980, N. Lampsakis7). The original Lake Cushman had a surface area of 500 acres (Henshaw et al. 1913) and had the potential to support sockeye salmon, prior to a dam being placed at its outlet.

Numerous introductions of Baker Lake, Cultus Lake, and an unknown stock of sockeye salmon have occurred in the Lake Washington/Lake Sammamish Basin (see Appendix Table D-2), and presently the largest population of sockeye salmon in the contiguous U.S. spawns in the Cedar River, the main tributary of Lake Washington (Royal and Seymour 1940, Kolb 1971, WDF et al. 1993). Historical accounts concerning the presence and distribution of sockeye salmon within the Lake Washington/Lake Sammamish drainage are equivocal (see discussion in "Information Specific to Sockeye Salmon Populations Under Review" section below). Kokanee were present within this drainage historically and are known to be native (Crawford 1979).

Construction of Elwha Dam in 1910 on the Elwha River on Washington's Olympic Peninsula reportedly eliminated a native sockeye salmon population that spawned and reared in Lake Sutherland (Brown 1982, Wunderlich et al. 1994, Hiss and Wunderlich 1994, NPS 1995). However, Gilbert (1914), in reference to Lake Sutherland, stated that

we are acquainted with certain colonies of dwarf redfish which have been inaccessible to the sea-run form for a very long period. Such are the colonies which inhabit Lakes Crescent and Sutherland, on the northern slopes of the Olympic Mountains in Washington. The outlets of these lakes open on the southern shore of the Straits of Fuca [sic]. No run of sockeyes occurs along this shore nor into any of the streams tributary to it.

Hiss and Wunderlich (1994) recorded that 7,128,000 kokanee from outside the Elwha Basin were released in Lake Sutherland between 1933 and 1964. In 1993, an estimated 3,174 kokanee of unknown heritage spawned in Lake Sutherland (Hiss and Wunderlich 1994).

Native sockeye salmon populations exist in Ozette and Quinault Lakes on the outer coast of the Olympic Peninsula in Washington (Evermann and Goldsborough 1907, Cobb 1911, Kemmerich 1945, Atkinson et al. 1967, WDF et al. 1993). Sockeye salmon currently exist in Lake Pleasant on the Olympic Peninsula of Washington, but whether this population is native or the result of introductions is uncertain (Kemmerich 1945, WDF et al. 1993). Numerous sources indicate that Ozette and Quinault Lake sockeye salmon were of great importance for subsistence and in ceremonies of the Makah and Quinault Indian cultures, respectively (Willoughby 1889, Lestelle and Workman 1990, Storm et al. 1990). Although sockeye salmon may not currently run up the Dickey River to Dickey Lake (a tributary system of the Quillayute River on the Olympic Peninsula), a sockeye salmon population existed in the lake historically (according to E. L. Brannon8).

Age Composition

Overall age of maturity in sockeye salmon ranges from 3 to 8 years. Male sockeye salmon are capable of maturing at any of 22 different combinations of freshwater and ocean ages, while female sockeye salmon may mature at any of 14 different age compositions (Healey 1986, 1987). Kokanee generally mature after either 2, 3, or 4 years in fresh water. Formulas for designating freshwater and ocean age in sockeye salmon have been reviewed by Koo (1962) and Foerster (1968). In this report, the European method of age designation will be used, in which a decimal point separates the number of winters spent in freshwater (minus the incubation period) from the number of winters spent in saltwater (Burgner 1991). Total age is calculated by adding 1 year to the total of freshwater and saltwater age. This is the method adopted by the Fisheries Research Institute (University of Washington) and the Pacific Salmon Commission to designate age of sockeye salmon. For example, an age 1.2 fish would have spent one winter in fresh water and 2 winters in the sea for a total age of 4 years.

A combination of both environmental and genetic factors is thought to influence age composition and age at maturity. Rogers (1987) reported that among sockeye salmon from Wood River, Alaska, ocean age was most often determined by parental ocean age, whereas environmental factors most often determined freshwater age. Godfrey (1958) and Ricker (1972) thought that hereditary factors were more important than environmental factors in determining age at maturity in sockeye salmon, whereas Peterman (1985) thought that environmental conditions during early marine life were of primary importance in determining age at maturity.

While age composition and total age at maturity among sockeye salmon populations may vary year-to-year within a population, due to environmental variation and maternal influences (Bilton 1970), age composition also varies between populations, both in different river systems and within river systems (Ricker 1972, Smirnov 1975, Peterman 1985, Healey 1987, Rogers 1987, Burgner 1991, Rutherford et al. 1992, Blair et al. 1993). Further complicating analyses of age structure comparisons between populations is the fact that for some populations the percent age composition is known to change over the course of a single year's run, and selective harvests can alter the age structure of escapements (Halupka et al. 1993). Available data on age composition (see Appendix Table C-1) or total age at maturity (see Appendix Table C-2) for sockeye salmon in Washington and selected British Columbia populations reveals temporal variability within populations, as well as geographical differences among populations.

With the exceptions of Quinault Lake, Lake Washington Basin, Lake Wenatchee, and Okanogan River, multiple-year freshwater/saltwater age composition data on populations of sockeye salmon in Washington and Oregon were extremely limited (see Appendix Table C-1). Figure 3 compares the overall mean percentage of returning sockeye salmon in each age category among eleven localities in Washington. Since these data were collected over different years and have been derived from both long-term (30 years, Quinault Lake) and short-term (1 year-Ozette Lake, 2 years-Lake Pleasant) data sets, they should be interpreted in light of the above-mentioned potential for temporal variation in population age structure. Figure 4 compares age composition of adult sockeye salmon in the Quinault River tribal fishery for the years 1912 to 1924 and 1974 to 1993.

In general, there has been a shift in sockeye salmon age at return to the Lake Washington Basin over the past 25 years, with adults appearing to return at an older age than they did in the 1970s (J. Ames9). Although return-year data for sockeye salmon in Lake Washington show large fluctuations, comparison of data from 1970 to 1994 indicates that in early years less than about 5% of returning sockeye salmon were 5-year-olds; presently, an average of 19.5% (range 8-63% between 1989 and 1994) are 5-year-olds (J. Ames10) (see Appendix Table C-1). Hendry (1995) and Hendry and Quinn (1997) concluded that in 1992 and 1993, a greater proportion of sockeye salmon of age 1.1 (3-year-olds) occurred in Big Bear and Cottage Lake Creeks than in the Cedar River or Issaquah Creek, within the Lake Washington Basin. However, the sampling methods used on the Cedar River could easily have missed the jacks because of their smaller size or potentially different migration timing. The weir on the Cedar River where sockeye salmon were caught was less than 100% "jack-tight" (J. Ames11). Therefore, Hendry's (1995) data may not characterize the normal sockeye salmon jack composition of Lake Washington stocks. Ricker (1972, p. 65) reports that a "large number" of 0-age or underyearling sockeye salmon smolt were observed leaving Lake Washington in summer 1966. About 4% of the sockeye salmon smolts leaving Lake Washington in 1996 were thought to be underyearlings (see Appendix Table C-4).

In both Lake Wenatchee and on the Okanogan River, adult sockeye salmon spawners are typically 4-year-olds; however, in some years fish of age 1.1 (3-year-olds) may be more abundant than 4-year-olds in the Okanogan River population (Mullan 1986, Chapman et al. 1995) (see Appendix Tables C-1 and C-2). Three-year-old sockeye salmon in the Okanogan River population are predominately males, however limited sex-ratio data of carcasses on the spawning grounds extracted from Allen and Meekin's (1980) Table 15, indicated that 3-year-old returns in 1971, 1972, and 1973 consisted of 20%, 28%, and 22% females, respectively. The population age structure of Lake Pleasant sockeye salmon is highly unusual, with both a high number of 3-year-old spawners (36-40%) and of smolts that have spent 2 years in freshwater (40-50%) (Fig. 3 and Appendix Table C-1).

Rutherford et al. (1992) found that among coastal British Columbia sockeye salmon lakes, Mikado and Mercer Lakes had greater than 50% freshwater age-2 spawners, and this freshwater age was more common in the northern mainland and Queen Charlotte Island populations than in populations from Vancouver Island or the southern mainland. Some small coastal lakes on Vancouver Island, like Cheewhat Lake and Muriel Lake, reportedly have a high proportion of both male (jacks) and female (jills) 3-year-old returns (K. Hyatt12).

Halupka et al. (1993) identified 14 populations out of 230 in southeast Alaska with substantial proportions of zero freshwater age (= sea-type sockeye salmon) individuals (Table 2). In addition, they found 4 populations dominated by fish having spent 2 years in freshwater prior to smoltification, including those from Benzeman Lake on Baranof Island, which had the shortest length of any sockeye salmon population investigated and were listed as unique. The Hasselborg River stock on Admiralty Island was listed as unique; it is dominated by sea-type individuals and spawns in a clearwater stream, whereas other sea-type stocks listed by Halupka et al. (1993) were restricted to glacial drainages.

Fecundity and Egg Size

For a given fish size, female sockeye salmon have the highest fecundity and the smallest egg size among the Pacific salmon (Burgner 1991). Average fecundity across the range of sockeye salmon is from 2,000 to 5,200, and from about 300 to slightly less than 2,000 for kokanee (Burgner 1991, Manzer and Miki 1985). Because larger females have higher fecundity than smaller females, any comparison of fecundity between populations is confounded by differences in female age and size (Rounsefell 1957, Bagenal 1978). However, studies have shown that once the size of females has been taken into account, differences between age classes are not significant (Beacham 1982, Manzer and Miki 1985). Consequently, comparisons of fecundity should be adjusted for size (Beacham 1982), which requires measurements of both size and fecundity from the same individuals. Available information that provides these measurements for naturally spawning sockeye salmon populations was insufficient to adequately evaluate patterns of relative fecundity among sockeye salmon populations in the Pacific Northwest (see Appendix Table C-3). Data on average fecundity were available for Okanogan River, Lake Wenatchee, Cedar River, and Quinault Lake sockeye salmon stocks in Washington (see Appendix Table C-3). Chapman et al. (1995) pointed out that sockeye salmon from Okanogan River and Lake Wenatchee in the mid-Columbia River have some of the lowest fecundity estimates reported in the literature, and that this low fecundity may be related to the long migration distance inherent to these populations. Quinault Lake sockeye salmon also have a relatively low average fecundity (see Appendix Table C-3).

In other areas, researchers have reported that fecundity can be effectively used to differentiate between sockeye salmon populations in different river systems (Hartman and Conkle 1960, Foerster 1968, Ivankov and Andreyev 1969, Manzer and Miki 1985, Burgner 1991) and between spawning locations within the same river system (Aro and Broadhead 1950, Gard et al. 1987, Beacham and Murray 1993, Blair et al. 1993). Manzer and Miki (1985) found that within British Columbia, coastal sockeye salmon stocks were about 18% more fecund than interior stocks. Beacham and Murray (1993) also found that fecundity was less in upper river stocks of sockeye salmon than lower river stocks in both the Skeena and Fraser Rivers, although the relationship was statistically significant only in the former.

With the exception of the Lake Washington Basin, no information on egg size was found for sockeye salmon populations in Washington and Oregon. Mean egg weight differed among sockeye salmon populations in the Lake Washington Basin (Cedar River, Issaquah Creek, Big Bear Creek, Cottage Lake Creek, and Lake Washington beach spawners) in both 1992 and 1993, although large inter-annual variation was evident in all populations except the Cedar River, which had larger sample sizes than the other populations (Hendry 1995). In the Cedar River, Quinn et al. (1995) found body length of females to be positively correlated with gonad weight, egg number, and egg weight. In other locations the size of sockeye salmon eggs has been used to differentiate populations (Robertson 1922, Brannon 1987, Beacham and Murray 1993). Within Fraser River sockeye salmon, upper river populations had smaller diameter and lighter eggs than did lower river populations (Beacham and Murray 1993), whereas sea-type sockeye salmon in Harrison Rapids (on the Fraser River) had larger eggs than did nearby lake-type populations (Robertson 1922, Beacham and Murray 1993). Quinn et al. (1995) reported very high correlations between egg weight and size composition of incubation gravels, whereas neither body length nor snout length were well correlated with egg weight among 18 Alaskan sockeye salmon populations.

Fry Emergence Timing and Fry Migration

Sockeye salmon populations may differ in spawn timing and rates of development and fry emergence as adaptations to different thermal regimes (Brannon 1987, Beacham and Murray 1989). Beacham and Murray (1989) observed that development rate (based on hatching and emergence timing) was faster for interior-spawning sockeye salmon in the Fraser River, which experienced colder temperatures, than for lower Fraser River sockeye salmon. Sockeye salmon eggs spawned in lake outlet tributaries or on lakeshore beaches are typically exposed to warmer temperatures than eggs spawned in inlet tributaries, since lakes cool more slowly in the fall. It has been postulated that in order to synchronize fry emergence in these three spawning habitats with optimal feeding and survival conditions in common lake environments, differences in egg size, incubation period, and spawn timing have arisen (Godin 1982, Brannon 1987, Burgner 1991). This has been suggested to explain the observation that inlet spawners typically spawn earlier than lakeshore beach and lake outlet spawners, which experience higher incubation temperatures (Burgner 1991).

Sockeye salmon fry emerge in the Cedar River from January through early June, with peak emergence occurring from early March to mid-May (Stober and Hamalainen 1979, 1980; Seiler and Kishimoto 1996). Within Lake Washington/Lake Sammamish populations, Hendry (1995) observed different hatching and emergence timing among Cedar River, Big Bear Creek, Cottage Lake Creek, Issaquah Creek, and Lake Washington beach spawners.

Emerging fry possess heritable rheotactic and directional responses that allow fry from outlet tributaries to move upstream and fry from inlet tributaries to move downstream, in order to reach the nursery lake habitat (Raleigh 1967, Brannon 1972a, Burgner 1991). Fry of some populations that spawn in side tributaries connected to lake outlet streams must first travel downstream and then reverse orientation and travel upstream to the nursery lake (Egorova 1970, Brannon 1972a). Fry spawned in rivers without nursery-lake habitat rear in spring areas, side channels, and sloughs or travel to the lower estuary to rear (Birtwell et al. 1987, Eiler et al. 1992). Raleigh (1967), Brannon (1972a, b), and Miller and Brannon (1982) indicated that fry migration patterns of sockeye salmon are under genetic control.

Quinn (1980, 1981) showed that sockeye salmon fry have innate directional preferences, with Cedar River fry displaying a northerly direction preference, corresponding to the direction they take in migrating into Lake Washington. Fry emigrating from the Cedar River do so primarily at night (less than 1% were seen to migrate in daylight) (Hamalainen 1978; Stober and Hamalainen 1979, 1980) and in normal flows from 24% to 98% of marked released fry migrated to Lake Washington in one night (Seiler and Kishimoto 1996). Hendry (1995) observed no difference between migration patterns of sockeye salmon fry from the Cedar River and those from Lake Washington beach spawners (both populations were positively rheotactic), although beach spawners were predicted to lack a particular rheotactic response based on studies of beach fry from Cultus Lake (Brannon 1972b).

Smolt Size and Outmigration Timing

Summaries of available data on sockeye salmon smolt size (see Appendix Table C-4) and smolt migration timing (see Appendix Table C-5) reveal differences between populations that are related to lake productivity, thermal regime, and altitude (Burgner 1991). Sockeye salmon smolt size is influenced by length of stay in the lake habitat and lake productivity. Smolts migrating earlier in the season tend to be larger than later migrants, and both survival at sea and age and size at maturity are dependent on smolt size (Burgner 1991). Unfertilized coastal lakes of British Columbia reportedly produce smaller smolts than more productive interior lakes (Pauley et al. 1989). Freshwater age-1 smolts from Lake Washington, Ozette Lake, Baker Lake, and Lake Osoyoos tend to be relatively large, whereas the smallest lake-type sockeye salmon smolts are found in glacial Owikeno Lake in coastal British Columbia (see Appendix Table C-4). Although large smolt size in the Lake Osoyoos population apparently results in a large proportion of small 3-year-old returns, large smolt size in the Baker River, Ozette Lake, and Lake Washington populations has not resulted in large numbers of 3-year-old returning adults in those systems (see Appendix Table C-1).

Since most sockeye salmon lakes in the north are ice-covered in the winter, and sockeye salmon migration begins soon after ice break-up, there is both a south-to-north cline and an altitude-dependent factor in sockeye salmon smolt outmigration timing (Hartman et al. 1967, Burgner 1991). Besides time of ice breakup, variations in outmigration timing can be affected by water temperatures; wind direction and its effects on the lake surface; and age, size, and physiological condition of the smolts (Burgner 1991).

Because of their responses to lake productivity, smolt size and outmigration timing have been influenced by anthropogenic activities that affect lake productivity, including artificial fertilization (Hyatt and Stockner 1985) and agricultural (Allen and Meekin 1980, Chapman et al. 1995) and municipal pollution (Edmondson and Lehman 1981).

Density-dependent processes may also lead to smolts leaving overcrowded lakes at a smaller than normal size (Hartman and Burgner 1972, Goodlad et al. 1974, Hyatt and Stockner 1985). These factors thoroughly complicate the assessment of any regional pattern that may exist for either smolt size or outmigration timing, since these activities have occurred throughout the range of sockeye salmon. Sampling design may also influence reported smolt sizes and outmigration timing.

Sockeye salmon smolts migrate from most nursery lake systems at night, with greatest numbers leaving between sunset and early morning (Burgner 1991). However, this migration pattern is reversed in Lake Washington, with most sockeye salmon smolts exiting the system during daylight hours (Warner 1997).

Adult Run-Timing

In general, river entry13 and spawn timing14 of sockeye salmon show considerable spatial and temporal variability. Sockeye salmon enter Puget Sound rivers from mid-June through August, while Columbia River populations begin river entry in May, passing Bonneville Dam from very late May to late August (see Appendix Table C-6). Sockeye salmon spawn in Puget Sound from late September to late December and occasionally into January, and in the Columbia River from late September to early November (see Appendix Table C-6). Small numbers of spawners are present in the Cedar River into February (WDFW 1996). Sockeye salmon on the western Olympic Peninsula of Washington and on Vancouver Island, British Columbia begin entering rivers much earlier than the above stocks, in April and May, and in the case of Quinault Lake as early as January. Sockeye salmon begin entering Cheewhat Lake on Vancouver Island in late February or early March, and the migration continues into September with a peak in mid-June to mid-July (K. Hyatt15). Sockeye salmon on the Olympic Peninsula may spend 3 to 6 months in fresh water before spawning, and in the extreme case of Quinault Lake from 3 to 10 months. Spawning on the Olympic Peninsula begins later in the fall and extends further into the new year than in Puget Sound or on the Columbia River (see Appendix Table C-6).

Fraser River sockeye salmon exhibit remarkably consistent chronological separation of river entry and spawn timing among individual spawning populations (Killick 1955; Gilhousen 1960, 1990). Two Fraser River sockeye salmon stocks, early Stuart and early Nadina, overlap in run-timing with Puget Sound stocks from Baker River and Lake Washington. Other Fraser River stocks overlap the run-timing of Okanogan River and Lake Wenatchee sockeye salmon (see Appendix Table C-6). Adaptation of round-the-clock video technology to sockeye salmon escapement counts at Tumwater Dam on the Wenatchee River revealed that up to 5% of the sockeye salmon were passing up-river at night and would not normally have been counted (Hatch et al. 1992).

Halupka et al. (1993) found no latitudinal pattern in run-timing in an analysis of 230 sockeye salmon stocks in southeast Alaska, although interior stocks had later mean migration dates and more compact run-timing than coastal mainland or island stocks. Five sockeye salmon populations in southeast Alaska were identified as having protracted run-timing, including two that may have had separate population segments and one interior stock that was identified as having run-timing lasting more than twice as long as any other interior stock (Halupka et al. 1993).

Egorova (1970) reported that in comparison with other Kamchatka sockeye salmon stocks, both river-entry timing and spawning duration are unusually protracted in sockeye salmon from the Ozernaya River on the southwestern coast of Kamchatka. River entry begins at the end of May, peaks in August, and is not over until the beginning of October. Spawning in this system lasts from the end of July to the beginning of February.

Burgner (1991) stated that since there is an optimum time for fry emergence that coincides with maximum conditions for juvenile survival, accurate spawn timing is crucial to allow these events to coincide. Spawn timing depends to some degree on spawning gravel temperature (Brannon 1987, Burgner 1991). Sockeye salmon river- and lake-entry timing is also influenced by other factors including river flow, as in Lake Pleasant (WDF et al. 1993), and river temperature, as in the Okanogan River (Major and Mighell 1966, Allen and Meekin 1980, Mullan 1986, Swan et al. 1994). Blackbourne (1987) reported that Fraser River and Quinault River sockeye salmon run-timing is closely correlated with winter-to-spring sea-surface temperatures in the Gulf of Alaska; presumably these stocks tend not to travel so far north in a cold year and thus reverse direction earlier, approaching their natal rivers earlier than usual. All these factors make determinations and comparisons of "average" or "peak" river entry and spawn timing difficult because of the high spatial and temporal variability exhibited within basins.

Available information on dates of occurrence and spawn timing for sockeye salmon observed in rivers without accessible lake rearing habitat in Washington is summarized in Appendix Table C-7.

Spawner Size

Like the other life history traits discussed above, adult spawner size16 in naturally spawning populations shows considerable spatial and temporal variability, which may obscure regional patterns of variation. Based on fishery catch data, which tends to select for larger fish than are present in the total run, Columbia River sockeye salmon average about 1.58 kg after two winters at sea; Fraser River sockeye salmon average 2.73 kg after a similar time at sea; Bristol Bay, Alaska sockeye salmon average 2.56 kg after two years at sea; and Chignik River, Alaska sockeye salmon average 3.16 kg after three winters at sea (Burgner 1991). At the same age, males are generally larger than females after two and three winters at sea (Burgner 1991). Halupka et al. (1993) found that within populations in southeast Alaska, age-1.3 and 2.3 males were larger than females, whereas age-1.2 females were larger than males of the same age.

Adult body size may also be affected by variations in stock abundance. McKinnell (1995) found that from 1912 to the late 1960s, the mean lengths of age-1.3 sockeye salmon from northern and central British Columbia stocks (Rivers Inlet, Nass, and Skeena Rivers) were significantly smaller (P = 0.05) in years when Bristol Bay sockeye salmon abundance was high. Density-dependent effects during ocean life on sockeye salmon growth and adult body length have also been reported by Rogers (1980), Peterman (1984), and Rogers and Ruggerone (1993).

Although the size at maturity of pink, coho, and chinook salmon caught in coastal British Columbia (Ricker 1982, 1995) and chum salmon caught in Asia and Alaska (Ogura et al. 1991, Helle and Hoffman 1995) has declined in recent decades, Ricker (1995) did not detect similar declines in length or weight at maturity in mixed-stock commercial catch data for Canadian sockeye salmon. However, Cox and Hinch (1997), using stock-specific data for 10 Fraser River sockeye salmon stocks, showed that size at maturity for females in all 10 stocks and for males in 8 of the 10 stocks has generally declined over the past 4 decades. These declines are similar to those documented for other Pacific salmon stocks. In general, growth and subsequent size at maturity for Fraser River sockeye salmon were reduced when sea surface temperatures were relatively warm (Hinch et al. 1995a,b; Cox and Hinch 1997).

Various methods are used to measure body lengths of adult sockeye salmon: the Alaska Department of Fish and Game uses the mid-eye to the tail fork (MEF), the Canadian Department of Fisheries and Oceans uses the post-orbit of the eye to the hypural plate (POH), the Quinault Indian Nation and Columbia River Intertribal Fisheries Commission use the snout to the tail fork (SNF), investigators reporting on Fraser River sockeye salmon have used the tip of the snout to the base of the caudal peduncle ("standard length," similar to POH) (STD), and other investigators (Woodey 1966, Hendry 1995) use the mid-eye to the hypural plate (MEH). Because SNF includes the length of the snout, which displays great sexual dimorphism in spawning sockeye salmon, and SNF and MEF lengths include a portion of the caudal fin, which erodes on the spawning grounds, direct conversions between these different measurements should be considered as gross approximations.

Since the majority of length data available for west coast sockeye salmon exist in the form of SNF, all adult body lengths in this report were converted to SNF length using generalized equations for converting sockeye salmon lengths (Ricker 1982, Pahlke 1989, Linley 1993) (see Appendix Table C-8). Although some of these conversion equations were developed from ocean-caught sockeye salmon in British Columbia (Ricker 1982) and Southeast Alaska (Pahlke 1989), they remain the best available conversions. Linear regression equations to convert MEH, MEF, and STD to SNF were not available, nor were separate sex-specific linear regression equations available to convert MEH or MEF to POH. The equations relating MEH and MEF to POH and POH to SNF and their correlation

coefficients and sample sizes, where available, are these:

POH = 0.891(MEF) 9.064; r2 = 0.977 (n = 820) (Pahlke 1989)

POH = 0.982(MEH) + 0.606; r2 = 0.986 (n = 820) (Pahlke 1989)

POH = 0.857 (STD) + 20.29 (Linley 1993)

males: SNF = 1.2605(POH) 28.47 (Ricker 1982)

females: SNF = 1.191(POH) + 0.24 (Ricker 1982).

Among coastal British Columbia sockeye salmon populations, Rutherford et al. (1992) showed that Skidegate Lake in the Queen Charlotte Islands and Cheewhat Lake on Vancouver Island had the smallest spawners at age 1.2, while the smallest average lengths were recorded for sockeye salmon in the Queen Charlotte Islands, and the largest tended to occur in the northern mainland populations (Rutherford et al. 1992). Lowe Lake, where adults have to ascend a 3-m-high falls, had the longest fish at both age 1.2 and 1.3 (Rutherford et al. 1992) (see Appendix Table C-8).

Halupka et al. (1993) found no significant differences in length between lake-type and river/sea-type sockeye salmon. In general, Halupka et al. (1993) determined that body length was a poor trait for identifying unique stocks due to high within-population variability, and they were unable to detect geographic or temporal trends in body length or weight. However, Blair et al. (1993) showed that six populations of sockeye salmon, from various spawning sites within the Iliamna Lake system, Alaska, varied significantly in size at age.

One population of sockeye salmon in southeast Alaska, out of 230 investigated, consisted of unusually small individuals (Halupka et al. 1993). Individuals in this population, which rears in Benzeman Lake in Necker Bay on Baranof Island, weigh on average about 1 kg (Moser 1899), have an average length of 460 mm at age 1.2, and 457 mm at age 2.2. Age 2.2 sockeye salmon predominate in this population (McPherson and McGregor 1986, Halupka et al. 1993). Similarly, Britton et al. (1982) reported that a "race of small sockeye averaging about 1.5 kg accounts for 60% of the escapement" in some years in Cridge Inlet Creek on the south coast of Pitt Island in coastal British Columbia.

In 1992 and 1993, average length of same-age male and female Baker River sockeye salmon was greater than that of Lake Washington fish (Hendry 1995). However, sockeye salmon of similar age did not differ significantly in length among these populations within the Lake Washington Basin: Big Bear Creek, Cottage Lake Creek, Cedar River, Issaquah Creek, and Lake Washington beach spawners (Hendry 1995). Likewise, within a common age-class and sex, length at maturity differs little between sockeye salmon from Lake Wenatchee and Okanogan River, although these populations reportedly have the smallest body size of any major stock of sockeye salmon (Chapman et al. 1995). In fact, in some years, average length of sockeye salmon from Quinault Lake is smaller than the length of fish of the same age from the mid-Columbia populations (see Appendix Table C-8).

Adult spawner lengths of sockeye salmon in the Lake Pleasant population are very small. In 1996, age-1.2 sockeye salmon from Lake Pleasant averaged approximately 464 mm in fork length for males and 456 mm for females (see Appendix Table C-8). The small body size of Lake Pleasant sockeye salmon is comparable to that noted above for the Benzeman Lake population in Southeast Alaska.

After noting differences in sample size and number of sampling years between stocks, as well as potential for interannual variations in lengths, Shaklee et al. (1996) classified age-1.2 sockeye salmon from populations in Washington into three size groups based on SNF length: large (males >520 mm and females >510 mm), small (males <515 mm and females <500 mm, but both sexes >460 mm), and very small (males and females <460 mm). Using these criteria, Shaklee et al. (1996) classified sockeye salmon adults from Ozette Lake, Baker River, Cedar River, and Lake Washington and Lake Sammamish tributaries as "large size"; Quinault Lake, Lake Wenatchee, Okanogan River, and Lake Washington beach locations as "small size"; and Lake Pleasant as "very small size."

Appendix Table C-9 summarizes data on kokanee adult spawner lengths in selected lakes in Washington, Oregon, Idaho, and British Columbia. With the exception of kokanee in Odell Lake, Oregon; Donner Lake, California; and Issaquah Creek, Washington, average size of kokanee at maturity is typically <300 mm fork length. Odell and Donner Lake kokanee are reportedly the remnants of transplanted populations, whereas early entry Issaquah Creek kokanee are native. Comparison of body length of native kokanee in Issaquah Creek between the early 1980s and 1993 indicates a recent decline in size for both males and females (Appendix Table C-9). Kimsey (1951) reported finding spawning kokanee in Donner Lake, California with an average length of 470 mm fork length, which is greater than the average length of sockeye salmon from Lake Pleasant, Washington and from Benzeman Lake, Alaska (see above).

Ocean Life and Migration

Populations of sockeye salmon have a genetic disposition to specific migratory patterns in the ocean (Burgner 1991). Ocean distribution of sockeye salmon has been studied using tagging, morphological, parasitological, serological, and scale pattern analyses (Margolis et al. 1966, French et al. 1976, Forrester 1987). Season, temperature, salinity, age, size, and prey distribution also affect sockeye salmon movements in the open ocean. Initially, sockeye salmon juveniles travel northward from Washington and British Columbia to the Gulf of Alaska staying in a migratory band relatively close to the coast (Hartt 1980). Fraser River sockeye salmon smolts migrate north through the Gulf of Georgia, either staying close to the mainland coast or crossing the Gulf and traveling north along the Gulf Islands, where they later rejoin the north migrating mainland coast smolts (Groot and Cooke 1987). The rate of travel for northward migrating Fraser River sockeye salmon juveniles was estimated at 18.5 km/day (Hartt and Dell 1986). Once in the Gulf of Alaska, offshore movement of juveniles is conjectured to occur in late autumn or winter.

Burgner (1991) reported that Blackbourn's (1987) study implies, although indirectly, that sockeye salmon have stock-specific winter distributions in the Gulf of Alaska. French et al. (1976) provided separate models of migration for Asian stocks, western Alaskan stocks, and northeastern Pacific stocks of sockeye salmon, but not for finer-scale stock separation. The ocean distribution of Asian and North American sockeye salmon appears to overlap a broad area of the Bering Sea and North Pacific Ocean, although in general the center of North American fish abundance is east of 175oE, and the center of Asian fish abundance is west of this longitude (French et al. 1976, Burgner 1991). Although there is also considerable overlap in distribution among sockeye salmon originating all the way from the Alaska Peninsula to the Columbia River, scale pattern analyses indicate that sockeye salmon from central Alaska are distributed much further to the west than populations from southeast Alaska, British Columbia, and Washington (French et al. 1976, Burgner 1991). British Columbian and Washington populations of sockeye salmon utilize the area east and south of Kodiak Island in concert with Alaskan stocks, but tend to be distributed further to the south than the Alaskan stocks (down to 46oN) (French et al. 1976, Burgner 1991). We found no data that could be used to distinguish between the general ocean distribution of Washington, Oregon, and British Columbia sockeye salmon or of individual stocks from these regions.


The occurrence of parasites and parasite resistance in sockeye salmon phenotypes are additional traits to consider when determining the ecological/genetic importance of salmon populations under the ESA (Waples 1991a, p. 14). Extensive work has been done in Russia and Canada utilizing prevalence of parasites acquired by juveniles in freshwater to differentiate local sockeye salmon populations (Margolis 1963, Margolis et al. 1966, Konovalov 1975). As a consequence of their long freshwater life as juveniles, sockeye salmon as a species host as many as 36 freshwater parasite species, and the occurrence of parasites specific to Bristol Bay and Kamchatkan-origin sockeye salmon, respectively, have been used to differentiate the continent of origin of fish samples on the high seas (Burgner 1991).

Bower and Margolis (1984) reported that populations of sockeye salmon (from the Fraser and Skeena Rivers) exhibit a genetic difference in susceptibility to infection by the hemoflagellate Cryptobia salmositica. In British Columbia, occurrence of the myxosporeans Myxobolus arcticus and Henneguya salminicola, parasites of the brain and musculature, respectively, have been utilized to differentiate stocks of sockeye salmon (Quinn et al. 1987; Wood et al. 1987a,b, 1988, 1989; Moles et al. 1990; Rutherford et al. 1992). Among British Columbia coastal lakes, Awun, Yakoun, Kitlope, Kimsquit, Skidegate, and Long Lakes were free of M. arcticus, while other populations had levels of infection varying up to 100%. Geographical patterns in prevalence of infection were not observed (Rutherford et al. 1992).

Information relative to parasite or disease prevalence in sockeye salmon stocks from Washington and Oregon was largely unavailable. However, Bailey and Margolis (1987) compared parasite fauna of juvenile sockeye salmon from Lake Washington with sockeye salmon populations from several British Columbia lakes. Populations from Lakes Washington, Nimpkish (on Vancouver Island), and Cultus (lower Fraser River) clustered together based on their particular parasite faunas. Based on this study, Bailey and Margolis (1987) stated that although "geography influences the characteristics of the parasite fauna . . . the trophic status of the lake and many biotic variables clearly have strong influences on the parasite faunas studied."

On to Next Section

Table of Contents