Pacific Salmon Life Histories - jeffersonco-treis.infojeffersonco-treis.info/s/3.04 Water... ·...

O •

Pacific Salmon Life Histories

EDITED BY C. GROOT AND L. MARGOLISDepartment of Fisheries and Oceans

Biological Sciences BranchPacific Biological Station, Nanaimo

British Columbia, Canada

UBCPressVancouver

Published in co-operation with the Government ofCanada, Department of Fisheries and Oceans

© UBC Press 1991Reprinted 1998

All rights reservedPrinted in Canada on acid-free paper °°

ISBN 0-7748-0359-2

Canadian Cataloguing in Publication DataMain entry under title:

Pacific salmon life histories

Includes bibliographical references and index.ISBN 0-7748-0359-2

i. Pacific salmon, i. Groot, C. (Cornells),1928- ii. Margolis, L., 1927-

QL638.S2P33 1991 597'-55 091-091310-2

UBC PressUniversity of British Columbia

6344 Memorial RdVancouver, BC v6r 122

(604) 822-5959Fax: 1-800-668-0821

CONTENTS

Colour Plates

PrefaceC. GROOT AND L. MARGOLIS

Contributors

Life History of Sockeye Salmon (Oncorhynchus nerka)ROBERT L. BURGNER

Life History of Pink Salmon (Oncorhynchus gorbuscha)WILLIAM R. HEARD

Life History of Chum Salmon (Oncorhynchus keta)E.O.SALO

Life History of Chinook Salmon (Oncorhynchus tshcrwytscha)M.C. HEALEY

Life History of Coho Salmon (Oncorhynchus kisutch)F.K. SANDERCOCK

Life Histories of Masu and Amago Salmon (Oncorhynchus masou and Oncorhynchus rhodurus)FUMIHIKO KATO

Geographical Index

Subject Index

VII

ix

xiii

119

231

3"

395

447

523

543

COLOUR PLATES

Following page 240

1 Pre-spawning sockeye salmon holding along the banks of the Adams River, British Columbia2 Developing pink salmon eggs and hatched alevins3 Mature male and female pink salmon, British Columbia

4 Sockeye salmon life history stages5 Ocean phase of sockeye salmon6 Spawning sockeye salmon, Iliamna Lake, Alaska

7 Pink salmon life history stages8 Pink salmon fry shortly after emergence from the gravel9 Albino pink salmon fry from Sashin Creek, southeastern Alaska

10 Adult pink-chum salmon hybrid

11 Chum salmon life history stages12 Chum salmon fry13 Adult chum salmon during spawning act, Big Beef Creek, Washington

14 Chinook salmon life history stages15 Maturing chinook salmon holding in a pool below Stamp River falls, British Columbia16 Maturing chinook salmon jumping three-metre falls during upstream migration, Stamp River,

British Columbia

17 Coho salmon life history stages18 Coho fry in full colour, Rosewall Creek, British Columbia19 Early coho fry in gravel bed prior to emerging, Rosewall Creek, British Columbia

20 Masu salmon life history stages21 Amago salmon life history stages

CONTENTS

INTRODUCTION 313RELATIVE ABUNDANCE 314SPAWNING POPULATIONS 315

Distribution and Abundance of Spawning Stocks 315Timing of Spawning Runs 318Timing of Spawning 320Redd Characteristics 321Length of Residence on the Redd 324

FECUNDITY 324SPAWNING, INCUBATION, AND

SURVIVAL 326Egg Deposition 326Survival during Incubation 327

Emergence 329Summary of Egg-to-Fry Survival 329

FRESHWATER RESIDENCE ANDDOWNSTREAM MIGRATION 331Fry Migrations 331Habitat Utilization in Fresh Water 334Fingerling Migrants 336Growth of Fingerlings 338Yearling Smolts 338Growth of Yearling Smolts 339

Summary of Freshwater Residence andDownstream Migration

MORTALITY AND ITS CAUSES DURINGFRESHWATER RESIDENCE

FOOD HABITS IN FRESH WATERUTILIZATION OF ESTUARINE HABITATS

Fry MigrantsFingerling and Yearling MigrantsFood Habits in EstuariesGrowth in Estuaries

OCEAN LIFEFish in their First Ocean Year

340

340341342342344346347350351

Fish in their Second and Subsequent Ocean Years 353Ocean Distribution in Relation to Temperature 356Vertical Distribution of Chinook 356Ocean Distribution in Relation to Area of Origin 357Summary of Ocean Distribution of Chinook 365Ocean Food Habits 367Ocean Growth 369

MATURATION AND RETURN TO SPAWN 375MORTALITY IN THE OCEAN 378HOMING AND STRAYING 379CONCLUDING REMARKS 381REFERENCES 383

312

LIFE HISTORY OF CHINOOK SALMON(Oncorhynchus tshawytscha)

M.C. Healey*

INTRODUCTION

THE GENUS Oncorhynchus dates at leastfrom the Pliocene (Smith 1975) and prob-ably originated from a stream- or lake-dwelling Salmo-like fish (Neave 1958).When the modern species evolved is un-

certain, but the chinook salmon (O. tshawytscha),and the other Pacific salmon species, may haveevolved as recently as 500,000 to 1,000,000 years ago(Neave 1958). A time span of one million years isextremely short for the differentiation of seven spe-cies. Nevertheless, Neave (1958) argued that therewere good geological reasons for believing thatsuch evolution had occurred. Most important, therestriction of Oncorhynchus species to the North Pa-cific suggested that their evolution postdated thefaunistic connection between Atlantic and Pacificthat existed during the late Pliocene. Thus, it seemspossible that the species had undergone their com-plete elaboration during the Pleistocene. Recentinvestigation of mitochondrial DNA, however, sug-gests that the species may be two to three millionyears old (Thomas et al. 1986), which would bemore in keeping with their elaboration during thePliocene.

Morphologically, the chinook is distinguishedfrom other Oncorhynchus species by its large size(adults may reach a weight of 45 kg), and by havingsmall black spots on both lobes of the caudal fin,black pigment along the base of the teeth (Plate 14),

'Department of Fisheries and Oceans, Biological SciencesBranch, Pacific Biological Station, Nanaimo, British ColumbiaV9R 5K6. Present address: Westwater Research Centre, Univer-sity of British Columbia, Vancouver, BC V6T 1Z2

and a large number of pyloric caeca (>100)(McPhail and Lindsey 1970; Hart 1973). Chinookalso differ from the other species by their variableflesh colour, from white through various shades ofpink to red.

Chinook fry and parr are distinguished by hav-ing large parr marks extending well below the lat-eral line (Plate 14). The adipose fin is normallyunpigmented in the centre, but edged with black.The anal fin is usually only slightly falcate, and theleading rays do not reach past the posterior inser-tion of the fin when folded against the body. Theanal fin has a white leading edge, but this is not setoff by a dark pigment line as it is in coho salmon.Juvenile characteristics are highly variable, how-ever, so that proper identification often requiresmeristic and pyloric caeca counts.

Within the species group, the chinook is mostclosely related to the coho (O. kisutch), with whichit forms one subgrouping. Sockeye (O. nerka), chum(O. keta), and pink (O. gorbuscha) form a second sub-grouping, and masu (O. masou) and amago (O. rho-durus), supposedly the most primitive, form a thirdsubgrouping. According to Tsuyuki et al. (1965) andTsuyuki and Roberts (1966), the probable evolu-tionary order of the species is: masou, kisutch, tsha-wytscha, keta, nerka, gorbuscha, although there issome question about the ordering of keta and nerka.Tsuyuki et al. (1965) and Tsuyuki and Roberts(1966) did not recognize rhodurus as a separate spe-cies, but see Kato, this volume. Oncorhynchus rhodu-rus is more primitive than O. masou.

The chinook, like all Oncorhynchus species, isanadromous and semelparous (i.e., dies after

313


it:

spawning once). Within this general life historystrategy, however, chinook display a broad array oftactics that includes variation in age at seaward mi-gration, variation in length of freshwater, estuarine,and oceanic residence, variation in ocean distribu-tion and ocean migratory patterns, and variation inage and season of spawning migration.

An important objective of this chapter is to de-velop a conceptual model of the life history of chi-nook that can encompass this degree of variation. Ishall argue that there must be two fundamentalcomponents of this model. The first component isracial (Healey 1983). (Here, "race" is used in thesense that Merrell (1981) used it, that is, to identifysubdivisions of a population that are geographi-cally separated to some degree and between whichgene flow is reduced.) A large part of the variationin chinook life history apparently derives from thefact that the species occurs in two behaviouralforms. One form, which has been designated"stream-type" (Gilbert 1913), is typical of Asianpopulations and of northern populations and head-water tributaries of southern populations in NorthAmerica. Stream-type chinook spend one or moreyears as fry or parr in fresh water before migratingto sea, perform extensive offshore oceanic migra-tions, and return to their natal river in the spring orsummer, several months prior to spawning (Figure1). Occasionally, males of this form mature preco-ciously without ever going to sea. The second form,which has been designated "ocean-type" ("sea-type" in Gilbert 1913), is typical of populations onthe North American coast south of 56°N. Ocean-type chinook migrate to sea during their first yearof life, normally within three months after emer-gence from the spawning gravel, spend most oftheir ocean life in coastal waters, and return to theirnatal river in the fall, a few days or weeks beforespawning (Figure 1).

The second component of the life history model

is tactical and encompasses variation within eachrace (Figure 1). This variation represents adaptationto uncertainties in juvenile survival and productiv-ity within particular freshwater and estuarinenursery habitats. Briefly, chinook appear to haveevolved a variety of juvenile and adult behaviourpatterns in order to spread the risk of mortalityacross years and across habitats (e.g., Stearns 1976;Real 1980). By so doing, they avoid the potentialdisaster associated with high mortality in a particu-lar year or habitat.

Species

Oncorhynchus tshawytscha(anadromous; semelparous)

Race: Stream-type(long freshwater residenceas juvenile; adult runs inspring and summer; adultsenter fresh water monthsbefore spawning

Race: Ocean-type(short freshwater residenceas juvenile; adult runs insummer and autumn; adultsspawn soon after enteringfresh water)

Variation in age of seawardmigration (years)

precocity

Variation in age ofmaturity (males andfemales)

Variation in time of seawardmigration (weeks)

Variation in length of estuarineresidence (weeks)

Variation in time of returnto natal stream (Februarythrough July)

Variation in age of maturity(males and females)

Variation in time of returnto natal stream (Julythrough December)

Variation in fecundity(high fecundity) Variation in fecundity

(low fecundity)

FIGURE 1Life history structure of chinook salmon showing the

division of the species into two races (ocean- andstream-type) and the range of tactical variation within

each race

RELATIVE ABUNDANCE

The chinook is a valuable commercial species.Fisheries for chinook are conducted by Japan on

the high seas west of 175°E in the North PacificOcean (west of 175°W during 1955-77) and west of

314

Life History of Chinook Salmon

175°W in the Bering Sea by means of gillnets. TheBering Sea fishery is to be phased out by 1994.Nationals of Canada, the United States, and theUSSR fish for chinook in their coastal waters andrivers. In the eastern USSR, fishing is by trap netand gillnet. In the coastal waters of the UnitedStates and Canada, fishing is by gillnet, purseseine, and troll (hook and line). Annual recordedcommercial landings averaged about 23,800 t be-tween 1970 and 1979 (INPFC 1972-82). The largestcatches are along the British Columbia coast, al-

though substantial catches are also made in south-eastern Alaska, in the Japanese mothership fishery,and along the coasts of Washington, Oregon, andCalifornia (Figure 2). In North America the chinookis also prized as a sport fish, and approximatelyone million are taken each year in sport fisheries(INPFC 1972-82; Department of Fisheries andOceans, Vancouver, British Columbia, unpub-lished data). Chinook are, however, the third leastabundant of the Pacific salmon, with only O. rhodu-rus and O. masou being less abundant.

70°N

- 60°N

40°

- 50"N

- 40°N

120°E 140°E 160°E 180° 160°W 140°W 120°W

FIGURE 2Map of the North Pacific Ocean with histograms showing the catch (hundreds of thousands of fish) of chinook in

major coastal fisheries from 1962-70. (Adapted from Major et al. 1978)

SPAWNING POPULATIONS

Distribution and Abundance of Spawning Stocks

Spawning stocks of chinook are known to be dis-tributed from northern Hokkaido to the AnadyrRiver on the Asian coast and from central Califor-

nia to Kotzebue Sound, Alaska, on the NorthAmerican coast (Figure 3) (McPhail and Lindsey1970; Major et al. 1978). Unconfirmed reports, how-ever, suggest that chinook salmon may be distrib-uted even further north and east on the Alaskan

315


coast, and Hart (1973) cited an unpublished reportof thirteen specimens from the Coppermine River(67°50'N, 115°00'W) in the Canadian Arctic.McLeod and O'Neil (1983) reported recovering asingle specimen from the Liard River in the upperMackenzie River drainage.

There are probably well in excess of a thousandspawning populations of chinook salmon on theNorth American coast (Atkinson et al. 1967; Aroand Shepard 1967) and an uncertain, but probablymuch lower, number on the Asian coast. Spawning

occurs from near tidewater to over 3,200 km up-stream in the headwaters of the Yukon River (Ma-jor et al. 1978). Individual spawning populations ofchinook are relatively small, not exceeding a fewtens of thousands. In British Columbia, where rec-ords have been kept on over three hundred chi-nook spawning populations for many decades, 80%of the populations have averaged fewer than onethousand spawners (Healey 1982a). Presumably,individual spawning populations are similarlysmall throughout the Chinook's range.

70° •

QueenChartotte\Islands

VancouverIsland

Columbia R.

Sacramento R

70°N

60°N

50°N

40°N

120°E 140°E 160°E 180° 160°W 140°W 120°W

FIGURE 3Map of the North Pacific Ocean and Bering Sea, showing the distribution of chinook spawning populations

(stippled) and some of the landmarks referred to in the text. The distribution of chinook spawning populationsnorth and east of Kotzebue Sound on the North American coast is unconfirmed (shown as question marks),

except for a positive identification in the Mackenzie drainage.

The largest rivers tend to support the largestaggregate runs of chinook and also tend to havethe largest individual spawning populations (Fig-ure 4). This is not surprising, because larger riversare likely to have more suitable spawning andrearing habitats than smaller rivers. What is some-what surprising, however, is that major rivers at

the northern and southern limits of the Chinook'srange support populations as large or larger thanthose in major rivers near the middle of the range.The Sacramento-San Joaquin River system, at thesouthern limit of the range of chinook, for example,had chinook runs of a million fish or more until theearly part of this century (Clark 1929). These rivers

316


100000-1

5000

2000cI 1000aOaat

I<a

500-

200-

100

50-

20

t •

50 100 200 500 1000 2000 5000 10000 20000

River discharge (c.f.s.)

FIGURE 4Relationship between average spawning populationsize (1952-76) and average river discharge for British

Columbia populations of chinook salmon (both axes areon a logarithmic scale). These data are for individual

spawning populations within each river system.

still have runs of several hundred thousand, eventhough habitat loss and water extraction havebeen so severe that the San Joaquin River wasvirtually unable to support chinook spawning,with escapement averaging less than 4,000 in themid-1970s (Kjelson et al. 1982). Since 1980, how-ever, escapements to the San Joaquin have recov-ered somewhat and escapement was 70,000 in 1985(M. Kjelson, u.s. Fish and Wildlife Service, Stock-ton, California, pers. comm.). Run estimates for theYukon and Nushagak rivers, near the northernlimit of the range of chinook, are very uncertain,but catches of chinook in western Alaska and es-capement estimates for these rivers indicate thatruns to both rivers are probably on the order of400,000 to 600,000 fish (Knudsen et al. 1983). Bycomparison, the Columbia River historically pro-duced only about twice as many fish as the Sacra-mento-San Joaquin River system, and the FraserRiver produced only about 200,000 chinook (Rich1942; M.C. Healey unpublished data). Both theColumbia and Fraser rivers are near the centre ofthe Chinook's range on the North American coast.

Chinook have been transplanted to a variety of

locations outside their normal range. Transplantsto the east coast of North America have been con-ducted for almost a century, initially with the hopeof supplementing declining Atlantic salmon (Salmosalar) runs, but also as a means to "control" land-locked smelt (Osmeridae) and alewife (Alosa pseudo-harengus) populations (Hoover 1936; Ricker andLoftus 1968). Most transplanted populations weremaintained only by artificial propagation/if theywere maintained at all. Recently, however, natu-rally spawning populations have become estab-lished in the Laurentian Great Lakes (Carl 1982).Both artificially and naturally produced chinookare now a valuable component of the sport fisher-ies in Lake Michigan and its tributary streams.

Chinook from California were successfullytransplanted to New Zealand around the turn ofthe century. The first transplants to New Zealandwere in the late 1800s, and intensive stocking be-gan in 1901. By 1907, adults were returning to NewZealand hatcheries, and by 1910 it was possible forthese hatcheries to export spawn to other parts ofNew Zealand and Tasmania. Chinook are nowwidespread in rivers along the east coast of thesouth island and occur in some rivers on the westand north coasts. Whether this has occurred as aresult of deliberate planting or natural straying isnot clear (Waugh 1980).

Although the original intent of the introductionof chinook to New Zealand was to create commer-cially exploitable runs of fish, established runshave provided only a low sustainable yield, andlittle commercial harvesting of salmon has oc-curred. Chinook are, however, an important sportfish. Limited suitable spawning and rearing areasin New Zealand streams, together with impound-ments, water extraction for irrigation, and pollu-tion, have served to limit smolt production (Waugh1980). Poor smolt production is presumably animportant reason for the relatively low productiv-ity of chinook in New Zealand.

Chinook have also been transplanted to south-ern Chile. Apparently, landlocked populationshave been established there, and the hope is thatmarine anadromous populations will also be suc-cessful in Chile. Some returns of anadromous chi-nook have been reported in sea ranchingoperations (Lindbergh 1982).

In this chapter, I shall concentrate on chinookwithin their natural range. This is not to discount

317


the possible impact of successful transplants. Theintroduction of chinook and other Pacific salmon tothe Great Lakes, for example, almost certainlymeans that they will ultimately invade AtlanticOcean drainages, with unknown consequences forlocal fish fauna. It is the intent here, however, toemphasize what is known about the biology of theanimal in its natural habitat.

Within the natural range of chinook salmon,stream- and ocean-type spawning populations aregeographically separated to a considerable degree(Healey 1983). All Asian stocks are apparentlystream-type (e.g., Knudsen et al. 1983). On theNorth American coast, chinook spawning popula-tions are wholly or predominantly stream-typethroughout Alaska. At the Alaska-British Columbiaborder, however, there is a rather abrupt shift incomposition. From the Nass and Skeena rivers(56°N) (Figure 3) southward, ocean-type chinookdominate all runs except, perhaps, the YakounRiver on the Queen Charlotte Islands (Table 1).Stream-type chinook make an important contribu-tion to runs to larger rivers south of 56°N(14%-48%) but are relatively scarce in smaller riv-ers (0%-12% of runs). Wherever they are sympatricwith ocean-type fish, stream-type fish tend to befound in headwater spawning areas and ocean-type fish in downstream spawning areas (Rich1925; Hallock et al. 1957; Healey and Jordan 1982).The geographic separation is not complete, how-ever, as the behavioural types are sympatric onmany spawning riffles.

Timing of Spawning Runs

Chinook salmon may return to their natal rivermouth during almost any month of the year(Snyder 1931; Rich 1942; Hallock et al. 1957). Thereare, however, typically one to three peaks of migra-tory activity. The timing and the number of migra-tory peaks varies among river systems (Figure 5).For northern river systems (Kamchatka River:Vronskiy 1972; Yukon River: Brady 1983; Cook Inlettributaries: Yancey and Thorsteinson 1963; Nassand Skeena rivers: Department of Fisheries andOceans, Vancouver, British Columbia, unpubl.data), a single peak of migratory activity duringJune appears typical, although the run may extendfrom April to August (Figure 5). Particular spawn-ing populations may return later in the season,

TABLE 1Occurrence of stream- and ocean-type chinook inspawning runs to rivers along the west coast of

North America

River system

AlaskaYukonCook Inlet riversTakuStikine

British ColumbiaNassSkeenaKitimatYakounBella CoolaDocee (Rivers Inlet)Quinsam(Campbell)Big QualicumFraserNanaimoNitinatChemainusCowichan

Washington/OregonColumbiaSixes

CaliforniaKlamathSacramento

Approximate N.latitude

62°30'61°30'58°30'56°40'

55°20'54°20'54°00'53°30'52°25'51°40'

50°00'49°25'49°20'49°10'48°50'48°50'48°50'

46°10'42°50"

41°30'38°00'

% of spawningruns

Stream

10097-99100100

42481257143

10

34510

10

2212

1410-18

Ocean

01-3

00

585288438697

99100669599

10090

7888

8682-90

Source: From Healey (1983) with additional data from Clark(1929)Note: The rivers are ordered in descending latitude.

however, and this may result in two peaks of mi-gratory activity (e.g., Kenai Peninsula: Yancey andThorsteinson 1963).

Further south, runs tend to occur progressivelylater. In the Bella Coola River, an early run in lateMay and June is followed by a second, but rela-tively smaller, run in August (Figure 5). In theFraser River there are two runs of almost the samesize: an early run peaking in July, and a late runpeaking in September/October. The Fraser Riverhas a third, smaller, run in August that corre-sponds in timing to the late run into the BellaCoola River (Figure 5) (Ball and Godfrey 1968a,1968b; Fraser et al. 1982). A late August run domi-

318


• Yukon R.A Cook Inlet tributarieso Nass R.

Bella Coola R.

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 ) 2 3 4

• Fraser R. •Columbia R.

o Klamath R.

1 2 3 4 1 2 3 4 1 2

MA May Jun Jul Aug Sep Oct N D

FIGURE 5The timing of spawning runs to rivers throughout the

North American range of chinook. The estimated abun-dance of fish during each quarter month is shown rela-

tive to the quarter monthly period with the greatestabundance offish. Data sources are: Brady (1983) for

the Yukon; Yancey and Thorsteinson (1963) for Cook In-let tributaries; Canada Dept. of Fisheries and Oceans

(unpublished data) for the Nass and Bella Coola rivers;Fraser et al. (1982) for the Fraser; Rich (1942) for the

Columbia; and Snyder (1931) for the Klamath

nates chinook returns to both the Columbia andKlamath rivers. The Columbia River also hasspring and summer runs, but these are numeri-cally small relative to the late August run (Figure5) (Rich 1942; Silliman 1950). The Klamath Riverhas a spring run, but it, too, is relatively small(Figure 5) (Snyder 1931). Historically, the KlamathRiver spring run and the Columbia River summerrun are believed to have been much larger, butthey have since been decimated by habitat lossand overfishing.

The early run in the Fraser River may be anala-gous to the summer run in the Columbia River andto the spring runs in more northern rivers. The lateFraser River run appears unique in its timing,

unless it is an analogue of the winter run into theSacramento River which reaches Redding from lateNovember through February, 350 km upstream(Slater 1963; Hallock and Fry 1967). The Sacra-mento River also has spring and fall runs of chi-nook, and a unique winter run. The winter andspring runs tend to overlap considerably in timingso that their separation is not precise during therun, but the two groups do separate on the basis ofspawning area and time of spawning. Winter-runfish spawn mainly in May and June in the uppermain stem of the Sacramento River. Spring-runfish, however, delay spawning until late August orSeptember, and their spawning areas are in theupper reaches of the main stem of the SacramentoRiver and of the principal tributaries. The majorityof chinook in the Sacramento River are fall-run fishthat enter the river in September and October.These fish spawn shortly after entering the river inthe middle and lower reaches (Hallock et al. 1957;Slater 1963; Hallock and Fry 1967; Kjelson et al.1981). Thus, Chinook in the Sacramento River ap-pear to exhibit typical stream-type (spring-run)and ocean-type (fall-run) behaviour. The winter-run fish are somewhat anomalous in that they havecharacteristics of both stream- and ocean-typeraces. They enter the river green and migrate farupstream. Spawning is delayed for some time afterriver entry. Young winter-run chinook, however,migrate to sea in November and December, afteronly four to seven months of river life. Withoutfurther information or details of the life history ofthese rare and interesting fish it is difficult todecide whether or not they fit the classification ofstream- and ocean-type races.

Despite the wide variation in run timing withinmost rivers, spawning times tend to be similaramong runs. Early-run fish normally delay spawn-ing and spawn in the fall, about the same time aslate-run fish. The winter run in the SacramentoRiver is an exception to this rule.

Detailed information on the racial compositionof spawning runs is available for the Fraser andColumbia rivers (Rich 1925; Ball and Godfrey1968a, 1968b). These data clearly demonstrate thealternation in timing of stream- and ocean-typeraces entering the two rivers (Figure 6). Stream-type chinook enter principally during the springand early summer, and ocean-type chinook enterduring the summer and fall. Given the fact that

319

1 Pacific Salmon Life Histories

100-1

S.

50-

Columbia River

Stream-type

Ocean-type -•x/

Fraser River

100

50-

Mar Apr May Jun Jul Aug Sep

FIGURE 6Seasonal changes in the percentage contribution of

stream- and ocean-type chinook to the total run in theColumbia and Fraser rivers. (Data from Ball and

Godfrey 1968a, 1968b, 1969,1970; Rich 1942)

stream-type fish migrate to headwater tributariesof these two systems, their early entry to freshwater permits them to take advantage of peaksummer flows to reach their spawning areas.Ocean-type fish, with shorter upriver migrationson average, can delay entry until after peak flowsand thus take advantage of a slightly longer oceanfeeding period. Stream-type fish, therefore, appearto suffer a double disadvantage. Not only do theylose feeding time in the sea, but they must alsomaintain their ion balance, without feeding, in theosmotically rigorous freshwater environment forseveral months before spawning. The adaptationsrequired to achieve this suggest that there must bemore than casual genetic separation betweenstream- and ocean-type chinook.

Timing of Spawning

The time at which chinook actually spawn is quitevariable, ranging from May/June for some more

northern populations and for the winter-run fishin the Sacramento River (Department of Fisheriesand Oceans, Vancouver, British Columbia, unpub-lished data; Hallock et al. 1957; Slater 1963) toDecember/January for fall-run chinook in the Sixesand Elk rivers, Oregon (Reimers 1971; Burck andReimers 1978), and in the Sacramento River (Hal-lock and Fry 1967). The trend throughout the rangeof chinook is to earlier spawning as one movesnorth (Figure 7), with northern populations tend-ing to spawn from July to September and southernpopulations from November to January. Withinriver systems, however, individual populationsmay spawn at widely different times. In the Sacra-mento River, for example, spring-run fish spawn inAugust to September, fall-run fish in October toDecember, and winter-run fish not until May andJune. Thus, there are (or were before the winterand spring runs were decimated) chinook spawn-ing virtually every month of the year in the Sacra-mento River. In the Fraser River (Department ofFisheries and Oceans, Vancouver, British Colum-bia, unpublished data), median spawning dates forindividual populations range from June to No-

8

300-

250-

200-

150

W £

35 40 45 50 55

Degrees N. latitude

60 65

FIGURE 7The relationship between the median spawning date

and the latitude of the spawning population forstream- and ocean-type chinook. Symbols are:

o = stream-type; * = ocean-type; • = mixed-type. Thecorrelation for all data is significant (r - -0.755, p <.01).

The races do not differ significantly (p >,05). Theregression reported on the figure is the regression

of Julian day on latitude.

320

Life'History of Chinook Salmon

vember, a span of six months. Spawning appearsto be less protracted in northern rivers. In theSkeena River the range of median spawning datesamong populations was 15 August to 25 Sep-tember, and in the Kamchatka River the rangeamong populations was 15 July to 15 August (De-partment of Fisheries and Oceans, Vancouver, Brit-ish Columbia, unpublished data; Vronskiy 1972).

There is some suggestion that stream-type fishspawn earlier than ocean-type fish, at least in thecentral and southern parts of the Chinook's rangewhere the two types are sympatric. Burner (1951)noted that spring-run (stream-type) ColumbiaRiver chinook spawned in late August whereassummer- and fall-run fish (mainly ocean-type)spawned in late September. Hallock et al. (1957)also noted earlier spawning of spring-run fish,which presumably were mainly stream-type, rela-tive to fall-run fish in the Sacramento River.Known populations of stream-type chinook in theFraser River, however, do not obviously spawnearlier than ocean-type chinook(Department ofFisheries and Oceans, Vancouver, British Colum-bia, unpublished data). Northern populations,which are all stream-type, do spawn early, but thisis not clearly related to race. If earlier spawningwas mainly a function of race, then one wouldexpect a discontinuity in the median date ofspawning at about 56°N latitude, the latitude atwhich the composition of spawning populationsshifts from predominantly ocean-type to predomi-nantly stream-type. The trend to earlier spawningas one moves north is, however, continuousthroughout the range of chinook, with no apparentdiscontinuity where population structure shifts tostream-type (Figure 7).

Redd Characteristics

The spawning beds chosen by chinook vary con-siderably in physical characteristics. Chinook willspawn in water depths from a few centimetres(Burner 1951; Vronskiy 1972) to several metres(Chapman 1943; Chapman et al. 1986). They willspawn in small tributaries two to three metreswide (Vronskiy 1972) and in the main stem of largerivers like the Columbia and Sacramento (Chap-man 1943; Hallock et al. 1957).

Several authors have reported in some detail onthe characteristics of chinook redds and spawning

beds (Chapman 1943; Burner 1951; Briggs 1953;Vronskiy 1972; Neilson and Banford 1983). In addi-tion, the expanding literature on instream flowrequirements provides depth and velocity criteriafor chinook spawning (Ceilings et al. 1972; Smith1973; Bovee 1978). Comparison among observa-tions by different authors is complicated by thefact that methods of measurement were different,and in some reports the methods were not given insufficient detail to permit an assessment of com-parability. The overriding impression is that, al-though there is good agreement among meanvalues for water depth and velocity in spawningbeds, the range in depths and velocities that chi-nook find acceptable is very broad (Table 2). Thereis little agreement among observers about eitherthe maximum or the minimum values for depthand velocity. For example, Burner (1951) observedchinook spawning in as little as 5 cm of water,whereas Collings et al. (1972) suggested that theminimum water depth for spawning is 30 cm. Max-imum depths of spawning range over an order ofmagnitude, from 41 to >700 cm. Velocity minimarange from 10 to 52 cm/s and maxima from 64.4 to150.0 cm/s. There is apparently no agreement as towhether depth and velocity characteristics used innest site selection differ between stream- andocean-type chinook. Some authors assign greaterdepth and velocity preferences to stream-type andothers to ocean-type chinook (Table 2).

In preparation for spawning, the female chinookdigs a shallow depression in the gravel of thestream bottom by performing vigorous swimmingmovements on her side near the bottom. Graveland sand thrown out of the depression accumu-late in a mound, or tailspill, at the downstreammargin of the depression. During the act of spawn-ing the female deposits a group or "pocket" of eggsin the depression and then covers them withgravel. Over the course of one to several days, thefemale deposits four or five such egg pockets in aline running upstream, enlarging the spawningexcavation in an upstream direction as she does so.The total area of excavation, including the tailspill,is here termed a "redd."

Four papers reported the size of the area of theredd excavation (Table 2). Burner (1951) observed arelatively narrow range of redd sizes in tributariesof the Columbia River, with stream-type chinookhaving the smaller redds. The redds of the stream-

321


TABLE 2Summary of published information on water depth and velocity in chinook spawning beds,

and area of the redd excavation

Water depth (cm) Water velocity (cm/s) Redd area (m2)

Source Type* Range Mean Range Mean Range Mean

Chapman (1943)

Burner (1951)

Briggs (1953)

S

0

30-460

5-122

28-41

31

32 30-76

2.4-4.0

3.9-6.5

Vronskiy (1972)Collings et al. (1972)

Smith (1973)

Bovee (1978)§

Nelson & Banford (1983)Chapman et al. (1986)

S0S0S0

13-72045-52

30-45

10-7010-120

to 700

56t

3138.93030

30-15052-68

30-68

21.7-64.418.6-80.5

10-10025-11515-10037-189

6154+

4349.7405056

>100

4.0-15.0$

0.5-27.52.1-44.8

9.517.0

Notes: 'Separate values are reported for stream- and ocean-type chinook when available. S - stream-type (spring-run chinook,0 - ocean-type (fall-run) chinooktGeometric mean of the ranges given by Vronskiy (1972) for different tributaries^Maximum redd areas. These were calculated as the product of maximum and minimum length and breadth measurements given inVronskiy (1972).§Values taken from probability-of-use curves in Bovee's (1978) report

type fish also appeared to be in areas of coarsergravel and were often characterized by having afew large cobbles in the bottom of the excavation.Vronskiy (1972) gave only length and width mea-surements for stream-type chinook spawningmounds in the Kamchatka River. Maximum reddareas, inferred from the product of minimum andmaximum measurements given by him, rangedfrom 4 to 15 m2. Vronskiy (1972) also commentedon the appearance of large cobbles in the bottom ofchinook redds in the Kamchatka River but notedthat in some tributaries chinook spawned in veryfine gravel. Neilson and Banford (1983) reported agreat range of sizes and large average size for reddsof stream-type chinook in the Nechako River (Brit-ish Columbia). Chapman et al. (1986) reportedredd areas for the Hanford reach of the ColumbiaRiver that ranged from 2.1 to 44.8 m2 and averaged17.0 m2. Neilson and Banford (1983) apparentlyestimated redd area as the product of maximumlength and width of the redd, as I did from Vrons-kiy's data, whereas Burner (1951) and Chapman et

al. (1986) took account of the oval shape of theredd. Thus, the measurements are not comparable,and Neilson and Banford's (1983) estimates of reddarea, as well as those that I made from Vronskiy'sdata, will be up to twice as large as those made byBurner (1951) or Chapman et al. (1986).

Both Vronskiy (1972) and Neilson and Banford(1983) observed that the depth of the redd excava-tion was negatively correlated with water velocityin the spawning area. According to Vronskiy(1972), the higher mound in the tailspill of reddsdug by chinook in low velocity water serves toimprove subgravel irrigation of the eggs. Low ve-locity areas were also likely to be characterized byfine gravel that, presumably, could be dug intomore easily by the fish, so that nests were deeper.

Briggs (1953) reported that chinook buried theireggs 20-36 cm deep (average 28 cm) in the gravel oftwo small streams in California. Vronskiy (1972)observed eggs buried from 10 to 80 cm deep in thegravel in the Kamchatka River, although he foundfew eggs below 50 cm. According to Vronskiy

322


(1972), the depth to which the eggs were buriedwas at least partly dependent on water flow, withthe eggs being buried more deeply where flow waslow. Chapman et al. (1986) found that the depth ofgravel over eggs and embryos ranged from 10 to 33cm and averaged 18.8 cm.

The range of depths and velocities within whichchinook have been observed to spawn suggeststhat establishing meaningful minimum and maxi-mum criteria for these factors is problematic. Al-though conventional wisdom states that chinookprefer deeper, faster rivers for spawning than theother Oncorhynchus species, measures of spawningarea characteristics for other species do not con-firm this wisdom (e.g., Burner 1951). Availablemeasurements do not suggest that chinook avoidshallow water and low flows. Chinook may, how-ever, spawn in water that is deeper and fasterflowing than that used by other species becausethey are large enough to hold position in the fastercurrent and to build a redd in the coarser gravelfound there.

Minimum spawning depth is presumably gov-erned by water depth needed for successful dig-ging and spawning; however, Burner's (1951)observations suggest that this can be accom-plished in as little as 5 cm of water. It is not clearwhy increasing depth should ever be a constraint,unless it is correlated with some other significantfactor, such as velocity. Chapman (1943) observedColumbia River fall (ocean-type) chinook spawn-ing below Kettle Falls in water 15 ft (4.6 m) deep. Inone spot, as many as thirty fish appeared to bespawning together in one large redd, although healso observed numerous individual redds andsome with a few fish digging together. Accordingto Indians whom he interviewed, this sort of mainstem spawning was commonplace in the ColumbiaRiver when chinook were abundant and suchspawning places had been important fishingplaces for them. Chapman (1943) considered suchspawning to be unusual, however, and stated that,in his experience, chinook normally spawned atthe head of a riffle in 0.3-1.2 m of water. Chapman(1943) speculated that the spawning below KettleFalls may have been stimulated by high subgravelflow rates below the falls.

Other authors have emphasized the importanceof subgravel flow in the choice of redd sites bychinook. Vronskiy's (1972) comment about the im-

portance of the tailspill gravel mound in stimulat-ing subgravel flow when redds are placed in lowvelocity water has already been mentioned. Inaddition, Vronskiy (1972), like Chapman (1943),observed that most redds (95%) were located at thehead of a riffle, just before the crest of the rapid.Vronskiy (1972) attributed the attractiveness ofthis location to the high subsurface flows thatoccurred there. Other spawning occurred in poolsbelow log jams where the log jam increased therate of subgravel flow. In the Nechako River mostchinook spawn on the upstream sides of largegravel dunes oriented across the river channel,presumably to take advantage of the subgravelflow stimulated by the dune (Russell et al. 1983).

Provided the condition of good subgravel flow ismet, chinook apparently will spawn in water thatis shallow or deep, slow or fast, and where thegravel is coarse or fine. The requirement for goodsubsurface flow is consistent with the probableincubation requirements of chinook relative to theother species. Chinook have the largest eggs(Rounsefell 1957) and, thus, their eggs have a smallsurface-to-volume ratio compared with the otherspecies of Pacific salmon. Their eggs should, there-fore, be more sensitive to reduced oxygen levelsand require a more certain rate of irrigation. Silveret al. (1963) observed that the size of chinook athatching was dependent on water velocity in theincubation apparatus even at velocities as high as1,350 cm/h, and on oxygen concentration evennear saturation levels, at least when the incubationtemperature was about 11°C.

The apparent preference of chinook for spawningareas with high subgravel flow may explain theirtendency to aggregate in particular locations forspawning and to ignore other, superficially similar,areas (Vronskiy 1972). Within areas of aggregationthe distribution of spawning nests is not randombut, rather, tends to an even distribution (Neilsonand Banford 1983). Burner (1951) suggested thateach spawning pair defends an area equal to aboutfour times the area of its redd, and he recommendedthat the available spawning area be divided by fourtimes the average redd area to arrive at an estimateof the maximum spawning population that the areacould support. According to Burner's (1951) esti-mates of redd size, stream-type chinook requireabout 16 m2 and ocean-type chinook require about24 m2 of gravel per spawning pair.

323


The Chinook's apparent need for strong subsur-face flow may mean that suitable chinook spawn-ing habitat is more limited in most rivers thansuperficial observation might suggest, so that athigh population density many chinook spawn inareas of low suitability, and their eggs conse-quently suffer high mortality. If this is the case, thecontinued high production of chinook in spite ofgreatly reduced spawning populations (Healey1982a) becomes more understandable, since theapparent reduction in spawning populations willnot have been accompanied by a correspondingreduction in fry production.

Length of Residence on the Redd

Chinook females in the Morice River, a tributary ofthe upper Skeena River drainage, spent between 4and 18 days defending their redd after they began

spawning (Neilson and Geen 1981). Chinook fe-males in the Nechako River, a tributary of theupper Fraser drainage, spent between 6 and 25days defending their redds (Neilson and Banford1983). The average length of residence on the reddsdeclined throughout the spawning period fromabout 14 days (early in the season) to about 5 days(late in the season) on the Morice River, and fromabout 15 days (early in the season) to about 4 days(late in the season) on the Nechako River. Both theMorice and Nechako river populations are mainlystream-type, although scale analysis indicates thepresence of ocean-type chinook as well. As far as Iam aware, these are the only measurements ofresidence on a redd for chinook. Apparently, nomeasurements exist for males, and, as Neilson andGeen (1981) pointed out, obtaining such measure-ments would be complicated by the fact that malesare not faithful to a single redd.

FECUNDITY

The earliest measurements of fecundity in chinooksalmon are those that McGregor (1922, 1923) re-ported for the Klamath and Sacramento rivers inCalifornia. McGregor's data indicated considerablevariation in fecundity within each population butan even larger difference between populations.McGregor (1923) went so far as to propose thatdifferences in fecundity would permit separationof Sacramento River and Klamath River chinookcaught at sea. Healey and Heard (1984) summa-rized more recent data on the fecundity of 16 addi-tional populations and confirmed the high intra-and interpopulation variation in chinook fecunditythat McGregor (1923) had described. Fecundity ofchinook females ranged from fewer than 2,000 eggsto more than 17,000 eggs. Fecundity was signifi-cantly correlated with female size in all but one ofthe populations examined to date. Size, however,explained only 50% or less of the variation in fe-cundity between individuals within a population(Figure 8). The slope of the fecundity on lengthrelationship for chinook was also low (generallyless than two in chinook compared with otherfishes in which it is generally greater than three)(Healey and Heard 1984).

In several instances, fecundity had been mea-sured for the same population over a number ofyears. Significant variation in average fecunditybetween years was evident in these populations,although the absolute interannual variation wasless than that observed between years in otherspecies (Healey and Heard 1984). The only otherfactor that appeared to contribute to within-popu-lation variation in fecundity was a small differencebetween fish of red and white flesh colours in theFraser River population (Godfrey 1968a). Age ap-parently contributed nothing to variation in fecun-dity beyond that predicted by the difference insizes between ages (Healey and Heard 1984).

Between-year and flesh colour variation in fe-cundity explained only a small additional amountof the within-population variation in fecundity, sothat a great deal of individual variation remains tobe explained. Healey and Heard (1984) speculatedthat this high variation may reflect an uncertaintrade-off between egg size and egg number in theoverall fitness of chinook populations. Unfortu-nately, there are no data on egg size in chinooksufficient to demonstrate that the more highlyfecund fish within a size class also have smaller

324


14

12-

10

r'_ 6-

Nushagak R1968

600 700 800 900 1000

10-

8-

Quinsam R.1975

I .'

F=0.00195L'

600 700 800 900

Post-orbit-hypural length (mm)

1000

FIGURE 8«iwu examples of the relationship between length and

Ljjteeundity for chinook salmon. Nushagak River chinookfe a stream-type race having high average fecundity,

i' and Quinsam River chinook are an ocean-type race{having low average fecundity. Although the fecundity

i length regressions are significant in both instances,!* the amount of variation in fecundity attributable to4j variation in length is relatively small.

Egg size has been shown to increase with&tnale size in Oregon chinook stocks (Nicholas

i Hankin 1988).Fecundity, therefore, appears to be less deter-led by body size in chinook than in other fishes.rtly, this may be due to the unresolved trade-off

ft. *ween egg size and egg number in chinook men-MlOrted above, so that egg size varies more between| Ihinook individuals than is usual for fishes. Also, itAppears that in the trade-off between body size

1 fecundity among older chinook, body size was|SWiore critical, so that more energy is devoted to

somatic growth and less to egg production than in

other fishes (Healey and Heard 1984). More energyis devoted to individual eggs among these largefish as well (Nicholas and Hankin 1988).

Although, within populations, interannual vari-ation in fecundity was significant, between-popu-lation variation was numerically greater (Healeyand Heard 1984; Nicholas and Hankin 1988). Aver-age fecundity at size varied approximately two-fold between populations (4,347-9,427 at 740 mmpost-orbit/hypural length). In general, fecundityincreased from south to north in the chinook'srange, contrary to Rounsefell's (1957) conclusionbased on a few samples (Figure 9). The SacramentoRiver population, at the southern limit of the Chi-nook's range, however, has an unusually high fe-

10

9-

1 8H8g£

6-

5-

o

••

t.»

40 60 63

Degrees N. latitude

FIGURE 9The relationship between average fecundity of chinookat 740 mm postorbit-hypural length and the latitude atwhich the population spawns (from data in Healey andHeard 1984; Nicholas and Hankin 1988). Symbols are:• = ocean-type populations having a common slope of

fecundity on length regression; A = ocean-type popula-tions having slopes of fecundity on length regressiondifferent from the above; o = stream-type populationshaving a common slope of fecundity on length regres-

sion. The overall correlation between fecundity and lati-tude is significant (r = 0.663, p <.05). If the populationswith symbol A are left out of the regression, the rela-tionship for ocean-type populations is also significant

(r = .642, p <.05). Within the stream-type race, however,there is no relationship between latitude and fecundity.

325


cundity. This population, as well as three others(Quinsam River, Puntledge River, Rivers Inlet, allin British Columbia), differed from all other popu-lations in the slope of the regression of fecundityon length (Healey and Heard 1984). Thus, it isdifficult to compare the Sacramento River fish di-rectly with most other populations. The high-fe-cundity populations near the northern limit of thechinook's range are all stream-type fish whichspend a year in fresh water before going to sea,whereas the low-fecundity populations in thesouth are mainly ocean-type fish which go to seaduring their first year of life. Thus, latitudinal dif-ferences in fecundity may partly reflect a racialdifference between stream- and ocean-type chi-nook rather than a latitudinal cline. In the Colum-

bia River, however, data are available on the fecun-dity of both stream- and ocean-type chinook. Al-though stream-type fish had a higher fecunditythan ocean-type, the difference between the raceswas not statistically significant (Galbreath andRidenhour 1966; Healey and Heard 1984). If thedata are segregated into stream- and ocean-typelife histories, there is still a latitudinal cline infecundity within the ocean-type life history, pro-vided the Sacramento, Puntledge, and Quinsamrivers and Rivers Inlet populations are excluded(r = .642, p < .05). The fecundity of stream-typepopulations alone is not significantly correlatedwith latitude, but, for all populations combined,the correlation is significant (r = .663, p < .05).

SPAWNING, INCUBATION, AND SURVIVAL

Egg Deposition

The fecundity of females represents only the po-tential for production of the next generation. Thispotential is subject to successive losses that, in astable population, ultimately result in an averageproduction of one adult female spawner for eachfemale spawner in the parent generation. Many ofthese successive losses have not been documentedfor chinook, except in an anecdotal way. They willbe listed here in order of their occurrence, togetherwith whatever estimates of loss are available in theliterature. As before, the emphasis will be on evi-dence of variation among populations as adapta-tions to local environments.

Females that die unspawned on the spawninggrounds, or that do not spawn all their eggs, repre-sent an important potential loss in egg production.Such losses are generally lumped together in esti-mates of unspawned egg retention, and these esti-mates are seldom large (Chapman et al. 1986).Vronskiy (1972) reported that egg retention wasgenerally about 0.6% of absolute fecundity for chi-nook; Major and Mighell (1969) reported it to beabout 0.5% for Yakima River chinook; and Shep-

herd (1975) reported it to be 1.3% for Morice Riverchinook. Paine et al.(1975), however, reported anaverage of 11.9% egg retention for Big QualicumRiver chinook, and Shepherd (1975) reported 25%egg retention for Bear River chinook with 9 of 47females unspawned and 20% egg retention for Ba-bine River chinook with 30 of 230 females un-spawned. Fish in these latter two rivers, however,had been subject to harassment. In 1965 about 25%of adult chinook died without spawning in aspawning channel at Priest Rapids, Washington.This mortality was apparently caused by an infec-tion of the gills with a protozoan of the genusDermocystidium (Pauley 1967).

Ovarian disease may be a cause of reduced fe-cundity or egg deposition. A condition termed"bad eggs" has been known for Columbia River fall(ocean-type) chinook since the early 1940s. Thecondition is characterized by a number of signs,including a pus-like discharge from the vent anddead white eggs, either individually or in clusters,on the surface of the ovary or near the point ofattachment. The condition may be present in oneor both ovaries and occurred in 2%-20% of femalesreturning to Columbia River hatcheries (Conrad

326


1965). Apparently, conditions of this type are not aproblem in British Columbia hatcheries or knownto be a problem in wild stocks (G. Hoskins, De-partment of Fisheries and Oceans, Pacific Biologi-cal Station, Nanaimo, British Columbia, pers.comm.).

Losses to eggs actually shed by females may oc-cur in a variety of ways. Some eggs will be sweptout of the redd during spawning and will be subjectto heavy predation. Some eggs will not be fertilized.Some will not be buried deeply enough and will beaccessible to vertebrate and invertebrate predators.Floods, siltation, freezing, desiccation, and diseasecan all take a toll. Poor gravel percolation or poorwater quality can cause mortality of eggs. Finally,the quality of eggs laid and the embryos produced,which themselves depend on the genetic and phe-notypic quality of the parents, may influence thesurvival of eggs. Few of these sources of egg andembryo loss have been quantified.

Opinions differ as to the quantity of eggs lostduring spawning by being swept out of the redd.Vronskiy (1972) believed that there was a very highloss of eggs during spawning. He based this con-clusion on the number of eggs recovered duringcomplete excavation of redds. No more than 30% ofthe average fecundity of chinook females was everrecovered from a redd, and the average number ofeggs recovered was about 12% of. average fecun-dity. Briggs (1953), however, believed that very feweggs were lost during spawning. The difference inopinion between these authors may be due to thefact that Vronskiy (1972) made his observations ata spawning area where velocity was very high,whereas Briggs (1953) conducted his investigationson small streams with lower velocity. Unpublishedobservations of my own, on a high-velocity riffle inthe Nanaimo River, British Columbia, indicatedthat few eggs were lost during spawning even infast flowing water. Both Vronskiy (1972) andBriggs (1953) commented that trout, charr, andother small fish may dart into a redd and steal afew eggs while the female is spawning, but the lossdue to this kind of predation was not quantified.

Most eggs deposited in redds appear to be fertil-ized. Briggs (1953) reported various studies thatdemonstrated that between 92% and 98% of eggswere successfully fertilized. Vronskiy (1972) re-ported fertilization in excess of 99% for KamchatkaRiver chinook.

Survival during Incubation

Information on mortality of fertilized eggs andagents of mortality comes from observations madeof both artificially planted eggs and natural redds.Shelton (1955) investigated the survival to hatch-ing and emergence of eggs planted at differentdepths in two sizes of gravel in artificial streamchannels and subjected to several rates of waterpercolation through the gravel. He concluded thatsurvival to hatching was greater than 97%, regard-less of planting depth in the gravel or gravel size,provided the percolation rate was at least 0.001 ft/s(0.03 cm/s). Emergence was 13% or less, however,from small gravel and when percolation was lessthan 0.002 ft/s (0.06 cm/s). Eighty-seven per cent offry emerged successfully from large gravel withadequate subgravel flows.

Alderdice and Velsen (1978) reviewed the avail-able information on rate of egg development andtemperature for chinook. Upper and lower temper-atures for 50% pre-hatch mortality were 16°C and2.5°-3°C, respectively, when the incubation tem-perature was constant. When incubation tempera-ture varied with the ambient temperature,development rate and survival were better at lowtemperatures than when incubation temperaturewas constantly low. Presumably, this better perfor-mance reflected the development of greater low-temperature tolerance after initial cell division.Time to 50% hatch ranged from about 159 days at3°C to 32 days at 16°C.

Alderdice and Velsen (1978) concluded that a loginverse form of Belehradek's equation (develop-ment rate = (temperature - C) exp b/K; where C, b,and K are constants) gave the best fit to the avail-able data on egg development in relation to tem-perature. Even this model, however, underesti-mated development rate at low temperatures. Formost practical purposes a simple thermal summodel (development time = 468.7/T; where T is theaverage temperature during incubation) was ade-quate for predicting time to hatching.

Gangmark and Bakkala (1960) recorded percola-tion rate, temperature, and oxygen concentrationnear eggs planted in Mill Creek, California. Normaltemperature fluctuations were not related to eggsurvival/but both percolation and oxygen concen-tration were. Mortality of eggs increased with de-creasing percolation rate, being 2.9% at 4.0 ft/h

M

327


(0.034 cm/s) and nearly 40% at 0.5 ft/h (0.0042 cm/s). Mortality also increased rapidly at dissolvedoxygen concentrations below 13 ppm, averaging3.9% at 13 ppm and 37.9% at less than 5 ppm.

The survival of eggs in undisturbed naturalredds appears to be quite good. Vronskiy (1972)reported survival of 97% to hatching, and Briggs(1953) reported 90% survival to the eyed stage and82% to hatching. Vronskiy's estimate is based onthe number of live and dead eggs and alevinsrecovered during excavation of total redds,whereas Briggs' estimates are based on samplesfrom redds. Both authors attempted to correct forthe disintegration of dead eggs, Briggs by plantingdead eggs and observing the losses over time, andVronskiy by counting egg shells as disintegratedeggs. Despite these corrections, the estimates ofsurvival to hatching must be taken as maximumestimates for successful redds. Neither authordealt with losses due to scouring or siltation.Briggs (1953) did, however, observe some interest-ing instances of high mortality in redds due toattacks by an undescribed species of oligochaeteworm. This is the only description of invertebratepredation on chinook eggs buried in a redd ofwhich I am aware, although, under the right cir-cumstances, eggs must be available to a variety ofinvertebrate predators.

Stream conditions during incubation can have adramatic effect on the survival of eggs to hatchingand emergence. In a series of experiments at MillCreek, California, Gangmark and Broad (1955) andGangmark and Bakkala (1960) demonstrated thatflooding in Mill Creek was an important cause ofhigh mortality of chinook eggs. Apart from loss ofeggs washed out of the gravel by floods, mortalitywas associated with low oxygen concentrations inspawning gravel «5 ppm) and poor percolation ofwater through spawning gravel, with increasingmortality below percolation rates of 4.0 ft/h (0.034cm/s).

Temperature has seldom been implicated in anysignificant loss of eggs during incubation. Coombsand Burrows (1957) speculated that salmon spawn-ing in cold headwater streams with temperaturesbelow 40°F (4.4°C) would suffer high mortality,and, at the other extreme, Slater (1963) concludedthat winter-run chinook spawn would suffer highmortality in some tributaries of the SacramentoRiver because of high water temperature during

incubation.Chinook normally begin spawning in late

summer and may begin spawning when tempera-tures are near 16°C, the upper temperature limitfor 50% egg mortality (Alderdice and Velsen 1978).As temperatures are falling rapidly at this time ofyear, however, the eggs are probably not exposedto near lethal temperatures for long. Similarly,temperatures may drop below 3°C part waythrough the incubation period in ice-covered riv-ers. The impact of seasonal exposure to extremetemperatures on survival and viability of eggs andalevins has not been studied systematically, butpresumably the embryos are able to survive condi-tions such as these, which are typical in spawningrivers.

Adequate water percolation through the spawn-ing gravels is essential for egg and alevin survival.There is no doubt that percolation is affected bysiltation and that siltation in spawning beds cancause high mortality (Shaw and Maga 1943; Wick-ett 1954; Shelton and Pollock 1966). There appearsto have been no systematic study of the interrela-tion between river discharge (velocity), sedimentload, and survival of chinook spawn. Of particularsignificance to any assessment of the probableeffects of siltation on survival is Shaw and Maga's(1943) observation that siltation resulted in great-est mortality when administered early in incuba-tion. Thus, siltation during winter freshets incoastal rivers or during summer peak discharge insnow-fed rivers may have a greater effect on theamount of suitable spawning gravel than on thesurvival of previously deposited spawn.

Becker et al. (1982,1983) investigated the effectsof dewatering artificial chinook redds on survivaland development rate of embryos at various stagesof development. Becker et al. (1982) defined stageof development in terms of accumulated thermalunits during incubation and studied four stages:cleavage eggs, incubated for 56-168 Celsius de-gree-days (DD), embryos (249-467 DD), eleuthe- ,roembryos (553-575 DD), and pre-emergent alevins(780-814 DD). Dewatering of redds is likely to occurin regulated rivers where discharge is varied tosatisfy some domestic or industrial need but couldalso occur in natural rivers. Alevins were mostsensitive to both periodic short-term dewateringand a prolonged single dewatering, surviving atless than 4% in periodic dewaterings of one hour or

328


a single dewatering of six hours. Eleutheroembryoswere less sensitive, and cleavage eggs and embryosleast sensitive. In fact, embryos apparently sufferedno ill effects from daily dewaterings of up to 22hours over a 20-day period. The development ratewas also reduced in those instances in which sur-vival was affected but not in instances when sur-vival was good. These results seem at variance withthe observation of Silver et al. (1963) that chinookembryo development is highly sensitive to anyreduction in oxygen concentration or percolationrate. Since the dewatered eggs and embryos re-mained damp, however, they probably suffered noshortage of oxygen. Elimination of metabolic wasteproducts may have been a problem.

Emergence

Estimating survival to emergence poses significantproblems with chinook, as some fish migratedownstream as fry whereas others rear for a vari-able length of time in the river before migratingdownstream. Counts of downstream migrants,therefore, provide only a minimum estimate of thenumber of fry that emerged. In Mill Creek,85%-100% of fertilized eggs deposited in plasticmesh bags in the gravel were lost prior to emer-gence. These losses were associated with floods inthe creek. In a channel with controlled flow, themortality of planted eyed eggs to emergence wasonly 40% (Gangmark and Bakkala 1960). In FallCreek, California, Wales and Coots (1954) andCoots (1957) found a 68%-93% mortality from eggdeposition to the emergent fry stage. The 93%mortality was associated with floods. From reddexcavation, Gebhards (1961) estimated that only42% of alevins would have emerged from a singleredd. Although not well documented, it appearsthat emergence may be a difficult time for fry.

Apparently gravel conditions can influence thesuccess of emergence. Shelton (1955) found thatonly 13% of hatched alevins emerged from experi-mental troughs in which eggs were planted in finegravel compared with 80%-90% emergence introughs with coarse gravel. Emergence from finegravel was further influenced by the depth ofplanting and water velocity through the gravel.Greater emergence occurred when the eggs wereplanted near the surface and when water velocitywas low.

Major and Mighell (1969) estimated that5.4%-16.4% of spring chinook survived to migrateas yearling smolts from the potential egg deposi-tion in the Yakima River, Washington. In the Cowi-chan River, British Columbia, 9.2% and 16.5% ofpotential egg deposition survived to migrate as fryand fingerlings (Lister et al. 1971), whereas in theBig Qualicum River survival from potential eggdeposition to fry and fingerling migrants was0.2%-7.0% prior to flow control and 12.0%-19.8%after flow control (Lister and Walker 1966; Paine etal. 1975). M.D. Bailey (Department of Fisheries andOceans, Vancouver, British Columbia, pers.comm.) estimated that 4%-50% of potential eggdeposition migrated as fry into the lower FraserRiver, British Columbia, but cautioned that theseestimates were based on poor data. Healey (1980b)estimated that about 12%-20% of potential eggsdeposited migrated downstream as fry in the Na-naimo River, British Columbia. All these values aredifficult to interpret because of uncertainty in esti-mates of both the potential eggs deposited and thenumbers of fry produced. The values do suggest,however, that, barring serious floods, egg-to-frysurvival in chinook is relatively good.

Summary of Egg-To-Fry Survival

Published estimates of the mortality rate betweenegg laying and fry emergence are so few and sovariable that it is difficult to draw any firm general-izations (Table 3). In particular, there is little evi-dence that can corroborate or refute my assertionsabout variability among chinook populations. Un-der natural conditions, 30% or less of the potentialeggs deposited resulted in emergent fry or fry andfingerling migrants in the systems studied (Table3). When and how eggs die or are lost to the popu-lation is uncertain. Two features do bear somecomment, however. Eggs properly buried in a reddthat remains undisturbed, or which are artificiallyplanted where subgravel percolation is good, ap-parently survive well. -Floods, which scour thebottom or result in heavy siltation, are generallyassociated with high egg mortality, as is dewater-ing of redds (but see Becker et al. 1982,1983). Flowcontrol appears to result in a significant increase inaverage survival. These observations suggest thategg-to-fry and fingerling mortality probably oc-curs either at the time of spawning or as a result of

329


TABLE 3Published estimates of mortality (%) of chinook to various development stages in fresh water

(mean of ranges in parentheses)

Riversystem

Mill Cr. (CA)

Fall Cr. (CA)Prairie Cr. (CA)

Yakima (WA)

Lemhi (ID)

Cowichan (BC)

Nanaimo (BC)

Big Qualicum (BC)

Skeena SystemBear R. (BC)Morice R. (BC)Babine R. (BC)

Kamchatka (USSR)

Eggsnot

spawned

1.0

12

251

201

Losses Spawning Spawning Spawning Spawningat to to to to

spawning eyed stage alevin emergence fry/smolt

85-100 (96)

40

68-93 (85)1.0 0-25.5(10) 14-25(18)

84-95 (89)

27 58

84-91 (87)

80-88 (84)

93-10080-88

88 1-6 (3)

Remarks

Planted eggs,flooding channel

Planted eggs,controlled flow

Natural spawningNatural spawning,

redd samplingStream-type, weir

counts of smoltsEmergence trap

over one reddRatio of fry/smolt

migrants to eggsRatio of fry/smolt

migrants to eggsBefore flow controlAfter flow control

Redd sampling

Source: See text for sources

redd disturbance due to floods. Losses at the timeof spawning in particular bear further investiga-tion, in view of Vronskiy's (1972) comment thathigh egg losses at spawning have been observedby Russian biologists for several species of Pacificsalmon, and the International Pacific Salmon Fish-eries Commission's observation that losses of eggsat spawning in Adams River sockeye were heavywhen the spawning stock was large (T. Gjernes,Department of Fisheries and Oceans, Pacific Bio-logical Station, Nanaimo, British Columbia, pers.comm.).

Major and Mighell's (1969) observations onspring chinook deserve comment, as their esti-mates of survival from egg deposition to smoltmigrants are comparable to estimates of survivalfrom egg deposition to fry and underyearling smoltmigrants for fall chinook in other systems. It istempting to assert that this is an example of thedichotomy between stream- and ocean-type chi-nook. There are, however, a number of possible

sources of error in the data on which these esti-mates were based. Major and Mighell (1969) esti-mated potential egg deposition by counting reddsand multiplying by the average fecundity of fe-males in the Columbia River. In my view, thisprocedure is liable to give a biased estimate ofpotential eggs deposited because false redds maybe counted as true redds, because redds may bemissed, or because the fecundity of local stocksdiffers from that of the general population. Theapparent high survival of spring chinook from eggto smolt may, therefore, be fortuitous. Egg-to-mi-grant survival was negatively correlated with reddcount and potential egg deposition in the Yakima.River, even excepting an unusual mortality of eggsin the upper Yakima River due to low flows in 1957(Figure 10). Such a relationship is consistent withthe notion that redd counts may have been inaccu-rate, so that variations in survival were more areflection of error in redd counts than real varia-tion in survival. It is also consistent, however, with

330


500

Number of redds

1000 1500 2000

15

10

ico

5-

(60) «o

(58)•o

(59)• o

•(57)e

0 1 2 3 4 5 6 7 8

Potential egg deposition (x10e)

FIGURE 10The relationship between the number of spring chinookredds counted and estimated egg-to-smolt survival (•),and between potential egg deposition and egg-to-smolt

survival (o) for the Yakima River, Washington. Broodyears are in brackets. (Data from Major and Mighell

1969)

the possibility suggested earlier that good reddsites are few in most rivers, and that fry and smoltproduction may be more related to the amount ofgood spawning area than to the number ofspawners. A third possibility is that yearling smoltproduction is independent of spawner density overa wide range of spawner densities. Lister andWalker (1966) noted a rather small variance in un-deryearling smolt production compared with fryproduction in the Big Qualicum River and specu-lated that rearing habitat was a limiting factor insmolt production.

FRESHWATER RESIDENCE AND DOWNSTREAM MIGRATION

Fry Migrations

Factors controlling the emergence of chinook fryfrom the spawning beds are .not well studied. Ac-cording to Reimers (1971), most emergence is atnight, although up to 20% of fry emerged duringthe day in his experimental troughs. Emergencewas reduced during a full moon. In Reimers' (1971)experiments, emergence peaked just after alevinsreached maximum weight.

Upon emergence, fry swim, or are displaced,downstream. Thomas et al. (1969) found that fallchinook fry go through a period of reduced swim-ming ability just before the time of complete yolk

absorption, and that this coincided with the timeof peak downstream migration. They hypothe-sized that reduced swimming ability was the causeof downstream migration.

Downstream movement of fry occurs mainly atnight, although small numbers may move duringthe day. Peak nightly catches of downstream mi-grants may occur before (Reimers 1971), at (Listeret al. 1971), or after midnight (Mains and Smith1964). Differences in time of peak catch most likelyreflect the distance of trapping sites below themain spawning area rather than differences in timeof emergence from the gravel. Reimers (1971) alsoobserved that downstream movement was inhi-

331


bited by bright moonlight, so that, on moonlitnights, peak trap catch was shifted from before toafter midnight.

Once started downstream, chinook fry may con-tinue migrating downstream to the river estuary,or may stop migrating and take up residence in thestream for a period of time ranging from a fewweeks to a year or more. What determines whetherfry will hold and rear in the river, or migrate down-stream to the estuary, is unknown. Kjelson et al.(1981) observed that peak catches of chinook fry inthe Sacramento-San Joaquin delta often followedflow increases associated with storm runoff. Theyspeculated that flow surges influence the numbersof fry that migrate from upper river spawninggrounds to the delta. Healey (1980b) also observedthat downstream movement of fry was correlatedwith river flow in the Nanaimo River. Reimers(1968) and Lister and Walker (1966), however, spec-ulated that social interaction or density-dependentmechanisms may cause fry to be displaced down-stream. Reimers (1968) observed lateral displays,chasing, nipping, fighting, fleeing, submission,and redirected aggression among juvenile fall chi-nook in stream tanks, where the agonistic behav-iour of one or a few dominant fish apparentlystimulated the downstream movement of subordi-nate fish. These same behaviours, with the excep-tion of nipping and redirected aggression, alsooccurred in natural stream populations of chinook.Lister and Walker (1966) observed that, in the BigQualicum River, fry migrants varied almost 100-fold in abundance between years, whereas finge-rling migrants varied only about ten-fold in abun-dance. They concluded that available freshwaterrearing area limited the number of fry that couldreside in the river, and that the rest were displaceddownstream. A similar conclusion could be drawnfrom Major and Mighell's (1969) observations onproduction of stream-type smolts in the YakimaRiver.

River discharge and intraspecific interactionmay both play a role in stimulating downstreammovement of chinook fry. Other factors, however,may also be important. Stein et al. (1972) observedthat juvenile coho apparently were dominant tochinook and grew faster in sympatric groupings instream troughs. Chinook were able to grow asrapidly as coho, however, when alone in thetroughs. Stein et al. (1972) speculated that interac-

tion with coho may influence the downstreammovement of chinook. Recently, Taylor and Larkin(1986) demonstrated that stream-type chinook fryshowed stronger positive rheotaxis and were moreaggressive towards conspecifics and coho fry thanwere ocean-type chinook fry. Taylor (1988) con-firmed for other stocks that stream-type chinookwere more aggressive but not that they hadstronger rheotaxis than ocean-type chinook. Thesebehaviour patterns are consistent with the ex-pected length of river residence of the two racesand suggest that river residency and its associatedbehaviour patterns may be inherited.

Lister and Genoe (1970) reported habitat segre-gation among juvenile chinook and coho in the BigQualicum River, as did Chapman and Bjornn(1969) and Everest and Chapman (1972) for chi-nook and steelhead (Oncorhynchus mykiss) in head-water tributaries of the Snake River, and Murphyet al. (1989) for chinook and coho in the Taku River.Everest and Chapman (1972) found that under-yearling chinook in summer occurred over all sub-strate types, at all depths, and in water of allvelocities (up to 1.2 m/s) studied, but that abun-dance generally declined with increasing substrateparticle size, increasing depth, and increasing wa-ter velocity. In these studies, chinook were largerand emerged earlier than the associated coho andsteelhead of the same brood year. Since the larger,older fish chose higher velocity habitats, there mayhave been little competition for space between thespecies. Habitat segregation in these studiesseemed to be a mechanism for reducing competi-tion rather than a result of competition.

A large downstream movement of chinook fryimmediately after emergence is typical of mostpopulations (e.g., Lister and Walker 1966; Bjornn1971; Reimers 1971; Healey 1980b; Kjelson et al.1982). The downstream migration of stream- andocean-type chinook fry, when spawning groundsare well upstream, is probably a dispersal mecha-nism that helps distribute fry among the suitablerearing habitats. In the case of ocean-type popula-tions that spawn close to tidewater, downstreammigrant fry may be swept to the river estuary in afew hours. It has been hypothesized that migrantfry swept to the estuary represent those that aresurplus to the carrying capacity of rearing habitatin the river (Lister and Genoe 1970). Often thesefry represent the majority of the emergent popula-

332


tion, however, and it seems doubtful that such awaste of reproductive potential would be adaptive(Figure 11). Furthermore, it is now known thatestuaries provide important nursery habitat forrecently emerged chinook fry (Northcote 1976;Healey 1980b, 1982b; Levy and Northcote 1982).

20

May June

FIGURE 11The temporal pattern in abundance of fry and under-yearling smolts migrating seaward in three VancouverIsland rivers. A - Cowichan River, 1967; B - Big Quali-cum River, 1967; C = Nanaimo River, 1980. Average forklength of migrants in the Cowichan and Nanaimo riversis shown and demonstrates the rapid switch from fry(35-40 mm) to smolt (60-70 mm) migrants in late May.

Downstream migration of fry in the lower Na-naimo River, British Columbia, for example, ap-pears not to be a consequence of limited rearingnabitat in the river. In some years, few chinook"emained to rear in the lower river after the frymigration, yet large numbers of fry consistently"eared in the river estuary (Healey 1980b; Healeymd Jordan 1982). It seems probable that the Na-naimo River estuary is the preferred rearing habi-tat for at least part of the Nanaimo River chinook

population rather than a final refuge for displacedfry. Chinook in the Nitinat River, British Columbia,also migrate to the estuary in large numbers as fry,and most of the underyearling smolts from boththe Nitinat and Nanaimo rivers are produced inthe estuary rather than in the river (Healey 1982a).Similarly, large numbers of ocean-type fry migrateseaward in the Sacramento River, the CowichanRiver (Figure 11), and the Fraser River (M.D. Bailey,Department of Fisheries and Oceans, Vancouver,British Columbia, pers. comm.) and rear in theriver estuaries (Kjelson et al. 1981, 1982; Healey1982a; Levy and Northcote 1982). It should benoted that the principal rearing areas in the Sacra-mento-San Joaquin and Fraser River estuaries areessentially fresh water whereas rearing areas inVancouver Island estuaries range up to 20 ppmsalinity.

In the Nanaimo River there are three geographi-cally separated spawning areas, and fry from thetwo most downstream areas drift down to theestuary. Many fry from the middle spawning area,however, rear in the river and migrate to sea aftersix to eight weeks, and fry from the upper spawn-ing area may spend up to a year in the river beforemigrating to sea (Healey and Jordan 1982). Analy-sis of polymorphic enzyme systems and body mor-phology suggested that the fry from the threespawning areas were genetically distinct and maybe programmed to migrate seaward at differentages (Carl and Healey 1984). Clarke et al. (1989)found that stream-type chinook required a periodof short day length before they would adapt to andgrow well in sea water, whereas ocean-type chi-nook did not. From these and the observations ofTaylor and Larkin (1986) and Taylor (1988), it ap-pears that the downstream movement of fry afteremergence may not be totally involuntary and thatlength of river residency may be dependent uponthe fish's genotype. It is reasonable that stream-and ocean-type chinook would differ in this regardsince length of freshwater residence is an impor-tant distinguishing characteristic for these races.

Downstream movement of fry is normally mostintense between February and May (Figures 11and 12), being earlier in more southern popula-tions. Rich (1920), for example, observed a few fryas early as December in the lower Columbia Riverand in October and November in the SacramentoRiver. The timing of peak downstream migration

i

i!333

1 Pacific Salmon Life Histories

0)0.

1962

Smolts

o

§a!a.

14-,

12-

10-

8-

6-

4-

2-

0

1972

10 20March

10 20April

10 20May

10 20June

FIGURE 12Examples of variation in downstream run timing of

ocean-type chinook in the Big Qualicum River (A) andthe Fraser River (B). Only fry migrants are shown for

the Fraser, but both fry and smolt migrants are shownfor the Big Qualicum. Note that large variation in fry

migration timing is not paralleled by variation in smoltmigration timing in the Big Qualicum.

can vary substantially from year to year in thesame system. Time of peak downstream movementvaried from the fourth week of March to the fourthweek of April in the Big Qualicum River (Lister andWalker 1966), and from mid-March to early Maynear the mouth of the Fraser River between 1964and 1977 (M.D. Bailey, Department of Fisheriesand Oceans, Vancouver, British Columbia, pers.comm.) (Figure 12). The beginning and end of therun appear to vary less from year to year, so that,when the run peaks early, the temporal patternhas a negative skew, and when the run peaks late,the temporal pattern has a positive skew.

In addition to annual variation in the peak of therun, there is tremendous day-to-day variation inthe abundance of downstream migrants (e.g., M.D.Bailey, Department of Fisheries and Oceans, Van-couver, British Columbia, pers. comm.; Healey and

Jordan 1982). The causes of both annual and dailyvariation in run are not well understood. As notedearlier, Kjelson et al. (1981) speculated that down-stream migration in the Sacramento River wasstimulated by high discharge. Mains and Smith(1964) also suggested that peaks in downstreammovement of chinook at Central Ferry on theSnake River were triggered by freshets. The great-est movement of chinook at this trapping locationwas at river temperatures between 4.5° and 13°C,but there appeared to be no relationship betweenmigration and temperature. At Byer's Landing onthe Columbia River, Mains and Smith (1964) foundno relationship between the migration of chinookand either discharge or temperature. At Byer'sLanding, temperatures ranged from 4.5° to 15.5°Cduring the run. In the Nanaimo River, daily varia-tion in the downstream run of chinook fry waspositively correlated with discharge while the runwas increasing in 1975 and 1976, but not while therun was decreasing; yet variation in discharge wascomparable over both the increasing and decreas-ing portions of the run. Greatest fry migrationoccurred at river temperatures of 6.0°-9.0°C in1975 and 8.0°-11.0°C in 1976, but variation in thedownstream run was not correlated with tempera-ture in either year (Healey 1980b). Irving (1986)found that simulated freshets in experimentalstream channels increased the numbers of frymoving downstream, provided water velocity atpeak flows exceeded 25 cm/s.

In larger rivers, chinook fry migrate more at theedges of the river than in the high velocity waternear the centre of the channel (Figure 13) and, whenthe river is deeper than about 3 m, they prefer thesurface (Mains and Smith 1964; M.D. Bailey, Depart-ment of Fisheries and Oceans, Vancouver, BritishColumbia, pers. comm.; Healey and Jordan 1982).These observations provide further support for myearlier suggestion that downstream movement offry is not simply a passive displacement controlledby water velocity, but that some active behaviour ofthe fry helps direct the migration.

Habitat Utilization in Fresh Water

The process by which chinook take up residence ina stream is not well studied. Reimers (1971) ob-served that, on the first night after emergence,virtually all fry drifted downstream in a stream

334


100-

80-

60-

8g 40-•o

Ol:rvo>

Nanaimo River

o00a 1 — 1

n 1L ffl i inCsS 100-

1 80-o>

£C

60-

40- So> o>

Station — ^1 21 oft hank > —

....

—

Snake River ^

m if! f3 4

1 — 1

^

—

5 6 7».Pinht hank

FIGURE 13Lateral distribution of downstream migrating chinookfry in the Nanaimo and Snake rivers showing the ten-dency of the fry to concentrate near the river banks.

Numbers of fry-per-unit-volume of water at eachstation across the river is shown as a percentage of thestation with the greatest fry-per-unit-volume. For both

rivers, stations were distributed with equal spacingfrom bank to bank.

trough. On each succeeding night, however, frythat were moved back upstream in the troughshowed a stronger and stronger tendency to holdposition in the trough. These observations suggestthat stream residence develops over a number ofdays and that fry could be displaced quite fardownstream before taking up residence. The ten-dency for all fry to drift downstream on the firstnight after emergence may partially explain whyfew fry were found rearing in the lower reaches ofthe Nanaimo River, where spawning occurs only afew kilometres from the estuary. This explanationis unsatisfactory, however, as fry were also ob-served drifting into the lower river from moreupstream spawning areas, and presumably thesecould have occupied habitat in the lower river ifthey had been so inclined. Resident fry were rela-tively abundant in suitable habitat near the up-stream spawning areas. Carl and Healey (1984)demonstrated that fry that migrated to the Na-naimo River estuary were genetically and morpho-logically different from those that reared in theriver. Furthermore, Taylor and Larkin (1986) andTaylor (1988) showed that stream-type chinook fry

displayed greater inter- and intraspecific aggres-sion than ocean-type fry and that some stream-type stocks also displayed stronger positive rheo-taxis. The behavioural mechanisms for holdingposition in the river after emergence thus appearto be better developed in some types of chinookfry than in others.

Lister and Genoe (1970) studied habitat segrega-tion among juvenile fall (ocean-type) chinook andcoho in the Big Qualicum River during the springof 1967, when flow in the river was held constant atabout 5.8 m3/s. They examined three sites alongthe river, and each site was subdivided into two orthree habitat types which differed in velocity,depth, and distance from shore. Chinook emergedduring March and April, whereas coho emergedduring May. Thus, chinook arrived in the studyareas at least six weeks earlier than coho. Chinookwere larger than coho at the time of emergenceand, because they emerged earlier, grew evenlarger before the coho fry appeared in the studysites. Smaller fry of both species inhabited mar-ginal areas of the river, particularly back eddies,behind fallen trees, undercut tree roots, or otherareas of bank cover. As they grew larger, bothspecies moved away from shore into midstreamand higher velocity areas. Although the correlationbetween size of fish captured and velocity of sam-pling site within species was weak, chinook werealways larger and more abundant than coho inhigh velocity subareas. Thus, there was importanthabitat segregation between chinook and coho inthe Big Qualicum River, and this was mainly aconsequence of the larger size and earlier emer-gence of the chinook salmon.

Chapman and- Bjornn (1969) and Everest andChapman (1972) reported qualitatively very sim-ilar observations on habitat segregation betweenstream-type chinook and steelhead in the SnakeRiver. Juvenile chinook were most abundant wheresubstrate particle size was small, velocity was low, .and depth was shallow, but were found in smallnumbers in virtually every habitat investigated.Fish size was positively correlated with water ve-locity and depth for both species, but the speciesdiffered in size owing to differences in emergencetiming and fry size between the species.

Murphy et al. (1989) sampled various habitattypes in the lower Taku River for chinook, coho,and riverine sockeye. They found that chinook

335


were mainly in riverine habitat and seldom inbeaver ponds or off-channel sloughs. Velocity andturbidity were the principal factors associatedwith chinook distributions. Chinook were rare instill water or where velocity was greater than 30cm/s. There was little overlap in chinook habitatwith that of coho or sockeye. Thus, habitat segre-gation appears to provide a mechanism for reduc-ing competition between cohabiting chinook andother stream salmonids, and the pattern of segre-gation is similar for stream- and ocean-type races.

The movement of fish offshore and into fasterwater represents a shift from predominantly sandysubstrate to predominantly boulder and rubblesubstrate. Chapman and Bjornn (1969) suggestedthat chinook prefer finer substrates than steelheadof comparable size, but both species showed astrong preference for the rubble type of habitat.Any interpretation of substrate preferences is con-founded by velocity preference, however, andneeds further investigation.

Edmundson et al. (1968) reported limited day-to-day movement of young chinook in a streamaquarium, suggesting strong fidelity to a particularsite. Reimers (1968) reported that juvenile chinookin the Sixes River were primarily solitary animalsand displayed aggressive behaviour towards otherchinook, suggesting the existence of defendedareas in the stream, at least during the day. Chap-man and his co-workers (Don Chapman Consul-tants 1989) observed temporary defence of feedingterritories by chinook in the evening. As with habi-tat preferences, these observations need to be sub-stantiated in other rivers and other situations.

Day and night distributions of chinook instreams may be quite different. Edmundson et al.(1968) and Don Chapman Consultants (1989)found that at night chinook moved inshore to quietwater over sandy substrates or into pools and thatmost settled to the bottom. With returning day-light, these fish returned to occupy the same riffleand glide areas that they had occupied on theprevious day.

Fingerling Migrants

Fish that elect to hold in the river after emergencemay migrate seaward almost any time of year. Inthe southern half of the Chinook's range, manystream dwellers migrate seaward as fingerlings

between April and June of their first year of life(Healey 1980b, 1982b; Kjelson et al. 1981/1982)(Figure 11). Although following close on the heelsof the fry migration, fingerling migrants are readilydistinguished by their larger size. In all the riversstudied, a sharp change in size of fish accompaniesthe change from fry to fingerling migrants (Figure11). Fry migrants normally range from 30 to 45 mmin fork length, although they have been recordedas small as 20 mm and as large as 55 mm (Mainsand Smith 1964; Lister et al. 1971; Healey et al.1977). Many fry migrants still have visible yolk andfew have begun feeding, although those above 44mm fork length may have some food in their stom-achs. Fingerling migrants, on the other hand, nor-mally range from 50 to 120 mm in fork length, andall have been actively feeding for some time (Mainsand Smith 1964; Lister et al. 1971; M.C. Healeyunpublished data).

The factors stimulating downstream movementof under yearling chinook are not known. Althoughit is well documented that June is a month of veryactive downstream migration for fingerlings, theyare known to migrate downstream at other timesof the year as well. The main fingerling migrationtends to be earlier in the southernmost parts of theChinook's range (Kjelson et al. 1981) and is influ-enced by the presence of populations with uniquespawning times (e.g., Slater 1963). In some Colum-bia River tributaries, juvenile chinook were foundto be resident as late as October but were gone inNovember (Reimers and Loeffel 1967). Bjornn(1971) observed downstream movement of stream-type chinook fingerlings in the Lemhi River(Idaho) during the fall months. He proposed thatthis migration represented a redistribution of fishto more suitable wintering habitat. Fingerling mi- 'gration is known to occur through August in theFraser River (Northcote 1976; M.D. Bailey, Depart-ment of Fisheries and Oceans, Vancouver, BritishColumbia, pers. comm.) but is generally observedto be complete by the end of June in most other '-rivers sampled (Lister and Walker 1966; Lister et al. *1971; Healey and Jordan 1982). Reimers and Loeffei ^(1967) were able to relate extended residence in j*Columbia River tributaries to slow growth, and fSjsuggested that size was an important variable in J|determining when fish will move downstream. j|>Nevertheless, downstream migrant fingerlings ^vary substantially in size, both within and be- ^

336


tween rivers (Table 4), so that some factor otherthan size must also play a role. Furthermore, it isknown that chinook may move out of tributariesand into a river main stem, or simply relocatedownstream with the approach of winter (Bell1958; Chapman and Bjornn 1969; Park 1969). Pre-sumably, as Bjornn (1971) suggested, suitablesummer habitat may not be suitable winter habi-tat. The disappearance of chinook from some Co-lumbia River tributaries in October or November(Reimers and Loeffel 1967), therefore, probablyindicates relocation to an instream wintering arearather than seaward migration.

Fingerlings migrate downstream throughout theday, but the majority migrate at night (Mains andSmith 1964; Lister et al. 1971). In the Columbia andSnake rivers, fingerlings apparently preferred theshoreline during migration (Mains and Smith1964). In the lower Fraser River and in the NanaimoRiver, however, most fingerlings migrated in thefastest water near the centre of the river (M.D.Bailey, Department of Fisheries and Oceans, Van-couver, British Columbia, pers. comm.; Healey andJordan 1982).

The rate of downstream migration of chinookfingerlings appears to be both time-and size-de-pendent and may also be related to river dischargeand the location of the chinooks in the river. Cra-mer and Lichatowich (1978) observed that migrat-ing spring chinook fingerlings in the Rogue River,Oregon, travelled downstream only about 0.3-5.0km/d in the upper reaches of the river but travelled6.1-24.0 km/d in the lower reaches during June-September. Rates of downstream migration in theRogue River were also related to fish size and timeof year. Larger chinook travelled downstream fas-ter, and the rate of migration increased with theseason. In 1975, a year of low and stable river flow,the rate of downstream migration was negativelycorrelated with discharge, whereas in 1976, whenflows were higher and more variable, the rate ofmigration was positively correlated with dis-charge. Cramer and Lichatowich (1978) interpretedthe negative correlation in 1975 to reflect a reduc-tion in rearing habitat as discharge dropped andinterpreted the positive correlation in 1976 to re-flect a direct effect of discharge on the migrationrate at higher discharge.

TABLE 4Fork length (mm) of age 0.0 and 1.0 riverine smolts in various rivers and years

Fork length

River

Age 0.0Sixes (OR)Nitinat (BC)Cowichan (BC)

Nanaimo (BC)

Big Qualicum (BC)

Age 1.0Yakima (WA)

gliake (OR)

Upper Columbia (OR)Jpu (BC, AK)brooked Cr. (AK)

Year

19691980196619671978197919801972

195919601961196219631954195519571958195519611961

Mean

62.052.777.372.163.868.863.566.5

125.5124.6127.0134.0132.6101.0101.068.467.984.273.393.5

Range

40-9144-67.563-9860-9157-7552-8449-76

105-170105-170105-17090-17090-16055-14755-14745-10550-9555-14045-11090-140

Source

Reimers (1971)Healey (unpubl. data)Lister etal. (1971)

n

Healey (unpubl. data)Healey & Jordan (1982)

II

Paine et al. (1975)

Major & Mighell (1969)tr

it

it

a

Mains & Smiths (1964)»

Bell (1985)H

Mains & Smith (1964)Meehan&Siniff(1962)Waite (1979)

337M


I

Growth of Fingerlings

Direct estimates of the growth of juvenile chinookin fresh water exist only for the Sacramento River(Kjelson et al. 1982). Inferences about growth inother river systems can be made from seasonalchanges in the size of resident chinook or from thesize of downstream migrants. These estimatesmust be viewed with caution, however, as thelength of freshwater residence is not known pre-cisely for either the downstream migrant fish or forthose captured during river residence.

Tagged chinook fry in the upper SacramentoRiver grew an average of 0.33 mm/d over a periodof 72 days. Fry that had migrated to the freshwaterSacramento-San Joaquin delta, however, grew sig-nificantly more quickly, averaging 0.53-0.86 mm/din two years of observation (Kjelson et al. 1982).Chinook that migrated seaward as fingerlingsmolts in June of their first year of life averaged52.7-77.3 mm fork length in four Vancouver Islandrivers and one Oregon river (Table 4). Assumingthat the length of river residence is indicated bythe difference in run timing between the fry andfingerling smolt migrants, these fish spent an aver-age of about 60 days in the river before migratingto sea. Growth rates thus ranged from a low of 0.21mm/d in the Nitinat River in 1980 to a high of 0.62mm/d in the Cowichan River in 1966. These ratesare comparable with those based on tagged fish inthe Sacramento River and delta.

Yearling Smolts

Stream-type chinook do not migrate to sea duringtheir first year of life but delay migration until thespring following their emergence from the graveland, in northern rivers, sometimes for an addi-tional year as well (Healey 1983). As noted earlier,stream-type chinook characteristically return totheir natal river in spring. Apparent exceptions tothis pattern occur in some Oregon rivers, such asthe Rogue, where spring-run adults produce un-der yearling smolts (Cramer and Lichatowich 1978;Nicholas and Hankin 1988). Chinook that overwin-ter in the larger rivers often move out of the tribu-tary streams and into the river main stem, wherethey occupy deep pools or crevices betweenboulders and rubble during the winter. In theNanaimo River, two lakes along the river main

stem are also used as overwintering areas. A fallredistribution offish, presumably from preferredsummer habitat to preferred winter habitat, hasbeen observed in some systems (Reimers and Loef-fel 1967; Chapman and Bjornn 1969; Bjornn 1971;Carl and Healey 1984; Don Chapman Consultants1989). Bjornn (1971) found that the number ofdownstream migrants in an experimental streamtrough in the fall was related to the presence ofsuitable substrata for overwintering in the experi-mental stream trough.

Yearling smolts normally migrate seaward in theearly spring, sometimes preceding the main migra-tions of fry and fingerlings and sometimes inter-mixed with them. In the Brownlee-Oxbow sectionof the Snake River, yearling smolts migrated down-stream from April to June with peak numbers inMay (Bell 1958). Bell also captured a few down-stream migrant fry, mainly in May. In the YakimaRiver, Washington, yearling smolts migratedmainly in April and May, with the time of peakmovement ranging from the third week of April tothe second week of May (Major and Mighell 1969).Underyearling smolts were not abundant in theYakima River until June. In the Taku River, yearlingsmolts migrated seaward from April to June, withpeak movement in early May (Meehan and Siniff1962). There were no under yearling migrants in theTaku River. The main period of yearling smoltoutmigration from the Kasilof River on the KenaiPeninsula, however, was July (Waite 1979).

Bell (1958) related the peak in migration of year-ling smolts to spring floods and increasing temper-atures. As with the underyearling migration,however, there has been no systematic study of thefactors triggering migration. Yearling migrants doappear to be less nocturnal than underyearlings.Meehan and Siniff (1962) could find no significantdifference in the abundance of migrants betweenday and night in the Taku River, although, onaverage, more smolts moved at night. Major andMighell (1969) also observed greater movement ofyearling smolts during the night, but Bell (1958)found that the greatest movement was during thedaylight hours.

Raymond (1968) found that the rate of down-stream migration of yearling smolts in the Colum-bia and Snake rivers was positively correlated withdischarge, but that rates of travel through free-flowing and impounded sections of these rivers

338


"f

were similar. At low discharge, the rate of migra-tion was 21 km/d, whereas at moderate dischargeit was 37 km/d. These rates of migration are con-siderably faster than those of underyearling smoltsobserved by Cramer and Lichatowich (1978). Therapid migration of smolts through impoundmentson the Columbia River indicates that yearlingsmolts undertake a directed migration that is inde-pendent of river flows.

Growth of Yearling Smolts

Yearling smolts vary greatly in size. In the YakimaRiver, they ranged from 100 to 160 mm in forklength, and the average increased from 124.6 mmto 134.0 mm between 1959 and 1962 (Table 4).These sizes suggest an average growth rate of 0.25mm/d, but the increase in size over time is perplex-ing. Major and Mighell (1969) could not explain theapparent increase in smolt size, but suggested thatit could be due to differential growth of separatetributary spawning populations coupled with thedifferential contribution of these populations tothe smolt run. This explanation is highly specula-tive.

Rich (1920) presented data for the Columbia Riverthat suggest a growth rate of about 0.20 mm/d forspring chinook during the period March - Sep-tember. During this period, the fish increased inlength from 40.0 to 74.5 mm. Mains and Smith (1964)observed that yearling smolts in the Columbia Riveraveraged about 84.2 mm fork length in 1955, for anannual growth rate of about 0.12 mm/d. In the SnakeRiver, by comparison, Mains and Smith (1964) foundyearling smolts to be about 101 mm in 1954 and1955. Thus, growth in the Snake River was close to0.17 mm/d. Bell (1958), on the other hand, foundyearling smolts in the Snake River to be only about68 mm fork length (growth = 0.077 mm/d) in May1957 and 1958. In April 1958, however, Bell (1958)observed a second, larger size mode at 100-104 mmamong the smolts captured in the Snake River. Fishin the larger mode were equivalent in size to thosecaptured by Mains and Smith (1964). During Mayand June, 1958, Bell observed only one size mode,but the position of the mode changed from 70-74mm in May to 85-89 mm in June. The modes in Maytod June appear to be a continuation of the smallerApril mode, with appropriate growth during eachmonth of about 0.33 mm/d.

In the Taku River the size range for yearlingsmolts was 50-105 mm with a mean of 73.3 mm(Meehan and Siniff 1962). Thus, in the Taku Riverthe average growth rate was only about 0.09 mm/d.Loftus and Lenon (1977) recorded mid-eye to forklengths of chinook smolts in the Salcha River (atributary in the upper Yukon River drainage,Alaska) to be 55-86 mm with a mean of 73 mm.These smolts were slightly larger than those in theTaku River when the difference in length measure-ments was taken into account. In Crooked Creek,on the Kenai Peninsula, Alaska, yearling smoltsaveraged 93.5 mm fork length (Waite 1979). Asthese smolts did not emigrate until late July, how-ever, they had the benefit of spring growth in theyear of migration. Assuming their river residencetime was about 430 days, the growth of CrookedCreek smolts averaged 0.124 mm/d.

In the Snake River, therefore, the average annualgrowth for smolts in the smaller mode was compa-rable to the growth in the Taku and Salcha rivers,whereas the fish in the larger mode had grown at arate more comparable to that in the Yakima River.Smolts captured in the Columbia River andCrooked Creek were intermediate in size andgrowth rate.

The existence of two distinct size groups of fishin the same run (e.g., the Snake River) suggeststhat there may be important differences in micro-habitat affecting chinook growth in rivers. If theincrease in size of fish in the smaller mode in theSnake River represents growth, then the rate ofgrowth during the spring months was 0.33 mm/d.Rich (1920) observed growth of 0.20 mm/d in theColumbia River. All these rates are slower than thegrowth rate observed for underyearling smolts.

Major and Mighell (1969) observed that, withinthe same year, the average size of yearling smoltsmigrating downstream in the Yakima River de-creased with time, and suggested that the largerfish migrated first. Bell (1958), on the other hand,observed that the larger fish of one group of year-ling smolts were caught later in the season in theSnake River, and Mains and Smith (1964) observedno systematic change in smolt size with time.These apparently conflicting results may simplyreflect the unresolvable interactions of differencesin stock growth and microhabitat as well as oppor-tunities for spring growth during and prior tomigration.

339


i :

Summary of Freshwater Residenceand Downstream Migration

Although there is some variation in timing, allpopulations of chinook appear to display similarmigratory behaviour. At the time of emergence,there is an extensive downstream dispersal of fry,although some fry apparently are able to take upresidence in the natal river at the spawning site.For populations that spawn close to tidewater, thisdownstream dispersal carries the fry to estuarinenursery areas, whereas in others it serves princi-pally to distribute the fry among suitable fresh-water nursery areas. Later in the spring, thereappears to be a second dispersal that carries somepopulations to the sea or simply redistributes thepopulation within the river system, presumably tomore suitable summer rearing areas. For thosepopulations that remain a year in fresh water thereis a third late fall redistribution to suitable over-

wintering habitat, usually from the tributaries tothe river main stem. Finally, in the spring there is amigration of yearling smolts to sea.

During the late spring and fall redistributions infresh water, the population tends to shift intodeeper water and to move seaward. These changesin habitat are consistent with the shorter termhabitat changes observed by Lister and Walker(1966) and Chapman and Bjornn (1969), in whichchinook moved into deeper, faster water as theygrew in size. The redistributions may punctuatedevelopmental stages as well as achieve more effi-cient utilization of freshwater nursery habitat. Thetendency for redistribution to carry the fish down-stream may be coincidental. Such a movementpattern may also be adaptive, however, by short-ening the length of spring migration for yearlingsmolts, particularly for headwater spawning popu-lations in larger rivers.

MORTALITY AND ITS CAUSES DURING FRESHWATER RESIDENCE

Rates of survival from fry to fingerling migrantstage and from fry to yearling migrant are un-known, with the exception of some recent datafrom the Sacramento River. Based on the oceanreturns of chinook from the same brood year -marked and released as both fry and smolts in theSacramento River at Red Bluff and in the Sacra-mento-San Joaquin River delta - survival from fryto smolt ranged from 3% to 34% for the 1980-82year classes (M. Kjelson, u.s. Fish and WildlifeService, Stockton, California, pers. comm.). Evi-dence from these and other releases of tagged fishin the Sacramento River system suggest that frythat rear in the upper river experience a highersurvival to smelting than fry that rear in the delta(Kjelson et al. 1982; Brown 1986). Survival of smoltspassing through the Sacramento-San JoaquinRiver delta was highly correlated with discharge ofthe Sacramento River.

Major and Mighell (1969) estimated that5.4%-16.4% of potential egg deposition survived tomigrate as yearling smolts in the Yakima River.

Assuming even a high average egg-to-fry survivalrate (30%), fry-to-smolt survival would have tohave been about 30% to account for these rates.This seems too high a rate of survival to be gener-ally true in other populations.

Even though estimates of fry and fingerling mor-tality rates are nonexistent, except for the Sacra-mento system, mortality is presumed to be heavyfn all rivers. Mortality rates of 70%-90% among fryand fingerlings are recorded for other species ofPacific salmon (Foerster and Ricker 1941; Hunter1959; Parker 1965), and, as these are similar to thelosses of chinook observed in the SacramentoRiver system, it seems reasonable to suppose thatchinook in other rivers suffer similar losses. Healey(1980b, 1982b, and unpublished data) could ac-count for only about 30% or less of downstreammigrant chinook fry in the estuaries of the Na-naimo and Nitinat rivers, suggesting that mortalitywas high during this stage of the Chinook's life.

Predators are commonly implicated as the prin-cipal agent of mortality among fry and fingerlings

340


of chinook and other species, and heavy losses dueto predators have been documented in some in-stances (Foerster and Ricker 1941; Hunter 1959).Patten (1971), however, was only able to infer scul-pin predation of l%-4% among chinook fingerlingsreleased from a hatchery in the Elokomin River,Washington, during 1962 and 1963. Most impor-tant predators were the prickly sculpin (Coitusasper) and the torrent sculpin (C rhotheus). Lessimportant were the reticulate sculpin (C. bairdi)and the coast range sculpin (C. aleuticus). However,the release of fingerlings occurred during a singlenight in 1962 (1.5 million) and during three nightsin 1963 (2.5 million), so that the young chinookwere available to the predators for only a briefperiod. Had the releases extended over two tothree weeks, as in a natural fingerling migration,losses to predators might have approached 0.5million (Patten 1971). Also, the large size of thefingerlings relative to the sculpins probably re-duced the efficiency of sculpin predation. Al-though other quantitative estimates of the rate ofpredation on chinook are lacking, various authorshave reported juvenile chinook in the stomachs ofpredatory fishes (Clemens and Munro 1934;Thompson 1959a, 1959b).

Other fish are generally considered to be themost important predators of juvenile salmon, butinvertebrate predators have occasionally been ob-served to kill or injure juvenile salmon. Eisler andSimon (1961), for example, observed that Hydraoligactus caused high mortality among recentlyhatched chinook alevins in hatchery troughs. Ex-

posure of alevins to 400 hydra in 500 ml of waterfor only five minutes was sufficient to cause mor-tality as high as 80%. Dead alevins showed symp-toms comparable to white spot disease. Cohoalevins exposed to as few as 20 hydra in 500 ml forseveral days suffered high mortality. Mortality ofthis magnitude due to hydra may be confined tohatcheries. Novotny and Mahnken (1971) ob-served a marine isopod, Rocinella belliceps pugetten-sis, attacking fry and fingerlings of chum, coho,and pink salmon in aquaria and in the field at nightin Puget Sound. The isopod normally attached tothe young fish on the side behind the dorsal fin.Young salmon so attacked would swim in a dart-ing, erratic, and twisting manner, presumably at-tempting to dislodge the isopod. Even if the isopodattacks were not fatal, the erratic swimming of thefry could attract other predators.

Since the behaviour during comparable life his-tory stages of populations which reside for differ-ent lengths of time in fresh water is essentially thesame, it seems likely that they are exposed tosimilar patterns of instream mortality during thesestages. The existence of different lengths of streamresidence, however, suggests that, at least in thepast, there must have been a survival advantage toprotracted stream residence in some situations andnot in others. The nature of that advantage is notimmediately apparent, particularly in situationssuch as those encountered in the Nanaimo River,where different behaviour patterns coexist in arelatively small river system.

FOOD HABITS IN FRESH WATER

3.1-

The principal foods of chinook while rearing infresh water appear to be larval and adult insects.Kjelson et al. (1982) found Cladocera, Diptera, Co-tpepoda, and Homoptera to be the dominant foods>of chinook fry in freshwater regions of the Sacra-mento-San Joaquin River delta. Chapman andQuistdorff (1938) found dipteran larvae, beetle lar-vae, stonefly nymphs, and leaf hoppers to be theroost abundant diet items of young chinook

(43-152 mm standard length) in tributaries of theColumbia River. Clemens (1934) found that youngchinook in Shuswap Lake fed primarily on terres-trial insects, small crustaceans (mainly Cladocera),and chironomid larvae, pupae, and adults. Herr-mann (1970) found young chinook in the lowerChehalis River feeding principally on crustaceanssuch as Corophium, and on immature and matureinsects. The presence of Corophium in the diet of

!if

341


these fish suggests that they had been feeding inestuarine waters. Rutter (1902) found young chi-nook in the Sacramento River feeding mainly onlarval and pupal insects. Loftus and Lenon (1977)found Diptera, Plecoptera, and Ephemeroptera tobe the most important components of the diet ofchinook smolts in the Salcha River, Alaska. In amore detailed study of the diet of chinook in freshwater, Becker (1973) found that insects constitutedover 95% of their diet in all seasons. Adult Chirono-midae were by far the most important dietarygroup, comprising 58%-63% of the diet. Followingthese, in order of importance, were: larval chiro-nomids (17%-18%), Trichoptera adults (3%-5%),Notonectidae (3%-5%), and Collembola (l%-5%).Some seasonal variation was apparent, with Dip-tera declining in importance from 99% to 70% be-tween March and May, then increasing again to

85% by July. Notonectids were most important inMay, Trichoptera in June and July, and Collembolain April and May. In contrast to these results,Craddock et al. (1976) found crustacean zooplank-ton, especially Cladocera, to be important in thediet of chinook during July-August in the lowerColumbia River. Insects predominated at othertimes of the year.

The importance of insects in the diet of chinookin fresh water indicates that chinook feed in thewater column or at the surface on drifting food.Their basic diet is similar to that of coho, steelhead,and other stream-dwelling salmonids (Mundie1969; Chapman and Bjornn 1969). Whether there iscompetition for food resources among the cohabit-ing species is not known; however, any such com-petition is presumably reduced by the habitatsegregation among species described earlier.

UTILIZATION OF ESTUARINE HABITATS

Fry Migrants

Many of the fry of ocean-type chinook that mi-grate downstream immediately after emergingfrom the spawning beds take up residence in theriver estuary and rear there to smolt size. Recentlyemerged chinook fry are known to rear in theSacramento and Columbia River estuaries (Rich1920; Kjelson et al. 1981,1982), in the Skagit Riverestuary (Congleton et al. 1981), in the Fra'ser Riverestuary (Dunford 1975; Goodman 1975; Levy andNorthcote 1981; 1982; Levings 1982; Gordon andLevings 1984), the Nanaimo River estuary, theCampbell River estuary and other estuaries on theeast coast of Vancouver Island (Healey 1980b,1982b; Levings et al. 1986), and the Nitinat andSomass River estuaries on the west coast of Van-couver Island (Birtwell 1978;Healey 1982b).

In some instances, the salinity of the estuarinerearing habitat is low (e.g., Sacramento River: Kjel-son et al. 1982; Fraser River: Levy and Northcote1981,1982) or is unknown, but observations on theCowichan, Nanaimo, Courtenay, Campbell, and

Nitinat River estuaries demonstrated that chinookfry will rear where salinity is commonly 15-20 ppmor more (Healey 1980b, 1982b; Levings et al. 1986).Thus, estuary rearing may be considered qualita-tively different from rearing in the river channelfurther upstream. Although many chinook fry ap-pear unable to survive immediate transfer to 30ppm salinity, they are clearly able to survivetransfer to 20 ppm or less, and osmoregulatorycapability develops quickly in fry exposed to inter-mediate salinities (Weisbart 1968; Wagner et al.1969; Clarke and Shelbourn 1985). I have trans-ferred chinook fry directly from downstream mi-grant traps on the Nanaimo River into sea water of32 ppm in the laboratory with no apparent short-term ill effects or retardation of growth comparedwith controls maintained in fresh water and brack-ish water of 15 ppm (M.C. Healey, unpublisheddata). Some chinook fry, therefore, appear to beable to tolerate immediate transfer to high salinity.

Rich (1920) reported observing chinook fry inthe Columbia River estuary as early as December,and earlier still, in October and November, in the

342


Sacramento River estuary. During March andApril, fry were abundant in the Columbia Riverestuary. Rich did not measure salinity at the cap-ture sites but did observe that the smallest fryappeared to avoid brackish water and were consis-tently associated with freshwater inflows to theestuary. Without associated data on downstreammigration of fry or estimates of relative abundance,little can be concluded from Rich's (1920) observa-tions other than that healthy fry were present inthe estuary. It is worth noting, however, that onthe basis of a few samples collected irregularlyfrom the Columbia, Sacramento, and other rivers,Rich (1920) hypothesized much of what we nowknow to be true about the early life history ofchinook salmon.

More recently, Kjelson et al. (1981, 1982) haveprovided much more detailed observations on theSacramento-San Joaquin River estuary. Fry arriveat the river delta mainly from January to Marchand reside there for about two months before mi-grating seaward. Most rearing occurs in freshwaterhabitats in the upper delta area, and the fry do notmove into brackish water until they smoltify. Levyand Northcote (1981, 1982) described similar be-haviour of chinook fry in freshwater marsh areas ofthe Fraser River delta. Fry migration to the deltawas later in the Fraser River, however, it occurredpredominantly in April and May.

Observations on a number of Vancouver Islandestuaries and on the Fraser River estuary (Healey1980b, 1982b; Levy and Northcote 1981,1982; Lev-ings 1982) showed that these estuaries were notonly important nursery areas for chinook, but alsothat the distribution of chinook changed season-ally and tidally. At high tide, the young chinookwere scattered along the edges of the marshes atthe highest points reached by the tide. As the tidereceded, the young chinook retreated into tidalchannels and creeks that dissect the marsh areas

(and retain water at low tide. With the incoming; tide, the chinook again dispersed along the edgesjsi the marshes. On the Fraser River, Levy andr Northcote (1982) found that chinook were amongthe last fish to vacate tidal channels in the marsh

,»Vfhen the channels dried up at low tide.3 ' The twice-daily pattern of migration from low-,tide refuges to the marsh areas and back again wasContinued throughout the period of residence offry in the estuaries. As the season progressed,

however, the major concentration of young fishmoved seaward through the delta area in Van-couver Island estuaries. This is partly due to thefact that larger fish appear to prefer deeper water,and that larger fish are able to osmoregulate inhigher salinities. The redistribution of fish may,however, also be associated with increasing tem-peratures in shallow tidal channels, particularly atlow tide. Healey (1980b) found that fry movedaway from sampling stations where temperaturesexceeded 20°-21°C. This seasonal, seaward move-ment was not apparent in the Fraser River estuary(Levy and Northcote 1981). Kjelson et al. (1982)reported that, in freshwater rearing areas of theSacramento-San Joaquin River delta, fry distribu-tion changed from day to night and with fish size.Fry were concentrated near shore in shallow waterduring the day but tended to move offshore atnight. Larger fish also tended to be further offshorethan smaller fry. During the day fry were concen-trated in the upper 3 m of the water column butbecame more randomly distributed in the watercolumn at night.

Fry remain in the estuarine nursery areas untilthey are about 70 mm fork length, after which theydisperse to nearby marine areas. In the FraserRiver estuary, peak abundance of fry in channelsthrough the marsh was April and May. Juvenileswere still abundant in major arms of the river inJune (Dunford 1975; Goodman 1975; Levy et al.1979). Most were gone from the river arms by Julybut remained abundant over the sand flats at thedelta front (Roberts and Sturgeon banks) through-out August (Goodman 1975; Gordon and Levings1984). In the Nanaimo River estuary and otherVancouver Island estuaries, fry were most abun-dant in April-June, but the time of peak abun-dance varied from year to year in accordance withchanges in the timing of the downstream run ofchinook fry (Healey 1980b, 1982b). Sasaki (1966)observed that young chinook salmon were mostabundant in the Sacramento-San Joaquin Riverdelta during April-June, similar to the timing ob-served in more northern deltas. However, Kjelsonet al. (1981, 1982) observed that fry were mostabundant in February and March in the Sacra-mento-San Joaquin River system, and that thesewere replaced by smolts from upriver in April toJune.

The proportion of downstream migrant fry that

- I S

343


find a place to rear in the estuary is not wellknown. For both the Nanaimo and Nitinat Riversystems on Vancouver Island, only 30% or less ofthe estimated downstream migrants could be ac-counted for in the estuary. The fate of the remain-ing 70% is unknown, but it was unlikely that theyreared elsewhere, despite their apparent ability tosurvive and grow in habitats with high salinity.Thorough sampling of other potential nurseryareas in the vicinity of the Nanaimo River in 1976and 1977 failed to reveal any significant numbers ofchinook fry outside the estuary in April and earlyMay (Healey 1980b, 1982b). Levy and Northcote(1981) estimated that the population of young chi-nook in the Ladner marsh complex of the FraserRiver was approximately 305,000 during mid-May,1979. Taking account of the fact that the Ladnermarsh represented less than one-half the marsharea of the Fraser River delta, that some otherimportant habitats had not been taken into consid-eration, and that the delta population would prob-ably turn over several times owing to the arrival ofnew downstream migrants and the emigration sea-ward of fish which had completed their rearing inthe delta, Levy and Northcote (1981) speculatedthat several millions of chinook probably reared inthe delta. The population accounted for in theFraser River delta is, nevertheless, small relative tothe estimated 64 million chinook fry that migratedthrough the lower Fraser River during March-June1979. The fate of the many millions of Fraser Riverchinook fry that do not rear in the delta is un-known. Thus, it appears that there may be highmortality of downstream migrant fry shortly afterthey complete their downstream migration. Theagents of this mortality are unknown.

The residence time of cohorts of fry in estuarinehabitats has been approximated by mark and re-capture studies in the Nanaimo, Nitinat, Fraser,Skagit, and Sacramento-San Joaquin River estuar-ies (Healey 1980b; 1982b; Congleton et al. 1-981;Levy and Northcote 1982; Kjelson et al. 1982). Onthe Nanaimo River estuary, recovery of marked frysuggested a maximum residence time of about 60days. The average length of residence of fry, basedon recaptures of marked fry and on rate of growthand maximum sizes of chinook in the inner estu-ary, was about 20-25 days (Healey 1980b). Resi-dence times in Nitinat Lake were similar to thosein the Nanaimo River estuary, the average in 1979

being about 21 days and in 1980 about 17 days.Levy and Northcote (1982) and Congleton et al.(1981) investigated residence in tidal channelsthrough marshes and found much shorter resi-dence times, about 8 days in the Fraser Rivermarshes and about 3 days in a single channel of theSkagit River marsh. The relatively short residencetimes observed by Levy and Northcote (1982) andby Congleton et al. (1981) may reflect the limitedarea sampled relative to the range of habitat avail-able to fry. In both instances, movement betweentidal channels in the marsh (observed by Levy andNorthcote 1982) and movement of cohorts of fryseaward with time (and thus out of the marsh butnot out of the estuary) (Healey 1980b) may havecontributed to shorter residence time estimates inthese studies. The maximum residence time ofchinook fry in the Sacramento-San Joaquin Riverdelta was 64 days in 1980 and 52 days in 1981(Kjelson et al. 1982). All the residence times so farcalculated include the combined effects of migra-tion and mortality.

Levy and Northcote (1981) investigated the rela-tionship between occurrence and abundance ofchinook fry in various marsh habitats according tothe physical characteristics of the habitat. Chinookabundance was significantly correlated withtwelve of twenty-two habitat characteristics. In amultiple regression analysis, however, only twocharacteristics (area of low tide refugia and eleva-tion of tidal channel banks) explained significantamounts of variation in chinook catch. A largenumber of correlations among the habitat charac-teristics may have confounded the analysis, but theresults suggest that young chinook prefer tidalchannels with low banks and many subtidal refu-gia. Chinook, and associated fish species, also ten-ded to be associated with larger tidal channels.

Fingerling and Yearling Migrants

The apparent movement of fry migrants away fromthe Nanaimo and Nitinat River estuaries in lateMay and June coincided with the downstream mi-gration of fingerling smolts, so that the fingerlingsmolts took over the habitat vacated by the fry. Inestuaries on the Oregon coast there appear to befew fry migrants, and the first chinook to enterthese estuaries are fingerling smolts. (Reimers1971; Reimers et al. 1979; Myers 1980; Myers and

344


Horton 1982). In the Sacramento-San Joaquin Riverestuary, as in the Vancouver Island estuaries, fin-gerling smolts replaced the fry in estuarinenursery areas (Kjelson et al. 1982). The fingerlingsmolts tend to occupy deeper water in the estuaryand to remain there for varying periods. In theNanaimo and Nitinat River estuaries, fingerlingsmolts were abundant during June and July butbegan to decline in abundance about mid-July andwere rare after August, although a few occurred

1001

year-round in the outer estuary of the Nanaimoand other rivers (Healey 1982b) (Figure 14). InOregon coastal estuaries, fingerling smolts appearto reside much longer-well into October (Reimers1971; Myers 1980; Myers and Horton 1982; Nicho-las and Hankin 1988) (Figure 14). The period ofgreatest abundance of fingerling smolts tends,however, to be June to August in Oregon estuaries.In the Sacramento-San Joaquin River estuary, fin-gerling smolts were most abundant from April to

Nanaimoestuary

1

4)tr

FIGURE 14Abundance of juvenile chinook at various locations in the Nanaimo andYaquina estuaries during March to January. Symbols are: • - inner estu-

ary beaches; o = outer estuary beaches; A =• outer estuary deep water.Note the different scales on the ordinates. (Adapted from Healey 1982b

and Myers 1980)

345it


mid-June but were scarce during summer months,apparently because of high water temperature inthe delta and bays (Kjelson et al. 1982). There was asmall secondary peak in smolt abundance in thefall, representing fish that had remained in coolerwater upstream over the summer (Kjelson et al.1982). It is possible that estuaries along the opencoast from Washington to California provide im-portant sheltered habitat for young fall chinookduring the summer and autumn, provided temper-atures in these esturies do not get too high. Shel-tered habitat is much more common along theBritish Columbia coast, and there may be less stim-ulus for young chinook to remain in British Colum-bia river estuaries.

Chinook that migrate to sea as yearling smoltsoften do so together with the emergent fry andthey, too, spend some time in the estuary of theirnatal stream. While the fry are concentrated in thedelta area, yearling smolts occupy the delta front,so that there is no spatial conflict between the twolife history types in the estuary. Not all down-stream migrant yearlings remain in the estuaryand some disperse to other nearshore areas adja-cent to the river mouth. Also, it appears that thelength of residence of yearling smolts in the estu-ary is relatively brief, as is their residence in shel-tered coastal waters in general (Healey 1980b,1982b, 1983; Levy and Northcote 1981).

Food Habits in Estuaries

The food habits of chinook in estuaries are docu-mented for a number of British Columbia estuaries(Dunford 1975; Goodman 1975; Birtwell 1978; Si-bert and Kask 1978; Fedorenko et al. 1979; Levy etal. 1979; Northcote et al. 1979; Healey 1980b, 1982b;Levy and Northcote 1981; Levings 1982), a numberof Oregon estuaries (Reimers et al. 1978; Myers1980; Bottom 1984), and the Sacramento-San Joa-quin River estuary (Sasaki 1966; Kjelson et al.1982). Diets vary considerably from estuary toestuary and from place to place within an estuary.Dunford (1975), Northcote et al. (1979), Levy et al.(1979), and Levy and Northcote (1981) reportedthat chironomid larvae and pupae were the mostimportant diet items of ocean-type chinook in tidalchannels throughout the Fraser River marshes. Ofsecondary importance were Daphnia, Eogammarus,Corophium, and Neomysis. Diets of chinook from the

north arm of the Fraser River tended to be morerestricted, with greater emphasis on Neomysis. Chi-nook captured in the main river channels or overRoberts and Sturgeon banks at the delta front werelarger than those found in tidal channels; and juve-nile herring, sticklebacks, and other small fish, aswell as cumaceans, insects, and Neomysis were im-portant in their diet (Dunford 1975; Goodman1975; Northcote et al. 1979; Levy and Northcote1981; Levings 1982). Yearling smolts (stream-type)were larger still and fed heavily on chum fry, aswell as on Eogammarus, Neomysis, Corophium, andchironomids. Northcote et al. (1979) noted thatchinook fry less than 50 mm long in the FraserRiver marshes demonstrated an intermediate tolong path food web dominated by benthic detriti-vores, but with significant input from other path-ways such as herbiverous zooplankton andterrestrial insects. Larger chinook characteristi-cally had long path food webs with multiple die-tary compartments, including benthic detritivores,zooplankton, and fish.

Chinook fry (ocean-type) in the Nitinat Riverestuary fed mainly on adult insects, gammarids,crab larvae, and Cladocera, whereas those in theintertidal zone of the Nanaimo River estuary fedmainly on crab larvae, mysids, adult insects, andharpacticoid copepods. As was observed in theFraser River estuary, chinook that had moved intodeeper water in the Nanaimo River estuary beganto feed heavily on fish (Healey 1980b). Insects andamphipods were the preferred diet items of ocean-type chinook in the Somass River estuary, exceptin some heavily industrialized areas where a pau-city of other fauna resulted in chinook feeding onoligochaetes (Birtwell 1978).

Research on a number of Oregon estuaries hasshown that benthic amphipods, particularly Coro-phium spp., and aquatic insects are the dominantfood of juvenile fall chinook (Reimers et al. 1978;Bottom 1984). By contrast, Myers (1980) found thatfishes, especially Engraulidae, Osmeridae, andClupeidae, dominated the diet of juvenile wildchinook in the Yaquina estuary, Oregon.

Insects and Crustacea dominated the diet ofyoung chinook in the Sacramento-San JoaquinRiver delta (Sasaki 1966; Kjelson et al. 1982). Sasaki(1966) found that chironomid larvae were impor-tant in the upstream areas of the delta, whereasNeomysis and Corophium were important in the

346


lower delta. By contrast, Kjelson et al. (1982) foundCladocera, Copepoda, and Diptera were the mostimportant foods in the upper and lower estuaryand that amphipods and mysids constituted only asmall percentage of the diet.

Seasonal changes in diet are typical, and pre-sumably reflect seasonal changes in the abundanceof prey organisms. In the Nitinat River estuary,chinook diet was dominated by insects and Eogam-marus during April. Towards the end of April, how-ever, larval herring became important and re-mained so until the end of June. About mid-Maycrab larvae began to comprise an important part ofthe diet and continued to be important until theend of June. Cladocera were not important untillate May and were a significant diet item until lateJune (Figure 15).

10090

80

70-

60

50-

40 '

30 '

20 -

10 -

0

I

Eogammarus

1 15April

30 15May

31 15June

30

I *t

FIGURE 15Seasonal changes in the diet of juvenile chinook in Nit-inat Lake. At top of figure, O refers to other diet items.

(From Healey 1982b)

Most studies of chinook diet are based on sam-ples taken during daylight hours. Studies of theSixes River estuary, however, indicate that juvenilechinook feed actively at night (Bottom 1984). Theirprincipal foods in this estuary were the amphip-ods Corophium and Eogammarus, which are inactiveduring the day but migrate into the water columnat night.

Sibert and Kask (1978) compared the diets ofchinook among the Fraser, Cowichan, Nanaimo,and Campbell River estuaries, and found little cor-relation among the diets of fish from comparablephysiographic regions of different estuaries oramong the diets of fish from different physiogra-

phic regions of the same estuary. The same, orclosely related organisms did tend to show upconsistently in the diets, however, although theirnumerical contribution to diet varied widely be-tween estuaries.

Chinook generally cohabit with other salmonidsin estuaries, in particular with chum salmon. Al-though they often eat the same organisms, thecorrelation between their diets was weak in theFraser and Nanaimo River estuaries (Dunford 1975;Sibert and Kask 1978). The diet of chinook alsocorrelated poorly with the diet of cohabiting cohoin the Nanaimo River estuary (Sibert and Kask1978). The diet of chinook was, in fact, more similarto the diet of some non-salmonids in these estuar-ies (e.g., herring, Clupea pallasi; stickleback, Gaster-osteus aculeatus; shiner perch, Cymatogaster aggregata;and sand lance, Ammodytes hexapterus) (Sibert andKask 1978). Myers (1980), on the other hand, foundthat there was often considerable overlap in thediets of wild chinook and hatchery coho in Ya-quina Bay, Oregon.

In general, chinook appear to be opportunisticfeeders in estuaries. Comparison of their diet withthat of other similar-sized salmonids in the samearea suggests that chinook prefer slightly largerorganisms and that larval and adult insects, as wellas amphipods of various sorts, are their preferredprey in the intertidal regions of most estuaries.Dunford (1975) found that chinook were more effi-cient predators of chironomid larvae than chumand were able to capture and eat Neomysis thatchum could not capture. In a mixed assemblage ofCladocera, chironomid larvae, and Neomysis, chi-nook fed preferentially on chironomids and Clad-ocera. As the chinook become larger and begin toinhabit deeper water, their dietary preference ap-pears to shift to larval and juvenile fishes.

Growth in Estuaries

In the Columbia River estuary, young fall chinook(ocean-type) increased in length from 38 mm inApril to 113 mm in October, an average daily in-crease of 0.44 mm, assuming that these fish werefrom the same cohort (Rich 1920). Chinook in theSacramento River estuary increased in length by0.48 mm/d between March and July, again assum-ing the same cohort was being sampled (Rich1920). The probability that these samples were

347


from the same cohort is rather small, however, asyoung chinook probably do not spend such a longtime in the estuary, and their departure from estu-aries appears to be size-related. Also, the "alterna-tion" of different juvenile life history types in theestuary makes it unlikely that these growth esti-mates relate to a single behavioural type. The fishcaptured later are likely to have reared in the riverfor some weeks or months. These estimates ofgrowth rate in estuaries must, therefore, be consid-ered minimum estimates. Kjelson et al. (1982) esti-mated growth rates of fry tagged with coded-wiretags in the Sacramento-San Joaquin River estuaryto be 0.86 mm/d in 1980 and 0.53 mm/d in 1981.These rates of growth are faster than those basedon unmarked fry.

During 1978 and 1979, young-of-the-year chi-nook in the Fraser River estuary increased from 40to 60-68 mm fork length between March and July(Levy and Northcote 1981). Their increase inlength was slow until mid-May, rapid from mid-May until the end of June, and then slow againduring July. From mid-May until the end of Junethey increased an average of 0.56 mm/d in 1978 and0.39 mm/d in 1979.

The slow increase in length of chinook fry in theFraser River estuary from March to mid-May wasalmost certainly due to the continued addition ofrecently emerged fry to the estuary population,since the main period of downstream migration offry was from mid-March to mid-May in both years(Levy and Northcote 1981). The apparent slowgrowth of chinook in July may have been due tomovement of larger chinook away from thebeaches and into deeper water. Levy and North-cote (1982) observed that young chinook capturedby purse seine in deep water in an old river chan-nel on the Fraser River foreshore were significantlylarger than those captured by beach seine alongthe margins of the channel. As noted earlier, Kjel-son et al. (1982) observed a similar distribution ofchinook in relation to size in the Sacramento-SanJoaquin River estuary.

Reimers (1971) observed the growth rate ofyoung fall chinook in the Sixes River estuary bothby sampling the general population and by ob-serving changes in the length of marked cohorts.Both sources of data gave similar results. Growthin the estuary from late April to early June wasrapid. During this period, chinook increased in

average length from 48 to 79 mm, an average of 0.9mm/d. From June to August, growth in the estuarywas poor, and the fish increased only 6 mm or 0.07mm/d during this period. From September to No-vember, growth was again rapid, averaging about0.5 mm/d. Reimers (1971) hypothesized that poorgrowth during the June-to-August period was dueto the large population of chinook in the estuary atthis time, and that the increased rate of growth inSeptember to November resulted from both a re-duction in population size and better utilization ofthe whole estuary. By means of otolith microstruc-ture, Neilson et al. (1985) confirmed that individualfish show a rapid growth until June in the SixesRiver estuary, after which their growth sloweddown. Neilson et al. (1985) suggested that thedecline in growth rate after June resulted from acombination of high temperatures in the estuarythat reduced growth efficiency and competition forfood. Neilson and Geen (1986) also noted thatchinook that entered the estuary at a large sizeremained large relative to members of the samecohort throughout their first year of ocean life.

The rate of growth of marked fry in the NanaimoRiver estuary averaged 1.32 mm/d (4%-5% bodyweight/d) (Figure 16). From average length data forthe general population, however, the rate of in-crease in length during April to June was onlyabout 0.5 mm/d (Healey 1980b), or less than halfthe rate indicated by marked fish. Other estuarieson the east coast of Vancouver Island showed ratesof growth based on average length data from0.22mm/d in the Cowichan River estuary to 0.61mm/d in the Courtenay River estuary (M.C. Hea-ley, unpublished data) and 0.46-0.55 mm/d in theCampbell River estuary (Levings et al. 1986). Pre-sumably, these are underestimates of the truegrowth rate in these estuaries. Average length inthe general population of chinook in the NitinatRiver estuary increased about 0.33 mm/d duringthe years 1975-77 (Fedorenko et al. 1979). In 1979,however, fish from a marked cohort increasedabout 0.62 mm/d (3% body weight/d) (Figure 16).Measurements based on increases in averagelength of the general population in estuaries,therefore, appear to underestimate the true growthrate by a factor of about two. The exception to thisis Reimers' (1971) data for the Sixes River. Variation .in growth rate among estuaries is also on the orderof two times, as is evident from data of the Na-

348


!

15

10-

5-

Nanaimo estuary

3 5-

I

Nitinat estuary

= 0.547ta038

15 21 27 3April

9 15 21 27May

FIGURE 16Growth of marked cohorts of chinook fry in the Na-

naimo and Nitinat River estuaries. Closed and open cir-cles refer to different tagged groups of fry. Regressionsare the regression of log weight (W) on days (t) follow-

ing marking.

naimo and Nitinat River estuaries. The availableevidence suggests that variation in growth ratesbetween estuaries and between years within anestuary is correlated with food supply (Healey1982b; Neilson et al. 1985).

Considering the limited information available onrates of growth of young chinook, particularlytheir growth in fresh water, no substantial compar-isons between freshwater and estuarine growthcan be made. Kjelson et al. (1982), however, dem-onstrated by means of fry tagged both in the riverand the estuary that fry grew more rapidly in theestuary. In many other instances, growth of chi-nook that rear in the river during their first springappears slower than growth of those that migrateto the estuary. Reimers (1971) and Rich (1920) in-ferred this from the closer spacing of circuli on thescales of fish that had reared in the river, andsuggested that circulus spacing could be used todistinguish between fish that reared to smolt sizem the river and those that reared in the estuary.

Reimers (1971) proposed that up to five, andSchluchter and Lichatowich (1977) proposed thatup to seven, different juvenile life history patternsinvolving different periods of river and estuarineresidence could be distinguished from patterns ofscale growth in the Sixes River and the RogueRiver, Oregon. These patterns of scale growthwere presumed to reflect differences in growthrate between fry residing in the river and in theestuary. In other instances, there appears to be nosubstantial difference in growth between fish thatrear to fingerling smolt size in the river and fishthat rear in the estuary (Vancouver Island estuar-ies, M.C. Healey 1980b, 1982b; M.C. Healey, un-published data). Fish from these systems thatreared in the estuary and fish that reared in theriver did not differ in the spacing of circuli on theirscales.

Apparently egg size may influence rate ofgrowth, at least in hatchery fry with abundantfood. Fowler (1972) found that fry from large eggswere larger at hatching and maintained that ad-vantage over 10-12 weeks after hatching. Rom-bough (1985) found that larger eggs producedalevins with greater maximum weight, but alsothat it took the alevins from larger eggs longer toreach their maximum weight after fertilization,compared with alevins from smaller eggs. Suchvariations in growth rate could have importantimplications for mortality if the larger fry outgrowpotential predators more quickly. In Fowler's(1972) experiments, however, the more rapidlygrowing fry from larger eggs also had a highermortality rate from unspecified causes. Thus, theadvantage of faster growth may be offset by otherunknown disadvantages associated with largereggs.

Healey (1982b) estimated the production andfood requirements of chinook and other salmonspecies in the Nanaimo and Nitinat River estuaries.Chinook were second to chum in production in theinner Nanaimo River estuary, producing about 200kg during the spring and early summer, comparedwith 1,750 kg for chum. In the Nitinat River estu-ary, however, chinook dominated juvenile salmonproduction, contributing 774 kg of an estimatedtotal production of 1,137 kg. These levels of pro-duction are small relative to the area of the twoestuaries, being about 0.031 g/m2 in the NanaimoRiver estuary and 0.025 g/m2 in the Nitinat River

349


estuary. Food resources required to support thisproduction were 0.093 g/m2 in the Nanaimo Riverestuary, consisting of mainly benthic foods butincluding about 30% insects; and 0.078 g/m2 in theNitinat River estuary, of which about 40% werebenthic organisms, 40% insects, and 20% plankton(Healey 1982b). The standing crops of food organ-isms in these estuaries appeared sufficient to sup-port considerably greater production. Growth ofchinook in the Nitinat River estuary was, however,considerably less than growth in the NanaimoRiver estuary, and this was correlated with lowerstanding crops of food organisms in the NitinatRiver estuary. Also, there was a positive correlationbetween abundance and growth rate in the Na-naimo River estuary and the fullness of chinookstomachs, suggesting that growth and productionwere positively related to food supply in this estu-ary (Healey 1982b).

Differences among stream- and ocean-type racesare evident in their utilization of estuarine habi-tats. Ocean-type fish make extensive use of estua-rine habitat, whereas stream-type fish spend littletime in the estuary of their natal stream. Amongthe ocean-type races there is a further dichotomy

between those that migrate to the estuary as fry inMarch or April and remain there until about Juneand those that migrate as fingerlings in May orJune and remain until August or later. This dichot-omy becomes blurred in Oregon, where a varietyof juvenile behaviour patterns has been identified(Reimers 1971; Schluchter and Lichatowich 1977;Nicholas and Hankin 1988) but is apparent againfurther south in the Sacramento-San Joaquin Riverestuary (Kjelson et al. 1981, 1982). The period ofestuarine occupancy by ocean-type chinook variesregionally, being greatest in the open coast estuar-ies of Washington to California, and least in thesheltered coastal estuaries of British Columbia.Estuaries apparently provide a rich feeding habitatfor the smaller fry and fingerling migrants, andgrowth in estuarine habitats, although variable, isrelatively rapid. Why stream-type chinook do notspend more time in estuarine habitats is not imme-diately apparent. The tendency for chinook to be-come piscivorous, however, as soon as their sizepermits, and the particular oceanic migratory pat-tern of stream-type chinook (Healey 1983), maypreclude a longer estuarine residence.

OCEAN LIFE

The distribution, seasonal abundance, and migra-tory behaviour of first ocean-year chinook salmonare described by Healey (1976, 1980a, 1980b) forthe Strait of Georgia (British Columbia), by Miller-et al. (1983) and Fisher et al. (1983, 1984) for thewaters off the coasts of Washington and Oregon,and by Hartt (1980), Healey (1983), and Hartt andDell (1986) for coastal and offshore waters from theColumbia River to the Bering Sea. The samplingdevice in all these studies was a small-mesh purseseine, so that the results pertain only to the surfacewaters (about 20 m deep).

Information on the distribution and migratoryhabits of older chinook derive from a wide varietyof sources. Sampling of commercial troll catchesfrom southeastern Alaska to California has pro-vided information on the general coastal distribu-

tion of chinook and their distribution by age andrace (Fry and Hughes 1951; Parker and Kirkness1956; Milne 1964; Ball and Godfrey 1968a, 1968b,1969,1970; Wright et al. 1972; Argue and Marshall1976; Nicholas and Hankin 1988). Sampling thecatch of the Japanese mothership and land basedgillnet fisheries has provided information on thedistribution of chinook in the western North Pa-cific Ocean. (Geographic distribution of these fish-eries is shown in Figures 19 and 20.)

Research cruises by Canadian, u.s., and Japanesevessels have provided more extensive informationon chinook salmon distribution throughout theNorth Pacific Ocean. Gear employed in researchcruises included floating longlines and gillnets(Major et al. 1978). As with the seine sampling forfirst ocean-year chinook, samples taken on the

350


high seas by gillnet and longline relate mainly tothe surface waters. Chinook are probably under-represented in these samples because of their ten-dency to be distributed deeper in the water col-umn than the other species of Pacific salmon(Milne 1955; Taylor 1969; Argue 1970). Despite po-tential sampling biases and difficulties in compar-ing among gear types, the observations of allauthors are consistent in that they indicate a majordifference in the distribution and migratory behav-iour of stream- and ocean-type chinook duringtheir ocean life.

More specific information on the distribution ofparticular stocks or groups of stocks of chinook isderived from tagging and recapture. Immature andmaturing salmon have been tagged in both coastaland high-seas fisheries, and tagged fish have beenrecaptured from a few weeks to several years aftertagging (Kondo et al. 1965; Hartt 1966; Godfrey1968b; Aro et al. 1971; Aro 1972, 1974). Compara-tively few fish tagged in the ocean have been re-captured after a sufficient length of time, however,to permit a precise determination of migratoryroutes and timing. Considerably more detailedinformation is available from tagging of hatchery-produced smolts and noting their recovery incoastal troll and net fisheries (e.g., Cleaver 1969;Wahle and Vreeland 1977; Dahlberg 1982; Werthei-mer and Dahlberg 1983,1984; Dahlberg and Fowler1985; Dahlberg et al. 1986; Nicholas and Hankin1988). A wealth of this kind of information hasbeen produced over the past decade as a result ofthe extensive application of coded-wire tagging inhatchery evaluation (Jefferts et al. 1963). Most ofthese data, however, remain completely unana-lysed. Furthermore, the question of whether thebehaviour of hatchery fish is comparable with thatof wild stocks remains unresolved, although Hea-ley and Groot (1987) found that wild and hatcherychinook from the east and west coasts of Van-couver Island had similar oceanic distributions.

Fish in their First Ocean Year

Observations in the Strait of Georgia indicatedthat, in these sheltered waters, young chinookbegan to disperse seaward from their natal estuaryshortly after completing their downstream migra-tion. Apparently, the first to do so were the stream-type smolts. In the Nanaimo River estuary, stream-

type smolts were rarely captured in the inner estu-ary but were common in the outer estuary and inother nearshore sampling stations in the vicinity ofNanaimo in June and July of 1975 and 1976, butwere rare after July (Healey 1980b) (Figure 17).

Apr May Jun Jul Aug ' Sep ' Oct ' Nov '

FIGURE 17Seasonal occurrence of 0.0,1.0, and 0.1-aged chinook inpurse seine catches near Nanaimo, BC, during 1975-77.

(From Healey 1980a)

Sampling from May to October, 1976, throughoutthe Gulf Islands (along the southeast coast of Van-couver Island) confirmed this seasonal pattern ofabundance of stream-type smolts outside the Na-naimo area (Healey 1980a). As stream-type juve-niles declined in abundance in the shelteredwaters of the Strait of Georgia, ocean-type juve-niles increased in abundance. In the Nanaimo area,ocean-type juveniles were about five times asabundant in purse seine samples as stream-typejuveniles, and ocean-type fish remained abundantfrom July until November (Figure 17). Seine netcatches were lower during the winter, but somechinook were present in surface waters and avail-able to the seine throughout the year (Healey1980b). In the Gulf Islands region of the Strait ofGeorgia, young chinook were rather constant inabundance from May to October, 1976. Samplestaken throughout the Strait of Georgia in Augustto September, 1975 and 1976, revealed that chinookwere most abundant in the region of the FraserRiver plume and less abundant elsewhere. Thisdistribution was much more pronounced in 1975than 1976 (Healey 1980a).

The general conclusion from these observationswas that stream-type chinook in their first oceanyear are common in the surface waters of the Strait

351


of Georgia only in the spring and early summer,but that ocean-type chinook in their first oceanyear are abundant throughout the summer andautumn, and some remain in the surface waters ofthe strait throughout their first ocean year. Therewas no indication of an outmigration of youngocean-type chinook from the Strait of Georgia be-fore November in 1976. Seine samples taken by theSnoquomish Indian band in Puget Sound in Octo-ber and November, 1976, indicated an abundanceof chinook in their first ocean year that was similarto that in the Strait of Georgia (Hartt and Dell1986).

The estuaries along the open coasts of Washing-ton, Oregon, and California afford the only shel-tered water habitat in these regions. Here youngocean-type chinook remain in estuaries consider-ably longer than was observed in British Columbiaestuaries (Reimers 1971; Myers 1980; Myers andHorton 1982) (Figure 14), although stream- andocean-type chinook also occur in coastal watersoutside the estuaries during the summer (Miller etal. 1983; Fisher et al. 1983, 1984). These observa-tions suggest that the affinity of young ocean-typechinook for sheltered waters is general throughoutthe range of chinook, but that there is also sbmeoffshore movement of these fish.

Off the coasts of Washington and Oregon, Milleret al. (1983) found young chinook in abundanceduring sampling from 27 May to 7 June, 1980.Catches were greatest off the Columbia Rivermouth, north of this river, and in seine sets madewithin 20 km of shore. Catches of chinook in theirfirst ocean year declined in July, presumably owingto high surface water temperatures at this time,and catches increased again in August-Septemberwhen surface temperatures were lower. In the Au-gust-September sampling, young chinook weredistributed about equally north and south of theColumbia River mouth but were still most abun-dant close to shore. Significantly, young springchinook (stream-type) from Columbia River hatch-eries were present only in the May-June samples,and then only in the northernmost sampling tran-sects, indicating a rapid migration of these fishnorth from the river mouth. In 1982 and 1983Fisher et al. (1983,1984) sampled an area similar tothat sampled by Miller et al. (1983) in 1980 andcaptured similar numbers of young chinook. Dur-ing 1982, chinook were most common in samples

352

taken in May and June but were rare in September.During 1983, catches were highest in May andSeptember but were low in June. There was noclear trend in abundance from south to north inthese samples; sometimes young chinook weremore abundant south of the Columbia River andsometimes north. Nor was there any clear evidencefrom the samples of Fisher et al. (1983, 1984) thatyoung chinook were more abundant close to thecoast.

Miller et al. (1983) compared catches in seine setsheld open towards the north or south. InMay-June, 80% of chinook were captured in setsheld open towards the south, indicating a signifi-cant northward dispersal at this time. During Julyand August-September sampling, however, thedirection of the set had no effect on catch. Thesedata indicated a northward dispersal of ocean-typechinook immediately on entry into the sea, butalso indicated that movements were not directedthroughout most of the summer. This is reminis-cent of the summer residency of young ocean-typechinook in the Strait of Georgia.

Hartt (1980) and Hartt and Dell (1986) reportedon an extensive series of samples taken by purseseine along the coast of North America from theColumbia River to Bristol Bay, throughout the Gulfof Alaska, along the Aleutian Islands chain, andinto the central Bering Sea. Hartt and his co-work-ers made a total of 3,073 sets during the 15-yearperiod, 1956-70, mostly during the summermonths, but extending from April to October. Thesampling was concentrated in coastal waters (Fig-ure 18), with offshore seining principally in thespring and early summer. Most young chinookwere caught near the coast, and catches overallwere greatest from the Columbia River to south-eastern Alaska. Smaller catches were made in cen-tral Alaska, along the Aleutian chain to AdakIsland, and in Bristol Bay. No chinook in their firstocean year were captured in the central Gulf ofAlaska or in the central Bering Sea. Greatestcatches were in the June-August period, and therewas an indication, although catches in all areaswere small, that chinook appeared in the catcheslater in the north than in the south.

An important feature of Hartt's catches wasthat, by far, the majority of chinook were stream-type (245/253 or 97%)(Healey 1983). Of the eightocean-type fish captured, six were captured off


7 : 14 ; 23 120.l: 15: 2 :. I2T- 'T39

I62./.22 i 56 U9-: * ; 20 -: 12 ! 29! 5 ! 6 i 4 :, . . . w . , . . . . . . . . . . . . . _^ ^

5 ; - : 19-I 10 i 5- i 13; :. 6

2. . - . . I . . ;

70°N

60°N

50°N

40°N

120°E 140°E 160°E 180° 160°W 140°W 120°W

FIGURE 18Numbers of small-mesh purse seine sets made by Hartt and his co-workers (Hartt and Dell 1986) between

1956 and 1970 in 2° x 5° geographic areas. These areas are the standard statistical recording areas of theInternational North Pacific Fisheries Commission (INPFC).

Cape Flattery and two off southeastern Alaska.Thus, it appeared that juvenile chinook in theirfirst ocean year, living along the outer coast, atleast from Cape Flattery and northward, were pre-dominantly stream-type fish, whereas those insheltered inside waters were predominantlyocean-type. The samples collected by Miller et al.(1983) and Fisher et al. (1983, 1984) suggest thatocean-type chinook were more common along theopen coast south of Cape Flattery.

Fish in their Second and Subsequent Ocean Years

With the exception of organized coastal fisheriesand the Japanese mothership and landbased fish-eries (Figure 19 and 20), comparatively few chi-nook have been captured in the open ocean, butthe distribution of captures has been very wide-spread (Figure 19). Chinook almost certainly occurfurther offshore south of 46°N on the North Ameri-can coast than is currently indicated, probably

even south of 40°N, considering the continuedhigh production of chinook from Californian riv-ers. Sampling, however, has not been adequate todemonstrate this.

Manzer et al. (1965) reported on the freshwaterages of 847 chinook captured in the Japanesemothership fishery between 175°E and 175°W.These fish were all stream-type. The majority wereprobably of Asian and Alaskan origin, althoughrecent returns of coded-wire tagged chinookshowed that fish from British Columbia, Washing-ton, and Oregon migrate as far west as 160°W-175°W longitude (Dahlberg 1982; Wertheimer andDahlberg 1983, 1984; Dahlberg and Fowler 1985;Dahlberg et al. 1986). It is possible, therefore, thatsome of the fish in the sample reported by Manzeret al. (1965) originated from populations south ofthe Alaskan Panhandle. If so, none was an ocean-type fish. More recent sampling of the Japanesehigh-seas fisheries (Knudsen et al. 1983; Myers etal. 1984) revealed a similarly low contribution of

f

353


70°N

60°N

50°N

40°N

120°E 140°E 160°E 180° 160°W 140°W 120°W

Mothershipfishing area1978-onward

70°N

60°N

FIGURE 19Known distribution of chinook in the North Pacific Ocean and Bering Sea based on captures in high-seas and coastalsampling (shaded 2° x 5° INPFC statistical areas). Areas in which sampling occurred but no chinook were captured are

shown (0). The Japanese mothership and landbased fishing areas prior to 1978 are shown.

ocean-type fish, although there are no absolutecriteria for distinguishing age 0. and age 1. fish,and there was some disagreement between Japa-nese and u.s. ageing of the fish reported by Knud-sen et al. (1983). The u.s. ageing team found someocean-type fish (68/2779) and the Japanese team,none.

Further east, ocean-type fish do make a contri-bution to the high-seas population. Of 80 chinookcaptured by Canadian research vessels fishingwith longlines north of 45°N and east of 170°Wduring 1961-67, 52 (65%) were stream-type and 26(35%) were ocean-type (Healey 1983). These fishwere almost certainly of North American origin,and were probably mainly from southeasternAlaska and rivers to the south. Since stream-typechinook appear to comprise no more than 25% ofall spawning populations from the SacramentoRiver to southeastern Alaska, the percentage ofstream-type fish in the Canadian research vesselcatch on the high seas was significantly greaterthan expected, if stream- and ocean-type chinookhave similar ocean distributions (X2 = 68.3; p <.001). The high percentage of stream-type fish in

50°N

40°N

140°E 150°E 160°E 170°E 180° 170°W

FIGURE 20Japanese mothership and landbased fishing areas

following renegotiation of agreements in 1978, whichwere in effect until 1986

354


this sample was not a consequence of the fishbeing captured mainly in the northern Gulf ofAlaska, as the catch was equally distributed northand south of 49°N. Furthermore, of the 28 ocean-type fish captured, 14 were captured north of49°N, and 14 south of this latitude. Thus, there wasno evidence of a higher proportion of ocean-typefish in the southern half of the sampling area, aswould be expected if ocean-type fish were welldistributed offshore, but they were concentratedin more southerly areas, in keeping with the distri-bution of their spawning populations (Healey1983). It appears, therefore, that stream-type fishconstitute a high proportion of the high-seas pop-ulation regardless of latitude, although the propor-tion is lower in the eastern than in the central andwestern North Pacific Ocean.

Chinook are abundant in coastal waters alongthe coast of North America from southeasternAlaska to California throughout their ocean lives.In these coastal waters, the representation ofstream- and ocean-type races is the opposite ofthat observed on the high seas, and ocean-typefish predominate (Healey 1983).

The proportion of stream-type chinook in thecommercial troll fisheries of Alaska, British Colum-bia, and Washington in recent years has rangedfrom 3% to 4% in the Strait of Georgia to 25% in theQueen Charlotte Islands area of British Columbia(Table 5). The proportion off Washington and offsoutheastern Alaska was similar at 15%-16%. In

recent years, therefore, stream-type fish havemade up a relatively small proportion of the oceantroll catch, generally less than 20%, and signifi-cantly less than one would expect from the propor-tion of stream-type fish in the regional spawningpopulations.

There is evidence that stream-type chinook may,historically, have constituted a greater percentageof the coastal troll catch than they do at present. Inthe Strait of Georgia, stream-type fish were 28% ofthe catch during 1911-20 and declined to 3%-4%by 1961-70 (Table 5). Off the west coast of Van-couver Island, stream-type fish were 20% of thecatch during 1921-30, declined to 3.9% during1961-70, and increased to 9.0% during 1981-85(Table 5). Off Washington and Oregon, stream-type chinook constituted more than 20% of thecatch before 1950 but only 15% of the catch after1960 (Table 5). These changes must be interpretedcautiously, however, as scale interpretation andfishing patterns may have changed over the years.For example, stream-type fish were a high percent-age of the recent net catch in major river estuariesof British Columbia (Table 5) and, as noted earlier,maturing stream-type chinook return to the estu-ary of their spawning river early in the year. Ifhistoric troll fisheries were nearer river mouths, orwere concentrated earlier in the year, then theircatch of stream-type fish might naturally havebeen higher. On the other hand, habitat modifica-tion, particularly damming of the Columbia and

TABLESStream-type chinook (as a percentage) in coastal troll catch and river mouth gillnet catch

during several decades

Fishery

S.B. Alaska trollNorth BC trollCentral BC trollVan. Is. trollGeorgia St. trollWash./Oreg. trollPraser R. netSkeena R. netNassR.net

Datasource*

122223444

1911-20 1921-30

23.0

20.028.0 17.5

22.0

1941-50 1951-60

15.75.6

12.4 10.96.5

28.4 34.4

1961-70

20.69.33.93.4

15.342.848.146.2

1971-80 1981-85

25.012.09.0

12.828.1

Notes: "1 Parker & Kirkness (1956)2 Milne (1964); Ball & Godfrey (1968a, 1968b, 1969,1970); Healey (1986)3 Wright et al. (1972); Rich (1925); Van Hyning (1951)'4 Godfrey (1968a); Ball & Godfrey (1968a, 1968b); Fraser et al. (1982); Ginetz (1976)

355


Sacramento rivers, but also water extraction forirrigation and hydraulic mining (Kjelson et al. 1981,1982), have probably had a much greater effect onstream-type chinook than ocean-type, owing totheir longer upriver migrations and longer riverresidence, both as adults and as young. It is, there-fore, quite conceivable that stream-type chinookwere relatively more abundant a few decades ago.The possibility that the changes in percentage ofstream-type fish in the troll catches represent a realreduction in the numbers of stream-type fish isfurther supported by the observation that springand summer spawning runs (mainly stream-typefish) have declined drastically in several rivers inthe southern half of the Chinook's range (Snyder1931; Rich 1942) and by the fact that the oldermaturing stream-type chinook are less able to sup-port an intensive fishery than younger maturingocean-type chinook (Hankin and Healey 1986).

The information on distribution of races indicatesthat stream-type chinook move offshore early intheir ocean life, whereas ocean-type chinook remainin sheltered coastal waters. Stream-type fish main-tain a more offshore distribution throughout theirocean life than do ocean-type. Although ocean-typefish are captured offshore in the eastern half of theNorth Pacific Ocean, they are much less commonthere than stream-type fish, whereas the reverse istrue close to the coast. Since only stream-type chi-nook occur in western Alaska and in Asia, it is notsurprising that catches in the western North PacificOcean are virtually all stream-type. Stream-typechinook are also common in river mouth fisheries ofBritish Columbia (Table 5) and in early seasoncatches in the Fraser and Columbia rivers (Figure 6).These observations suggest that maturing stream-type fish move rather quickly through the coastaltroll fisheries to the river estuaries and so are avail-able for only a relatively short time to the trollfisheries. Healey and Groot (1987) noted that matur-ing chinook returning to their native river travelledmore than 45 km/d, or close to their optimal cruisingspeed, on a direct course towards the river. Theyappear to remain in the river estuaries for sometime, however, and while there are vulnerable to theriver mouth gillnet fisheries.

Ocean Distribution in Relation to Temperature

High-seas catches of chinook by research vessels

generally have been too small to permit analysis ofchinook distribution in relation to ocean tempera-ture except for catches by Japanese research ves-sels in the western North Pacific Ocean. There,during 1962-70, chinook were encountered at tem-peratures ranging from 1° to 15°C (Major et al.1978). There was no evidence for a preferred tem-perature within this range, except that, in Apriland May, relatively fewer chinook were encoun-tered at temperatures below about 5°C. These datarefer, of course, to Asian and Alaskan stream-typechinook. This may mean that chinook are indiffer-ent to temperature over this range. The data maynot be a true reflection of temperature selection (orlack thereof) by chinook, however, as the samplesoffish and temperature measurements came onlyfrom surface waters. As will be demonstrated later,chinook are not concentrated at the surface. Sur-face catches may, therefore, not be representativeof temperature selection occurring at depth.

Vertical Distribution of Chinook

Information on the depth distribution of chinookcomes primarily from two sources, neither ofwhich distinguished between stream- and ocean-type fish. Sampling was also conducted only dur-ing daylight hours. Taylor (1969) reported thedepth of capture of immature chinook off the eastand west coasts of Vancouver Island during surveycruises for Pacific herring. Most samples came fromthe west coast of the island, and these will beemphasized here. Chinook were captured in trawlsfishing deeper than 60 fm (110 m), but most fishwere captured above 40 fm (73 m) (Table 6). Chi-nook were not concentrated near the surface butwere most abundant in the 30-40 fm (57-73 m)stratum. Chinook captured in trawls in the Straitof Georgia were more abundant near the surface,with most (51/53) captured between 10 and 20 fm(20-37 m) (Table 6). Argue (1970) reported a moreextensive set of data based on troll sampling inJuan de Fuca Strait. Argue's (1970) samples onlyextended to 55 m, and most chinook (54%) werecaptured in the 48-55 m stratum (Table 6). Possibly,catches would have been as large at even greaterdepth. Both young chinook (younger than .2) andmaturing chinook (older than .2) were capturedmore frequently at shallower depths than wereolder immature chinook (Table 6). There appeared

356


TABLE 6Depth distribution of chinook (as a percentage of catch) captured off the west (W) and east (E)

coasts of Vancouver Island during herring surveys (Taylor 1969), and in Juan de Fuca Straitduring a troll fishery investigation (Argue 1970)

Taylor 1969 Argue 1970*

Immature

fm (m)-194 N-53W E

All agesN-150

0.0+1.0N-50

0.1N-76

0.2N-71

Maturing

N=19

0-5 (0-9) 1023

Note: *Argue's data are segregated by age and maturity

to be a seasonal change in depth distribution, withthe average depth of capture dropping from 33 min June and July to 41 m in August-October (Figure21). Taylor (1969) did not indicate the freshwaterages of the fish captured. The fish in Argue's sam-ples, however, were virtually all ocean-type.

Ocean Distribution in Relation to Area of Origin

There are two sources of information on oceanicdistribution of chinook in relation to area of origin:tagging studies and analysis of scale patterns. Tag-ging studies include both the tagging of immatureand maturing chinook in the ocean with subse-quent recapture in the ocean or on the spawningbeds, and the tagging of hatchery smolts at thetime of seaward migration with subsequent recap-ture in the ocean. Scale pattern analysis attemptsto determine the area of origin of chinook capturedon the high seas from characteristic features oftheir scales. Neither source of information gives a

30 -i

35-

40-

45-

50

11

6-10

11-15

16-20

21-25

26-30

31-35

36-40

41-45

46-50

>50

(11-18) J 3 7 8 6

(20-27) ) 7 14 21 217 96

(29-37) J 17 21 29 10

(38-46) ] 16 32 16 257 2

(48-55) J 54 24 14 38

(57-64) )46

(66-73) J

(72-82) }6

(84-91) J

(>91) 10

17

44

18

0

10

Jim Sep OctJul Aug

FIGURE 21Seasonal changes in the average depth of capture of

chinook in Juan de Fuca Strait. Bars show two standarderrors around mean. (Data from Argue 1970)

357


very satisfactory picture of the ocean distributionof the many stocks of chinook salmon. The major-ity of ocean tagging of chinook has been conduc-ted along the Pacific coast of North America fromsoutheastern Alaska to California. Godfrey (1968b)summarized data from ocean tagging in this areaduring the period 1924-64. In addition, a few tagshave been recovered from chinook tagged duringhigh-seas fishing operations in the northern NorthPacific Ocean and the Bering Sea (Major et al. 1978;Knudsen et al. 1983). The ocean distribution ofchinook tagged as smolts and recaptured in coastalfisheries has been described by Cleaver (1969),Wahle and Vreeland (1977), Wahle et al. (1981), andHealey and Groot (1987):

A total of 21,566 chinook salmon were tagged incoastal waters during the 1924-64 period, and re-captures numbered 2,418 (Godfrey 1968b). Recov-eries of tags came from recreational and commer-cial fisheries and from spawning ground surveys.Because of unequal sampling and recovery effort,the tag returns provide only a general indication ofthe stock composition in various regions of thecoast. Some important generalizations are, how-ever, possible. Most recaptures were made either

in the area of tagging or to the south, except for thetagging off California, where it was only possibleto recover tags in the area of tagging or to thenorth (Figure 22). When tagging was conducted ininside sheltered waters, most recaptures camefrom the area of tagging or immediately adjacentareas. When tagging was conducted in more openwaters, however, such as off the west coast ofsoutheastern Alaska or off the west coast of Van-couver Island, recaptures were more widely dis-tributed geographically (Figure 22). Fish tagged offsoutheastern Alaska were recaptured as far southas Oregon coastal streams, whereas fish tagged offCalifornia were recaptured as far north as theWashington coast. These data suggest a northwarddispersal of juveniles along the coast, followed by asouthward homing migration of maturing adults.The more limited dispersal offish tagged in insidewaters is, however, perplexing. It may indicate thepresence of "resident" stocks that do not undergolong-distance migrations. It may also be a reflec-tion of a greater number of stream-type chinook inthe groups tagged in outside waters, coupled withthe apparent longer migrations of this race. Recentanalysis of tag returns from hatchery and wild

\, TAGGING

^ECOVERY^.

S.E. Alaska

S.E.

Alaska

Inside Outside

NorthernBC

Inside Outside

SouthernBC

Inside Outside

M n

Wash. Oreg. California

Northern BC

Southern BC

Washington

Oregon

California

FIGURE 22The distribution of recaptures from chinook tagged as immatures in coastal waters from southeastern Alaska toCalifornia. Tagging location in each diagram is marked (x). Proportion of recaptures in each region is indicatedby the width of the diagram at that point. Inside and outside tagging areas for some regions refer to inside or

outside coastal island chains.

358


chinook stocks on the east and west sides of Van-couver Island (Healey and Groot 1987) indicatethat "inside" ocean-type stocks have a more re-stricted migration than "outside" ocean-typestocks.

Coastal Oregon stocks of chinook show diverseocean migration patterns that do not correspondto the general northward migrating behaviour onother parts of the North American coast. Stocksthat spawn in rivers on the central and northernparts of the Oregon coast (from the Elk Rivernorth) show the typical northward migration asimmatures, and these stocks contribute to fisheriesfrom Oregon to Alaska. Stocks that spawn in riverson the southern part of the Oregon coast (from theRogue River south), by contrast, disperse mainlysouth and contribute to fisheries off Oregon andnorthern California. One stock, the Umpqua Riverspring-run chinook, disperse both north and southfrom their natal river and are harvested fromnorthern California to Alaska (Nicholas and Han-kin 1988).

The distribution of particular stocks by age isrevealed by returns of fish tagged from hatcheryreleases. Releases from Columbia River hatcheriesdispersed mainly north of the Columbia River,although some were also captured south of theriver mouth (Figure 23) (Cleaver 1969). Fish agedtwo years were caught mainly off the Washingtoncoast; fish aged three years, off Washington andsouthern British Columbia; fish aged four years, offWashington and southern British Columbia butwith a few returns from northern British ColumbiaRECOVERY

AREAAGE

0.1 0.2 0.3 0.4

S.E. Alaska

Northern BC

Southern BC

Washington

Oregon

California

FIGURE 23Distribution of recaptures by age for ocean-type

chinook from Columbia River hatcheries (solid lines)and Robertson Creek, BC (dashed lines). Proportion ofrecaptures in each region is indicated by the width of

the diagram.

and southeastern Alaska; and fish aged five years,mainly from southern British Columbia, but ex-tending into northern British Columbia and south-eastern Alaska (Figure 23). Chinook from theRobertson Creek Hatchery on the west coast ofVancouver Island also dispersed mainly north asthey became older, but there was somewhatgreater southward dispersal than was apparent forthe Columbia River fish. Robertson Creek fish wereclearly distributed further north than ColumbiaRiver fish (Figure 23). Robertson Creek fish, how-ever, also dispersed further north than Big Quali-cum River fish from eastern Vancouver Island(Healey and Groot 1987). It appears that there may,in some instances, be considerable differences inthe ocean distribution of stocks from the samegeographic area.

Wahle and Vreeland (1977) and Wahle et al.(1981) described the ocean distribution of fall(ocean-type) and spring (stream-type) chinookfrom various Columbia River hatcheries by meansof recaptures in coastal fisheries. Fall chinook wererecaptured mainly in British Columbia and Wash-ington fisheries and, secondarily, in ColumbiaRiver estuary fisheries. Except for the KalamaHatchery (Washington) release, there were no re-coveries of fall chinook from southeastern Alaskaand relatively few recoveries from south of theColumbia River (Figure 24). There were some inter-esting, and perhaps significant, differences in thecontribution of different hatchery stocks to thedifferent fisheries. For example, the proportionrecovered in British Columbia fisheries rangedfrom 12% to 50%, and the proportion recovered inthe Columbia River estuary ranged from 2% to 26%(Figure 24). Spring chinook, in general, had a widerdistribution than fall chinook. For example, up to20% were recaptured in southeastern Alaska, andall stocks contributed to the Alaskan catch. Morespring chinook tended to be caught south of theColumbia River and in the river estuary as well. Aswith the fall chinook stocks, there were interestingdifferences in the apparent ocean distributionamong spring chinook hatchery stocks.

To a large extent, the distribution of tag recover-ies of various stocks has been determined by thelocation of intensive fisheries for chinook. This isclearly revealed by the recent recovery of tags fromchinook captured incidentally in foreign trawl fish-eries within the United States 200-mile conserva-

359


OCEAN-TYPE CHINOOK

HATCHERY

RECOVERYAREA

.0.2COO

co

_ojco

_c

o_oLU

O.0XO one

0)-a8WCOO

S

O

^

0)0)0

COOom

sIE

_g) £

I] CO

AlaskaBritish ColumbiaWashingtonColumbia RiverOregonCalifornia

STREAM-TYPE CHINOCK

HATCHERY

RECOVERYAREA

co &«CO

CO

0)Q

egla

.Q.O=>0

CD_^co

CO

AlaskaBritish ColumbiaWashingtonColumbia RiverOregonCalifornia

FIGURE 24Distribution of recaptures of chinook from ten Columbia River hatcheries producing ocean-type chinook and

six hatcheries producing stream-type chinook. Proportion of recaptures in each region is indicated by thewidth of the diagram.

tion zone in Alaska (Dahlberg 1982; Wertheimerand Dahlberg 1983, 1984; Dahlberg and Fowler1985; Dahlberg et al. 1986). Approximately 60,000chinook have been examined by United Statesobservers aboard foreign trawlers fishing mainlysouth of the Alaska Peninsula and in the southeast-ern Bering Sea, and these have yielded a total of244 tagged chinook that originated from riversranging from central Alaska to California. The dis-tribution of recaptures extends from northernsoutheastern Alaska westward through centralAlaska, along the Alaska Peninsula and AleutianIslands to about 175°W, and northward into thesoutheastern Bering Sea (Figure 25).

Of the recaptured chinook, 70 were from stocksin southeastern and central Alaska, 75 from stocksin British Columbia, 43 from stocks in Washington,54 from stocks in Oregon, and one each fromstocks in California and Idaho. The recaptures ofAlaskan stream-type chinook in these coastal trawlfisheries provide evidence that at least some fishfrom northern stream-type stocks remain close tothe coast for one to two years following their sea-ward migration, similar to the behaviour of ocean-type chinook from further south. The recaptures ofchinook from stocks in British Columbia and fur-ther south represent definite range extensions forthese stocks. Most of the recaptures were of ocean-

360


,60°N^ /? ALASKA

oOVi

BC=52WA=28OR=28CA=1

AK=5BC=4OR=4 AK=4

BC=3WA=1OR=2

W7044 W6544 W6044 W5544 W4544 W4O44W3544 W3044

170°W 160°W 140°W 130°W150°W

FIGURE 25Numbers of tagged chinook from different coastal origins recovered as by-catch in the foreign trawl fleet fishing

within the u.s. 200-mile zone off Alaska. Catches are shown within INPFC statistical areas.

type chinook, and the previous maximum extentof their range along the coast had been determinedfrom recaptures in the troll fisheries of southeast-ern Alaska. These new recaptures extend the rangeof these stocks westward along the coast to nearly175°W and, for British Columbian and Oregonchinook, northward into the Bering Sea (Figure25).

Although the recaptures from incidental catch ofchinook in foreign trawl fisheries indicate signifi-cant range extensions for numerous stocks of chi-nook, the incidence of tagged chinook amongthose landed is very low (0.4% overall, 0.3% if onlyocean-type stocks are considered), indicating thatonly a small proportion of these stocks travels sofar west and north. Furthermore, these recaptures,together with a single return to the ColumbiaRiver from chinook tagged south of Adak Island,appear to define the probable western limit ofstocks from south of the Alaskan Panhandle atabout 180°, because examination of almost 53,000chinook captured in the Japanese mothershipfishery operating within the United States 200-mile conservation zone west of 175°E produced notagged chinook. Nor did examination of over

20,000 chinook captured by Japanese research ves-sels fishing in the northern North Pacific Oceanand Bering Sea produce any tagged chinook (Dahl-berg 1982; Wertheimer and Dahlberg 1983, 1984;Dahlberg and Fowler 1985; Dahlberg et al. 1986).

The information presented to this point refersonly to coastal distribution of particular stocks ofchinook salmon. Information on the open oceandistribution is much more sketchy and speculative,as most of it is not based on tag returns from fish ofknown origin.

Approximately 2,099 chinook older than age 1.have been tagged and released on the high seaswest of 150° W and in the Bering Sea between 1956and 1984 (C. Harrison, Fisheries Research Insti-tute, University of Washington, Seattle, Washing-ton, pers. comm.). Thirty-three of these have beenrecaptured (Figure 26), and these recaptures pro-vide some information on the ocean migrations ofstream-type chinook. Recaptures up to 1972 werediscussed by Major et al. (1978).

Twenty-one fish tagged in the central BeringSea were recaptured within the Bering Sea andits coastal areas (Figure 26). Five had moved to-wards the west (northwest to southwest) from the

361

1 i

i !! i

PANEL ARECAPTURED DURING YEAR

OF RELEASE

|E5536 Ee036| E653e| E703a|E753e| 6036 W5536 W503a| W4538 W4038JW3538 W3038| \

PANEL BRECAPTURED YEAR FOLLOWINGRELEASE

=4036 ]E453a|E503a EBS36 E903B | E653B E703B | E7S3a^ 6038 |w7S36 W70M [wBS36 W003a|wftB3e W503B [w453S W4038 |tfV3538 W3036

1 1 1 1 1 T

PANEL CRECAPTURED TWO OR MORE

YEARS FOLLOWING RELEASE

E553fl E6038 |E6538 E7Q38 |E753B 8038 |w7538 IV7038 |»V653B W603e|w553a W503a|w4536

160'E 170'E 170'W 180-W 150-W 140-W 130-W

FIGURE 26The location of tagging ( • ) and recapture (— ) of chinook tagged on the high seas in the central Bering Sea,

in the N.W. Pacific, and south of Adak Island in the Aleutian chain. Panel A, chinook recaptured in the year of tagging.Panel B, chinook recaptured one year after tagging. Panel C, chinook recaptured two or more years after tagging

362


point of release, and three a moderate distancenorth or northeast. Three of these were immatureat the time of recapture and the other four were ofundetermined maturity. Eight had moved east toriver mouths in western Alaska, and these were allmature at the time of recapture, most being recap-tured two years after tagging. The remaining fivefish were recaptured only a short distance fromtheir point of tagging. Scant as these results are,they suggest a north and westward movement ofimmatures and an eastward movement of matures,at least of those bound for western Alaskan rivers.

Of five chinook tagged and recaptured in thewestern North Pacific Ocean, three were recap-tured south or southwest, one north, and one westof their tagging sites (Figure 26). Their maturity atthe time of recapture was not determined. Fourfish tagged south of Adak Island in the Aleutianchain were recaptured, three as maturing fish andone of undetermined maturity (Figure 26). Of thematuring fish, one was recaptured in Bristol Bay,one in southeastern Alaska, and one at the Colum-bia River mouth. The fourth fish was recapturedsouthwest of Adak Island. One other fish taggedsouth of the Aleutians was recaptured off the eastcoast of Kamchatka one year later. These returnsreveal only that chinook from the northern NorthPacific Ocean are from diverse origins and that fishfrom the Columbia River may migrate as far westas Adak Island.

The costliness, and the limitations, of ocean tag-ging have led investigators, particularly those inthe United States, to explore other techniques foridentifying the origin of chinook captured on thehigh seas. One of these techniques is the estima-tion of stock composition by discriminant analysisof scale measurements.

The earliest attempts to classify high-seascatches of chinook by this technique were de-scribed by Major et al. (1978). In these early studiesonly western Alaskan and Asian stock groupingswere recognized. The analysis was expanded byKnudsen et al. (1983) and Myers et al. (1984) toInclude classification of central Alaskan, southeast-ftfn Alaskan/British Columbian, and Washington/fJregon/Californian stock groupings. The Wash-ttlgton/Oregon/Californian stock grouping wasilinunated from further consideration by theseWithers when they found that it made an insignifi-

contribution to high-seas catches. Most re-

cently, Ito et al. (1985,1986) and Myers (1986) haveinvestigated how altering the proportion of thedifferent Asian stocks used as known standards indeveloping the discriminant functions affects theclassification of unknown samples from the Japa-nese mothership and landbased fisheries. Myers etal. (1987) summarized findings from recent u.s.analyses.

The results of these analyses are controversial,particularly when they are used to estimatenumbers of fish of different origins landed by aparticular fishery. Before presenting some of theresults of these analyses, the most important criti-cisms of the methodology will be summarized.There can be no doubt that the quantitative resultsof the discriminant function analysis are sensitiveto the composition of the standards used to de-velop the discriminant functions, and to assump-tions about which stocks or stock groupingscontribute to the catch in any high-seas fishery.Because the analysis is sensitive to these factors, itis my view that it is inappropriate at this time touse the discriminant function analysis to estimatestock composition quantitatively. It is also myview, however, that the analyses are sufficientlyconsistent, particularly with respect to mixingproportions of some stocks, to provide a reason-able basis for speculation about qualitative fea-tures of the high-seas distribution of chinook.

There are numerous technical difficulties in-volved in applying discriminant function analysisof scale features to the analysis of a sample ofscales from fish of unknown origin. A set of scalesappropriate for such an analysis must be compiledfrom known populations. Features of the scalesthat are unambiguous in terms of their measure-ment, and that differ between populations but arereasonably consistent within populations, must bediscovered if classification by discriminant func-tion analysis is to be successful. The scales of un-known origin that are to be classified must becomparable with those used to develop the origi-nal discriminant functions and must comprise anunknown mixture of only those stocks for whichdiscriminant functions have been calculated.Failure to meet these and other technical require-ments in the data can result in errors in classifica-tion of unknown seriousness. Myers et al. (1984)and Myers (1986) report rather careful screening ofthe data to minimize the possibility of such errors,

363


such as ensuring that the scales used in the analy-sis were from the preferred area of the body, andthat the discriminant functions were brood yearspecific. It is virtually impossible, however, toeliminate all such sources of error.

An important potential source of error in theclassification of an unknown mixture of scales isthe presence of scales from stocks not included inthe standards used to develop the discriminantfunctions. Such scales will be classified into one ormore of the stocks specified in the analysis and, ifabundant, could greatly bias the estimate of stockcomposition.

The success in classifying the scales of knownorigin from which the discriminant functions werecalculated provides one indicator of the magnitudeof error that might occur in classifying a sample ofunknown scales. It should be remembered, how-ever, that there may be greater error in the classifi-cation of individual scales than in the estimate ofoverall stock composition.

The overall success in classifying fish of knownorigin by the discriminant functions developedfrom those fish ranged from 69.4% to 79.7% in thestudies reported by Myers et al. (1984) and Myers(1986), and was 60.9% in the study reported by Itoet al. (1985). Although these are reasonably highrates of successful classification, there is still roomfor considerable error in quantitative estimation ofstock composition. For individual stock groupings,the success in classifying known samples rangedfrom 58.5% to 95.5% for the Asian grouping, from64.6% to 89.9% for the western Alaskan grouping,from 47.0% to 67.7% for the central Alaskan group-ing, and from 70.5% to 83.5% for the southeasternAlaskan/British Columbian grouping (Myers et al.1984, Ito et al. 1985; Myers 1986). The central Alas-kan stock grouping was the grouping most likelyto be misclassified in all analyses, and these fishtended to be misclassified as belonging either tothe southeastern Alaskan/British Columbian orAsian stock groupings rather than to the westernAlaskan grouping (Myers et al. 1984; Ito et al. 1985,1986; Myers 1986). The directions and degree ofmisclassification of central Alaskan scales de-pended to a considerable extent on the stock pro-portions in scales used for the Asian standard(Myers 1986).

Considering the potential sources of bias anderror it is not surprising that the quantitative esti-

mates of stock composition produced by differentanalyses differ considerably. In Figure 27 the esti-mates from recent publications of stock composi-tion of immature chinook captured in July insubareas of the Japanese mothership fishery areshown (Myers et al. 1984; Ito et al. 1986; Myers1986). Despite the quantitative disagreement in theresults of different analyses, there appears to beconsiderable qualitative agreement.

In the Bering Sea, all the analyses agree thatwestern Alaskan (including the Canadian Yukon)chinook are the most abundant stock grouping.Next in abundance are Asian and central Alaskanchinook, which are each about half as abundant aswestern Alaskan chinook. The southeastern Alas-kan/British Columbian stock grouping is rare inthe Bering Sea (Figure 27).

In the western North Pacific Ocean, the greatestdisagreement is over the relative contribution ofcentral Alaskan and Asian stock groupings; theanalyses of Myers et al. (1984) and Myers (1986)suggest that central Alaskan chinook often domi-nate in this area, whereas the analyses of Ito et al.(1985,1986) suggest that Asian chinook dominate(Figure 27). This disagreement is not, as yet, tech-nically resolvable. The very high percentage ofcentral Alaskan chinook in the western North Pa-cific Ocean suggested by Myers et al. (1984) (some-times as high as 100%) seems inconsistent with theapparent abundance of this species in spawningescapements in central Alaska. Myers et al. (1987),however, suggested that chinook may be moreabundant in central Alaska than previouslythought. Comparison of measured characteristicssuggests similarity between Asian and central Al-askan scales in a number of features. It seemspossible that there may be considerable misclassi-fication of Asian scales as central Alaskan and viceversa, depending on the scale characters and thestocks used to calculate the discriminant func-tions. Leaving aside the relative contributions ofAsian and central Alaskan chinook, all the analysessuggest that the chinook population of the west-ern North Pacific Ocean is comprised mainly ofstocks from Asia, western Alaska, and centralAlaska, and that each of these stocks makes asubstantial contribution to the population. South- .eastern Alaskan/British Columbian chinook, al-though perhaps slightly more abundant than inthe Bering Sea, still constitute only a small per- t

364


160°E 165°E 170°E 175°E 180° 175°W

60°N

AS

WAK

CAK

SE/BC

AS

WAK

CAK

SE/BC

48°N

AS

WAK

CAK

SE/BC

/tn°N

A B C

(1)

0-13 15-57

7-7 9-34

80-83 27-39

0-10 7-12

(11)

0-10 12-55 7-23

0-18 19-33 7-23

82-96 9-41 50-80

0-0 6-14 3-7

A B C

H)

0 17

75 66

25 17

0 0

(3)

0-26 8-68 4-32

0-46 19-53 19-52

34-79 10-36 12-56

0-6 1-3 3-13

A B C

(6)

10-38 7-30 2-15

46-83 58-77 68-77

0-14 5-9 13-19

0-7 7-7 1-3

(5)

0-39 18-44 18-32

0-28 46-48 23-49

33-100 6-30 14-49

0-25 2-6 4-21

A B C

(8)

0-48 10-24 2-35

40-82 55-63 40-84

0-31 19-25 9-32

0-7 2-2 0-7.5

(7)

0-37 17-36 16-34

6-41 28-36 26-61

16-78 28-35 13-38

0-12 8-12 3-14

A B C

(10)

1-18 6-42 0-21

44-94 30-59 57-85

0-39 25-32 8-18

0-6 2-3 0-8

(9)

13-78 18-46 5-54

5-43 27-34 15-72

17-66 16-39 13-59

0-1 7-9 0-12

62°N

52"N

46°N

FIGURE 27Estimates by various authors of stock composition of chinook captured in subdivisions of the Japanese mothership

fishery based on discriminant function analysis of scale characteristics showing variation in results. AS = Asia,WAK - Western Alaska, CAK - Central Alaska, SE/BC - southeastern Alaska/BC. The numbers (l)-(ll) represent sub-

divisions of the Japanese mothership fishing area. A = estimates of stock composition from Myers et al.'s (1984) dis-criminant analysis of scale characters; range of values for brood years 1971-77. B - estimates of stock composition

from Ito et al.'s (1985,1986) discriminant analysis of scale characters; range of values for different assumptions aboutAsian stock origins; brood year 1970. C - estimates of stock composition from Myers's (1986) discriminant analysis ofscale characters; range of values for different assumptions about Asian stock origins; brood years 1973, 1974, and 1976

centage of the population in the western NorthPacific Ocean (Figure 27).

Summary of Ocean Distribution of Chinook

Despite the many extensive and intensive investi-gations that have been conducted, our understand-ing of the ocean migratory and distributionpatterns of chinook is still very sketchy. Informa-tion on the distribution of ocean-type populationshas been derived almost exclusively from the in-tensive ocean troll fisheries off Oregon, Washing-ton, and British Columbia. Nevertheless, therelatively small number of recaptures of southern

British Columbian and Columbia River chinook insoutheastern Alaska, where there is an active trollfishery, suggests that most ocean-type chinook donot disperse more than about 1,000 km from theirnatal river. Some go much further, of course, asrevealed by the recaptures from south of the Aleu-tian chain and in the southeastern Bering Sea. Allrecaptures to date are from near the coast, and thedispersal of ocean-type populations offshore isrevealed only in catches by Canadian researchvessels fishing in the eastern North Pacific Ocean.The low catches of chinook in this part of theNorth Pacific Ocean, and the low proportion ofocean-type chinook among the few that were cap-

365


Ocean-type ChinookStream-type Chinook - Central Alaska & SouthStream-type Chinook - Western AlaskaStream-type Chinook - Asian

70°N

60°N

—X : 50°N

-40°N

120° 140°E 160°E 180 160°W 140°W 120°W

FIGURE 28The likely ocean distributions and relative abundance in different parts of the North Pacific Ocean and Bering Sea

of ocean- and stream-type chinook from various regions of the North American and Asian coasts. Stream-typechinook from western Alaska include Canadian Yukon River chinook. See text for details regarding the basis of

these distributions and difficulties with the data.

tured, suggests that ocean-type chinook do notoften wander far from shore (Figure 28) or thatthey are found at greater depths than those fishedby the research vessel gear.

The dispersal of stream-type chinook appears tobe much broader, as revealed by their contributionto both coastal and high-seas catches. Recapturesof tagged stream-type chinook from the ColumbiaRiver in coastal troll fisheries show greater disper-sal of this race, both north and south of the rivermouth, than of ocean-type populations. The highcontribution of stream-type fish to the high-seascatches throughout the North Pacific Ocean andBering Sea further indicates much greater offshoredispersal of this race (Figure 28).

The high-seas distribution of stream-type chi-nook from different regions of the North Americancoast and Asia is a matter of considerable debate.Nevertheless, I believe a few speculations may bemade about such distributions, with the under-

standing that these will be subject to modificationas more information becomes available.

Asian chinook are probably distributed through-out the Bering Sea but are also probably concen-trated west of 180°. In the northern North PacificOcean they appear to be distributed at least as fareast as 175°W, and probably further, and to berelatively more abundant in the western NorthPacific Ocean than in the Bering Sea. The southernlimit of their distribution is not known but is atleast 40°N latitude and perhaps further south (Fig-ure 28).

Western Alaskan chinook (including CandianYukon chinook) are also distributed throughoutthe Bering Sea. Chinook are probably much moreabundant in western Alaska than in Asia, so thatchinook from western Alaska tend to dominateBering Sea catches, even in the western half of thesea. Alaskan chinook are, nevertheless, probablyrelatively more abundant in the eastern and cen-

366


tral than in the western Bering Sea. Western Alas-kan chinook apparently also migrate south of theAleutian chain into the North Pacific Ocean. Thelimits of their distribution there are not known, butthey may occur as far westward as 160°E, andprobably extend considerably east of 175°W. West-ern Alaskan chinook may also be distributed as farsouth as 40°N (Figure 28). I believe that westernAlaskan chinook will ultimately be shown to havea very wide distribution in the western NorthPacific Ocean, but they are probably no moreabundant than Asian chinook in these waters(Knudsen et al. 1983; Myers et al. 1984) (Figure 27).

Central Alaskan chinook are also probablywidely distributed in the central and westernNorth Pacific Ocean and Bering Sea (Figure 28). AsI noted earlier, the relative abundance of centralAlaskan chinook in these waters is controversial. Isuspect that they will ultimately be shown to beless abundant than Asian chinook in the BeringSea and western North Pacific Ocean.

Southeastern Alaskan/British Columbian chi-nook appear to be relatively rare in both the BeringSea and the western North Pacific Ocean, at leaston the basis of scale analysis (Figure 27). Sincevirtually all chinook captured in the Bering Seaand western North Pacific Ocean are stream-type,any fish of the southeastern Alaskan/British Co-lumbian stock grouping captured there must havebeen of the stream-type race. The southeasternAlaskan/British Columbian chinook, as well asthose from Washington, Oregon, and California,are probably distributed mainly in the easternNorth Pacific with the greatest concentrations overthe continental shelf waters along the North Amer-ican coast (Healey 1983).

Ocean Food Habits

Most data on food habits of chinook in the oceanare from samples taken in the commercial fisheryand, therefore, relate to larger fish. Healey (1980a),however, reported on the diets of chinook of 10-30cm fork length captured in the Strait of Georgia inlate summer, 1975 and 1976. Small fish, particularlyherring, pelagic amphipods, and crab megalopamade up between 70% and 92% of the diet, withfish being the largest single contributor at28%-63% of the diet. Adult insects were also im-portant in the diet of chinook captured in 1976 but

not in 1975. There was evidence for regional varia-tion in diet within the Strait of Georgia in that fishcomposed about 79% of stomach contents in theGulf Islands region but only 37% in the FraserRiver plume and central region of the Strait ofGeorgia. Pelagic amphipods were relatively unim-portant in the Gulf Islands region but constituted24% of stomach contents in the Fraser River plumeand 14% of stomach contents in the central Strait ofGeorgia. Crab megalopa increased in importancefrom 11.5% of stomach contents in the Gulf Islandsregion to 46.4% of stomach contents in the centralStrait of Georgia. Insect adults were only impor-tant in the Fraser River plume.

For larger chinook, diet information is availablefrom southeastern Alaska south to the Californiacoast. The data span a number of years and timesof year and thus provide evidence for regional,annual, and seasonal changes in diet. Techniquesof data collection and analysis have not been con-sistent among investigators, however, so thatquantitative comparisons must be viewed withcaution.

Reid (1961) described the diet of chinook cap-tured in the troll fishery off southeastern Alaska in1957 and 1958. Stomachs were sampled from mid-June to mid-September. The volume of stomachcontents was measured, and organisms were iden-tified taxonomically and counted. Thirteen differ-ent taxa were identified in the stomachs, but by farthe most important diet item was herring(60%-68% of volume). There were some indicationsof differences in diet between years. For example,squid were about ten times as common in the dietin 1958 as in 1957, whereas fishes other thanherring were much more important in 1957. Therewas also an indication that squid were more im-portant in sheltered inside waters than in the moreopen water fishing areas. The importance ofherring relative to other foods was related to sizeof chinook, increasing from less than 20% of dietfor chinook less than 64 cm to more than 60% forchinook more than 100 cm in length.

Pritchard and Tester (1944) and Prakash (1962)described the diet of chinook salmon in BritishColumbian waters. In both studies, samples werecollected from the commercial fisheries. Pritchardand Tester (1944) sampled from the whole coastduring 1939-1941, whereas Prakash (1962) sampledonly the west coast of Vancouver Island, the Juan

367


de Fuca Strait, and the Fraser River estuary. Con-tribution of various taxonomic groupings to dietwas estimated as a percentage of the total stomachvolume in both studies.

Pritchard and Tester (1944) recorded 21 differenttaxonomic groupings in the diet of chinook duringthe three years of sampling. Fish of various sorts,but especially herring and sand lance (Ammodyteshexapterus), dominated the diet. Invertebrate taxanever exceeded 6% of the diet. Some between-yearvariation in diet was evident, most notably in theabsence of pilchards (Sardinops caerulea) from thediet in 1939 and their relative importance in 1940and 1941. This change appeared correlated withthe strength of the pilchard stocks in southernBritish Columbia. Regional variation in diet wasalso apparent but chiefly in the alternation in rela-tive importance of sand lance, herring, and pil-chards. Off the northwest coast of the QueenCharlotte Islands, for example, herring and sandlance were almost equal in importance and pil-chards were unimportant. On the northern main-land coast, herring were relatively important onlyin 1939 and were replaced by pilchards in 1940 and1941. Off the northwest coast of Vancouver Island,sand lance predominated, constituting about 70%of the diet. Various invertebrates were also some-times important in this region. On the southwestcoast of Vancouver Island, herring predominatedtogether with pilchards in 1940 and 1941. In 1939,sticklebacks were important (presumably replac-ing pilchards). In the northern Strait of Georgia,herring was the dominant food; and in the south-ern Strait of Georgia, pilchards were dominant.These regional differences in diet are unlikely toreflect patterns of dietary preference among re-gions. Rather, they probably reflect the relativeabundance of potential prey between regions andyears. Pritchard and Tester (1944) were of the opin-ion that chinook were largely opportunistic feed-ers. Apart from the Chinook's apparent preferencefor fish as prey, this interpretation is probablycorrect.

Prakash (1962) found herring to be the mostimportant diet item of chinook off southern Van-couver Island (72.5% of the diet). Larger herringwere taken off the west coast of the Island than offthe east coast (13-23 cm compared with 5-10 cm).As Reid (1961) noted in Alaska, large chinook ten-ded to have a higher percentage of herring in their

diets than did small chinook. Consequently, stom-ach samples from the west coast of VancouverIsland, which were from larger chinook, indicateda higher proportion of herring in the diet thanstomach samples from the east coast of the island.Invertebrate food organisms (mainly euphausiids)dominated the diet of chinook from the east coastof Vancouver Island in May and June, but other-wise there was no evidence for significant seasonalchanges in diet.

Robinson et al. (1982) reported on the stomachcontents of 27 chinook captured in the QualicumRiver area of the Strait of Georgia during April andMay, 1981. They ranged from 29 to 72 cm in length.Their principal diet items were chum salmon fry,larval herring, sand lance, euphausiids, and adultherring. This is a rare instance in which predationby chinook upon the juveniles of another salmonspecies has been documented.

Silliman (1941) reported on the diet of chinookcaptured by trolling off the coasts of Washingtonand Vancouver Island in 1938. Stomach contentswere measured gravimetrically. Herring domi-nated the diet of chinook captured off VancouverIsland (Pritchard and Tester 1944; Prakash 1962),but euphausiids dominated off Westport, Washing-ton (43% of the diet). Fish of various sorts stillmade up most of the diet in the Westport area, butno one species overshadowed the rest.

Heg and Van Hyning (1951) described the diet ofchinook captured by trolling off the Oregon coastin 1948-50 and of chinook captured by sports fish-ermen in 1950. Clupeids and clupeid remains dom-inated the diet, and anchovies (Engraulis mordax)were the most important single species. Inverte-brates (predominantly euphausiids) constitutedonly 4%-5% of the diet.

Merkel (1957) reported on the diets of chinookcaptured in the ocean sport troll fishery off SanFrancisco. Sampling extended from February toNovember, and samples were adequate to permitmonthly comparisons of diet as well as comparisonby size of fish and location of capture. Northernanchovy (29.1%) and various juvenile rockfishes(22.5%) were the most important diet items, al-though euphausiids (14.9%) and herring (12.7%)also contributed significantly. Seasonal variationsin diet were dramatic. Anchovies dominated thediet from August to November, rockfishes duringJune and July, and herring in February and March.

368


During April and May, invertebrate foods, espe-cially euphausiids but including squid and crabmegalopa, were dominant. The importance of in-vertebrate foods in April and May, but also to someextent in June, is reminiscent of Prakash's (1962)observation that invertebrates dominated the dietof chinook off southeastern Vancouver Island inMay and June.

Diet was also related to location of capture (Mer-kel 1957). In water shallower than 20 fm (38 m),anchovies dominated (91%-92% of diet). In waterdeeper than 20 fm (38 m), rockfishes dominated(71%-74%); but a wider variety of organisms oc-curred in stomachs, and herring and euphausiidswere significant diet items. Anchovies were lessthan 5% of the diet of chinook taken over deepwater and rockfishes less than 1% of the diet ofchinook taken in shallow water. As other authorshave observed, small chinook fed more heavily oninvertebrates than large chinook. Chinook lessthan 25 in. (63 cm) long captured near the FarallonIslands (California) had 85% invertebrates in theirstomachs, whereas chinook more than 25 in. (63cm) long had only 52% invertebrates in their stom-achs.

Viewed together, the coastwide data on chinookdiet suggest some regional trends (Figure 29). Ingeneral, the importance of herring and sand lanceincreases from south to north, whereas the impor-tance of rockfishes and anchovies decreases. Thesetrends probably reflect the relative abundance ofthe prey along the coast. Pilchards were importantin the central parts of the chinook's range prior tothe collapse of the pilchard stocks. More recentobservations suggest that pilchards are not impor-tant in the diet of chinook at present, but thatanchovies are more important further north than isindicated by the historic data. Euphausiids areimportant from time to time, and there is no cleargeographic trend for their importance. Other preyitems are incidental. The importance offish in thediet of chinook is apparent in virtually all studiesand, in fact, chinook appear to be the Oncorhynchusspecies most dependent on fish as food (Healey1976).

Not only are seasonal and regional variations iniiet composition apparent but so are seasonal andregional variations in feeding intensity. Healey,1982c) found that the weight of stomach contentssf first ocean-year chinook in the Strait of Georgia

ranged from 0.4%-1.73% of body weight betweenregions and years of sampling. Stomach contentstended to be highest in areas where catch washighest, suggesting that the young chinook werecongregating in good feeding areas. For commer-cial-sized fish, Prakash (1962) found that stomachcontents were greater off the west coast of Van-couver Island than off the east coast during May toSeptember, and that contents tended to be great-est in August (Figure 30). The percentage of stom-achs with food was uniformly high (greater than80%) among troll-caught fish, except for samplesfrom the west coast of Vancouver Island and theJuan de Fuca Strait in September. Chinook sam-pled from the Fraser River gillnet fishery had littlein their stomachs, and the percentage of stomachswith food declined from about 50% in July to 0% inOctober. Stomach contents of chinook captured offthe west coast of Vancouver Island and coastalWashington during April-October 1938 rangedfrom 0.1 to 63.7 g/fish (Silliman 1941). Contentsincreased during the sampling period off the westcoast of Vancouver Island but decreased off thecoast of Washington (Figure 30). Silliman (1941)found a significant relationship between salmontroll catch and the weight of fish in the diet ofchinook. He interpreted this to reflect the relativeattractiveness of troll lures, which mimic preyfishes, during periods when the chinook werefeeding on fish. Finally, Merkel (1957) observedthat stomach contents of chinook off Californiawere least in spring and fall and greatest in earlysummer (Figure 30).

Although highly variable, the data on stomachcontents suggest that chinook feed most activelyin spring and summer. They also suggest that thebest feeding periods are during July and Augustoff southwest British Columbia but earlier, duringApril to June, from Washington to California. Thesedifferences in seasonal feeding patterns may, inturn, reflect the differences in diet along the coastdescribed earlier (e.g., the greater importance ofanchovy and rockfishes south of Washington andthe greater importance of herring and sand lancenorth of Washington).

Ocean Growth

Healey (1980a) and Miller et al. (1983) provideddata on the growth of chinook salmon during their

369

§ 40

H S P R A O ' E S O

1

FISH INVERTEBRATES

FIGURE 29Diet composition of chinook from southeastern Alaska to California

370


Ioto

1£"5«Eo

60

50-

40 -

30-

20 -

10-

East coast Vancouver I. (1957)West coast Vancouver I. (1938)West coast Vancouver I. (1957)

Coastal Washington (1938)

Coastal California (1954-55)

M

FIGURE 30Seasonal changes in the weight of stomach contents" of chinook captured off

Vancouver Island, Washington, and California

first summer at sea. For chinook in the Strait ofGeorgia in 1976, modal size groups increased inlength by 0.75-0.85 mm/d. Chinook captured offthe Washington/Oregon coast in May-September1980 increased in length by 1.74 mm/d. These datamust be interpreted cautiously, as there is littlelikelihood that the same cohort of chinook wassampled throughout the time series of sampling ineither study. In particular, the increase in length ofchinook off the Washington/Oregon coast seemshigh. By comparison, chinook sampled by Fisheret al. (1984) in the same area showed little increase

in mean length (0.3 mm/d).Loeffel and Wendler (1969) summarized the

known information on growth of chinook at seabased on time series of samples from a single loca-tion and on back-calculations from scales. Figure31, adapted from their report, represents a com-posite picture of growth of stream- and ocean-typeraces. The data suggest a definite seasonal patternof growth in both races, with rapid summergrowth and slow winter growth. Faster growingfish mature at a younger age as is clearly shown bythe back-calculated lengths of fish that matured at

' I'M

1?

371


1000-

900-

800-

E 700-

g 600-

£ 500-

400-

300-

Ocean-typechinook

MM|.RV

c*R

_ R

F R *"""-•'" MV

...-r'.fl MV

,-R R

1000-

900-

-^ 800-

E.J£ 700-D>

i 600-

£500-

400-

300-

200-

100

Stream-typechinook

PF

bRR

R--

..P' P

,R

DEC DEC DEC DEC DEC DEC2 3 . 4 5 6

Year of lifeFIGURE 31

Growth curves (fork length at age) for ocean- and stream-type chinook. Lettersrefer to data from different sources. (Adapted from Loeffel and Wendler 1969)

different ages (Figure 32). Grachev (1967) observedthe same pattern in Kamchatka River chinook, andNeilson and Geen (1986) observed that age at ma-turity was negatively correlated with the size ofchinook at the end of their first year in the ocean.The apparent seasonality of growth (Figure 31)

may be accentuated, therefore, by the disappear-ance from the ocean population each summer andautumn of the larger members of each age classthat are returning to fresh water to spawn. Ocean-type chinook are larger than stream-type chinookat every calendar age. This is because the stream-

I

372


E

o

"5

<c

en

.§

1000-n

900-

800-

700-

600-

500-

400-

300-

200-

100-

0-

Ocean-typechinook

2 3 4

Age at capture

c

Stream-typechinook

~r2 3 4

Age at capture

—i6

FIGURE 32Length at annulus formation for ocean- and stream-type chinook which mature at different ages. Data from south-

eastern Alaska (•), Strait of Georgia (o), west coast of Vancouver Island (A), and the Fraser River (n)

type fish grow more slowly than ocean-type dur-ing their first year of life. Rates of growth of thetwo races during ocean life are similar (Figure 31).

Taken at face value, the available data suggestthat, for North American populations, length atage is greater in the south than in the north (Figure33). The data are a little difficult to interpret, how-ever, as the Alaskan samples were from the oceantroll fishery and include many immature fish, ofwhich some would be from Fraser River, ColumbiaRiver, and other, more southern stocks (Figures22-24) (Parker and Kirkness 1956). Size at age ofKamchatka River chinook was similar to chinookcaught off southeastern Alaska (Grachev 1967),whereas on the basis of latitude I would have ex-pected them to be more similar to Yukon Riverchinook. Also, the growth rates offish aged two tosix years are similar for all areas sampled (0.35-0.57mm/d). In fact, the Columbia and SacramentoRiver populations of both races had the slowestgrowth rate over the ages two to six years. The

differences in size between regions are entirelydue to differences in size at age 0.1 in ocean-typechinook or age 1.1 in stream-type chinook. Con-ceivably, the differences in size at these ages couldbe due to differences among investigators in inter-pretation of annuli. Chinook are notoriously diffi-cult to age from scales (Godfrey et al. 1968).Conversely, it is possible that growth during thefirst ocean year differs dramatically along thecoast. The apparently greater growth of firstocean-year fish off the Columbia River, comparedwith those in the Strait of Georgia (see earlier), isconsistent with this explanation.

Grachev (1967) reported back-calculated growthincrements for Kamchatka River chinook spanning23 brood years, from 1933 to 1955. The most com-plete samples were of males representing ages 1.2,1.3, and 1.4. Mean lengths in the year of migrationvaried from 58.3 to 70.4 cm for fish aged 1.2, andfrom 72.7 to 88.5 cm for fish aged 1.3. Total lengthwas most closely correlated with the length incre-

i

t i l

II t :

373


E

O)

£

S.E. Alaskao S. British ColumbiaA Columbia RiverA Sacramento River

Stream-type

E,

O)

100

Yukon River• S.E. Alaskao S. British ColumbiaA Columbia River

—i8

Age (years)

FIGURE 33Length at age for chinook captured in the southeastern Alaska troll fishery andrivermouth fisheries in southern BC and the Columbia and Sacramento rivers

ment laid down during the final ocean year, sug-gesting that feeding conditions during this lastyear at sea were most important in determiningfinal adult size of each age class.

All of the above analyses of growth are incom-plete and superficial because they do not take fullyinto account the complex interrelationships among

growth, migration, and maturation in chinook.Samples taken in ocean fisheries consist of repre-sentatives of numerous spawning stocks. Theyounger fish in a sample will be mainly fromnearby spawning populations and the older fish,mainly from distant spawning populations. Forchinook older than the age of first maturity, those

374


remaining at sea will be the smaller, slower grow-ing component of any spawning stock. Samplestaken on the spawning grounds suffer from theopposite problem. The spawning fish in theyounger age groups will be the faster growing

members of the age class. Any model of growth forchinook must take account of these phenomena.Current data are inadequate for the task. Morerecent tag recovery data may prove sufficient, butthese data are as yet largely unanalysed.

MATURATION AND RETURN TO SPAWN

As was the case with growth rate, estimating age-specific rates of maturity in chinook is a complexproblem. Again, current data are inadequate forthe task. Three approaches have been taken todate:1 The proportion of each age class in samples

captured at sea that will mature in the year ofcapture is determined from gonad/body weightratios or egg size (Rich 1925; Borque and Pitre1972; Wright and Bernhard 1972; Ito et al. 1974;Baranski 1979).

2 Maturity schedules are estimated from tag re-turns (Parker and Kirkness 1956; Cleaver 1969).

3 Maturity schedules are inferred from the agecomposition of spawning runs (various authors,but see Godfrey 1968a; Loeffel and Wendler 1969;Hankin and Healey 1986).Mature chinook captured on the spawning

grounds range from 0.1 to 0.5 in age for ocean-typestocks and from 1.0 to 2.5 for stream-type stocks.Not all populations have all ages, however, andmost populations are dominated by a few ageclasses. Two examples of age-frequency curves willsuffice to illustrate the differences between sys-tems, races, and sexes (Figure 34). In the YukonRiver (Brady 1983), which has only stream-typechinook, six-year-old fish of both sexes predomi-nate in the spawning run. Males dominate theyounger ages (ages 4 and 5), however, and femalesthe older ages (ages 7 and 8). In the Fraser River,which has both stream- and ocean-type fish (God-frey 1968a), four-year-old fish predominate in therun, and this is the most abundant age class formale and female ocean-type and male stream-typechinook. Female stream-type fish were most abun-dant at age 5. For both races, males were more

FemalesMales

A Females (ocean-type)O Males (ocean-type)

Females (stream-type)Males (stream-type)

1 2 3 4 5 6 7

Years of life

FIGURE 34Composition by age (as a percentage) of spawning runsto the Yukon and Fraser rivers for ocean- and stream-

type chinook

common among the younger fish, and femalesamong the older fish.

The average age of mature male ocean-type chi-nook ranged from 3.02 to 3.88 years in river sys-

375


H i

tems for which comparative data exist, and from3.98 to 4.32 years for mature female ocean-typechinook (Table 7, see table for references). Forstream-type chinook, the range was from 3.69 to5.64 years for males and from 4.39 to 6.12 years forfemales (Table 7). Thus, males mature at a youngeraverage age than females, regardless of race. Thedifference in age between males and females wassmall in some instances (0.1 y) but in other instan-ces it was large (1.3 y) (Table 7). From the Sacra-mento River (37°45'N) to the Nass River (55°00'N)there was no consistent trend in the age of matu-rity among sexes and races. There was a weakpositive correlation between age and latitude forstream-type females and a weak negative correla-tion between age and latitude for ocean-typemales, but neither correlation was significant.North of the Nass River, however, and in Kam-chatka, chinook appear to mature about a yearolder, on average (Table 7). Thus, the stream-typeraces, at least on a geographic scale, have a longergeneration time than the ocean-type, presumablyowing to their longer freshwater residence.

The standard deviation of age at maturity formales ranged from 0.601 to 1.058, with no consis-tent geographic or racial trend. For females, thestandard deviation of age at maturity ranged from0.206 to 0.698, again with no consistent racial orgeographic trend. The variation in male age of

return was, therefore, consistently larger than thevariance in female age of return. This trend isreflected in the high percentage of females return-ing to spawn at a single age, compared with males(Figure 34).

The age composition of spawning runs providesa measure of the ages over which chinook maturebut gives only a rough indication of age-specificmaturation rates. These latter rates depend notonly on a knowledge of the number of mature fishin each age group but also on the number of imma-tures. Such data are provided by sampling popula-tions at sea and assessing the proportion ofmaturing fish in each age class, and by analysingtag returns to estimate both mortality and matu-rity rates.

On the basis of tag returns from Columbia Riverhatchery chinook, Cleaver (1969) estimated matura-tion rates at various ages for the 1922 brood year,1954-55 brood years, and the 1961 brood year.Rates of maturation ranged from 1% to 5% for chi-nook in their second year, from 18% to 35% forchinook in their third year, from 61% to 96% forchinook in their fourth year, and were 100% forchinook in their fifth year (Table 8). There wasevidence for a greater proportion of mature fishamong the youngest age groups in recent years.

On the basis of tag returns from ocean taggingoff southeastern Alaska, Parker and Kirkness (1956)

TABLE 7Mean age (years) and standard deviation of age for mature stream- and ocean-type chinook from rivers

throughout the geographic range of the species

Mean age Standard deviation

River N. Latitude Race Male Female Male Female Source

SacramentoKlamath

Columbia

Fraser

Skeena

Nass

TakuYukonKamchatka

37°45'41°30'

46°15'

49°15'

54°15'

. 55°00'

58°30'62°31'S6°00'

OceanStreamOceanStreamOceanStreamOceanStreamOceanStreamOceanStreamStreamStream

4.403.883.983.554.133.594.133.473.693.02

5.64

4.184.603.984.394.294.564.004.564.194.824.32

5.296.12

5.39

0.7640.6180.6671.0580.6010.7320^6770.8270.8760.745

0.803

0.7790.5920.4580.5470.6980.5630.2060.4850.4410.4050.476

0.5390.425

0.618

Clark (1929)Snyder (1931)

Rich (1925)

Godfrey (1968a)

Godfrey (1968a)

Godfrey (1968a)

Kissner (1973)McBride et al. (1983)Vronskiy (1972)

Note: Sexes not distinguished in Sacramento, Taku, and Kamchatka rivers

376


TABLE 8Age-specific maturity schedules (proportion maturing at age) for male and female ocean-type chinook

as calculated by various authors for various times and locations

Age

Location

Columbia R.*

Wash, coastt

Puget Sd.t

Vancouver Is.t

S.E. Alaska*Columbia R.J

Year(s)

19221954-5519611972

1975-77

1967-70

1950-521919

Sex

male/femalemale/femalemale/femalemalefemalemalefemalemalefemalemale/femalefemale (S)§female (O)§

2

0,0090.0500.0090.3000.0300.1200.0300.3600.020

-0.0000.000

3

0.1810.3500.3380.7000.4900.4800.2400.4400.1200.3200.0000.150

4

0.7540.6100.9590.9900.9900.9500.9500.7000.6300.5450.7700.870

5

1.000-

1.0001.0001.000

----

0.9050.9601.000

Source

Cleaver (1969)

Wright & Bernhard (1972)

Baranski (1979)

Borque & Pitre (1972)

Parker & Kirkness (1956)Rich (1925)

Notes: 'Results based on tag recapturestResults based on ocean sampling and maturity indices^Results based on analysis of egg size§S - stream-type; O - ocean-type

estimated maturity rates for chinook aged three tofive years that were similar to, but slightly lowerthan, Cleaver's (1969) estimates for 1954-55 broodyear Columbia River chinook.

Unfortunately, the tag-return data analysed byParker and Kirkness (1956) and Cleaver (1969) donot permit separate estimates of maturity rate formales and females. Estimates based on maturityindices for fish captured in the ocean do, however,permit separate estimates of maturity by sex.These estimates are consistent with the maturationpattern established from spawning runs in thatmales show a greater proportion maturing at ayounger age than do females (Table 8). Maturityrates of 12%-36% were observed among males intheir second year of life, as were maturity rates of0%-3% for females in their second year and 0%-49%for females in their third year. Although two-year-old jack males are common in some populations,maturity rates of 12%-36% for this age group, as ageneral proposition, are unlikely, as that wouldmean two-year-old males would be a dominant agegroup among mature males in spawning runs tomost rivers rather than a minor component (Figure34). Similarly, two-year-old mature females arevirtually unknown, and three-year-old mature fe-males are uncommon in spawning runs except in a

few Oregon stocks (Nicholas and Hankin 1988); yetthe maturity rate estimates from gonad size indices(Wright and Bernhard 1972; Borque and Pitre 1972;Baranski 1979) suggest that two-year-old maturefemales should be relatively common and thatthree-year-old mature females should be a domi-nant component of spawning runs. Rich's (1925)estimates of female age-specific maturity rates,which are based on egg size rather than gonadsize, seem much more appropriate, as they indi-cate no maturation of two-year-old females, 15%maturation among three-year-old ocean-type fe-males, and high rates of maturation among four-and five-year-old females of both ocean- andstream-type races. Also, the patterns of maturityamong stream- and ocean-type females describedby Rich (1925) are consistent with the patternobserved in spawning runs (Figure 34). EvenRich's (1925) estimates, however, produce run age-composition estimates that include too manyyoung fish in comparison with observed run com-positions. Clearly, there is some fundamental prob-lem with the commonly used indices of maturityfor fish captured in the ocean, or the samples thathave been analysed to date are not representativeof the general maturity schedules for chinook. Onepossible factor contributing to the apparent high

I

vl

377


proportion of maturing fish in the ocean samplesoff southwest Vancouver Island, Washington, andin Puget Sound may be the observation by Ito et al.(1974) and Major et al. (1978) that maturing fishconstitute a higher proportion of the nearshorecatch of chinook. However, this alone is not asufficient explanation for the high proportion ofmaturing fish in the younger age groups, particu-larly the observed numbers of maturing two- andthree-year-old females.

Although most chinook do not mature until theyhave spent at least one summer at sea, a few malechinook may mature without migrating to sea(Rutter 1902; Rich 1920; Burck 1967). Such malesmay mature at the end of their first summer or,more commonly, after their second summer infresh water. It is generally the largest males in anysample that show evidence of maturity (Rich 1920).This phenomenon is documented for the Columbiaand Sacramento rivers, but probably occurs else-where as well. The proportion of stream-dwellingmales that become mature is unknown, but Rich

(1920) noted that 10%-12% of males in theMcCloud River, California, which were stream-resident, matured precociously.

Precocious maturation of males is associatedwith stream-resident populations in headwatertributaries, suggesting that it is a characteristic ofstream-type chinook. It is not known whether pre-cocious males contribute to reproduction, al-though J.W. Mullan (u.s. Fish and Wildlife Service,Leavenworth, Washington, pers. comm.) suggeststhat they do. They cannot always do so, however,as some mature outside the normal spawning timefor sea-run fish. Nor is it known for sure whetherprecocious males die after maturing. Rich (1920)claimed that at least some recovered, but Burck(1967) found that, of 259 specimens of mature pre-cocious males that he held in a downstream mi-grant trap, all died. J.W. Mullan (u.s. Fish andWildlife Service, Leavenworth, Washington, pers.comm.) suggests that most precocial age 0. malessurvive whereas most precocial 1. males die.

!!:; MORTALITY IN THE OCEAN

H .H:

Estimates of natural and fishing mortality of chi-nook are derived from several sources. Parker andKirkness (1956) were apparently the first to at-tempt such an estimate. Their estimate, based onrecaptures of fish tagged off southeastern Alaska,was 34.1% annual mortality for all age classes.

Cleaver (1969) estimated probable values of nat-ural mortality and maturity from recaptures oftagged 1961-brood chinook released from severalColumbia River hatcheries. Unfortunately, matu-rity and mortality are confounded in the data sothat it was not possible to estimate either uniquely.According to Cleaver (1969), the most realistic solu-tion to the joint estimate of maturity and mortalityoccurred when instantaneous annual natural mor-tality was 0.45 (36% annual mortality). This is veryclose to the value obtained by Parker and Kirkness(1956).

Henry (1978) explored the mortality and matu-rity schedules for 1961 and 1962 brood year Co-

lumbia River hatchery chinook. His analysisdiffered from Cleaver's (1969) in that he assumed afixed maturity schedule and calculated age-spe-cific mortality rates. Henry's (1978) estimates sug-gested some differences in natural mortalitybetween brood years and much higher naturalmortality during the first ocean year than duringlater ocean years. Although these results are intui-tively reasonable, the differences in mortality be-tween brood years could be an artifact of theassumption that maturity was fixed. There seemsto be no greater reason for assuming a fixed matu-rity schedule than for assuming a fixed mortalityschedule.

Ricker (1976) reviewed all published approachesto estimating the natural mortality rate of Pacificsalmon at sea/discussed their strengths and weak-nesses, and assessed their biases. The publishedestimates of natural mortality for chinook, accord-ing to Ricker (1976), are likely to be too large, both

378


because of tag loss and mortality of tagged fish,and because they do not take full account of mor-tality due to catch and release in some fisheries.This latter uncontrolled mortality source is partic-ularly important for chinook, most of which arecaptured as immature fish in hook-and-line fisher-ies, and for which there are size limits that forbidlanding small fish. At this point in time, it is impos-sible to state what is the real natural mortality rateof chinook salmon in the ocean. In all probability,the value is considerably less than 35% per year,probably closer to the value of 20% per year esti-mated to be a reasonable average for sockeye(Ricker 1976). It is also probable that age-specificmortality declines with age in chinook so that mostocean mortality occurs during the first year or twoof ocean life.

The available data are too scanty to determinewhether ocean mortality schedules differ amongpopulations. In addition, all of these data refer tothe ocean-type race so that no interracial compari-son may be made. Parker (1962) argued that it waslogical to suppose that causes of mortality (espe-

cially predators) were more concentrated in thecoastal zone, so that one would expect greatermortality among salmon during their residence incoastal waters. Parker (1962) further argued thatsmaller salmon should be more vulnerable to pred-ators and should, therefore, suffer greater mortalitythan larger salmon. If these arguments hold, thenocean mortality of ocean-type chinook should begreater than that of stream-type chinook becausestream-type chinook enter the ocean at a large sizeand move offshore quickly, whereas ocean-typechinook enter the ocean at a small size and spendmost of their ocean life in coastal waters. Even iftheir natural mortality rates do not differ substan-tially, I would expect fishing mortality to be lowerfor the stream-type race, at least for stream-typepopulations on the North American coast south ofcentral Alaska. These populations appear not tocontribute substantially to the Japanese high-seasfishery and, because they are distributed furtheroffshore than ocean-type chinook, they do notcontribute heavily to the intensive coastal trollfishery (Healey 1983).

HOMING AND STRAYING

Fish that survive to mature and return to freshwater to spawn must select a spawning stream,ascend it (Plate 15), and then select an appropriatespawning riffle. Upstream migration of maturechinook apparently occurs mainly during daylighthours (Plate 16), at least for the ocean-type race(Neave 1943). A few fish, however, do migrate up-stream at night.

Salmon, in general, have well-developed homingbehaviour, apparently returning to their natalstream to spawn with considerable fidelity. Thechoice of spawning river, tributary, and even riffleappears to be guided by long-term memory ofspecific odours. Groves et al. (1968) demonstratedthe apparent importance of olfaction for chinookreturning to the Spring Creek Hatchery on theColumbia River. They took chinook that had al-ready returned to the hatchery, occluded their

olfactory or visual senses, and released the treated(and untreated control fish) twenty or more ki-lometres downstream and upstream from thehatchery. The results of the study are summarizedin Table 9. About 49% of control fish returned toSpring Creek from downstream releases and 37%from upstream releases. This difference was notstatistically significant. About 23% of chinook withtheir vision destroyed found their way back toSpring Creek, significantly fewer than the controlfish (X2 = 28.7; p < .001). Only 6 of 193 (3%) chinookwith their nasal sacs plugged, however, returnedto Spring Creek, significantly less than either thecontrols (X2 = 101.8; p < .001) or chinook with theirvision destroyed (X2 = 46.1; p < .01). Chinook withboth vision and olfaction occluded were hardlyrecaptured at all, but two still found their way backto Spring Creek. The implication of these results is

379

Pacific balmon Lite Histories

TABLE 9Chinook captured at Spring Creek, tagged, and relased upstream and downstream from Spring Creek, and

recaptured at Spring Creek, other Columbia River hatcheries, and in river sport and domestic fisheries

Treatment

Number released downstreamNumber recaptured at:

Spring CreekOther hatcheriesFisheries

Number released upstreamNumber recaptured at:

Spring CreekOther hatcheriesFisheries

Control

192

94451

49

1881

Olfactoryocclusion

152

6271641

012

Visualocclusion

192

46510

49

1042

Olfactory andvisual occlusion

150

21

1541

033

Source: From Groves et al. (1968)

that both olfaction and vision are important inselection of home stream, but that olfaction is byfar the more important sense. In a related study,Hara et al. (1965) demonstrated an electrophysio-logical response of the olfactory bulb of chinook toinfusion with home-stream water, implying recog-nition of home-stream odour. The interpretation ofelectrophysiological results has been cast intodoubt, however, by Bodznick (1975) who couldfind no correlation between the electrophysiologi-cal response and the actual migratory behaviour ofsockeye salmon.

Quite a large number of the fish released byGroves et al. (1968) were recovered at hatcheriesother than Spring Creek, even among the controlfish. These cannot be regarded as true strays, how-ever, as most of the hatcheries were operatingweirs, and once the fish had entered a hatcherystream, they could not get out. Thus, the fish wereunable to correct any mistakes in homing. Sim-ilarly, the fish originally selected for the study hadbeen "trapped" by a weir at Spring Creek, andsome may not really have been Spring Creek fish.

More representative data on straying are pro-vided by Rich and Holmes (1928) and Mclsaac andQuinn (1988), who reported on the return ofadults, which had been tagged as fingerlings, tovarious Columbia River hatcheries and tributaries.None of the fish investigated by Rich and Holmes(1928) was released into a tributary from which itsparents had come, whereas Mclsaac & Quinn

(1988) investigated return rates for transplantedand resident fish. Although all recoveries weremade within the Columbia River, Rich and Holmes(1928) found that stream-type chinook showedrather poor fidelity to their particular release tribu-tary. Ocean-type chinook, on the other hand,showed very strong fidelity to their release site.Mclsaac and Quinn (1988) found that chinook froman upriver population transplanted to the Bonne-ville Hatchery returned poorly to the hatchery andthat many recaptures came from well upstream.Control fish, by contrast, homed faithfully. Richand Holmes (1928) suggested that their resultsimplied a contribution of both heredity and learn-ing to homing behaviour, but also that the suitabil-ity of the release tributary for chinook wasimportant. It may also be the case that the time ofimprinting to home-stream odour of stream-typechinook is different from that of ocean-type. Sincestream-type fish undergo several in-stream migra-tions during their freshwater life, imprinting to thehome stream may occur very early, whereas ocean-type fish may delay imprinting until just beforeocean migration. Thus, stream-type fish reared forseveral weeks in a hatchery before being releasedinto a tributary may not imprint to the tributary.Mclssac and Quinn's (1988) results, however, sug-gest much more strongly that heredity is impor-tant.

Detailed information on straying of ocean-typechinook is given by Quinn and Fresh (1984) for fish

380


from the Cowlitz River Hatchery on the ColumbiaRiver drainage. These authors found that home-stream fidelity among four brood years of chinookaveraged 98.6%, and that most strays were fromspawning areas close to the Cowlitz River. Tenchinook, however, were recovered from spawningareas in Puget Sound and the Juan de Fuca Strait.Such long-distance strays, although few innumber, could have important consequences forthe genetic composition of regional spawning pop-ulations. In contrast to the results for the CowlitzRiver, Uremovich (1977) reported considerablestraying of Elk River hatchery chinook into theadjacent Sixes River.

Among the categories of chinook returning tothe Cowlitz River, older chinook strayed more than

younger chinook, and males that had spent onlyone summer at sea were the most faithful of all.These males also returned to the river later in theseason than older age classes. The proportion ofstrays varied among brood years, and the amountof straying appeared to be related to brood yearsuccess. Higher straying was observed amongbrood years with the poorest overall returns, sug-gesting that conditions that were poor for survivalwere also poor for home-stream fidelity (Quinnand Fresh 1984). Such a behaviour pattern couldbe coincidental. It could also be a mechanism toprovide for wide distribution of spawners duringpoor survival years. Uremovich (1977) also re-ported variation in straying between years butoffered no explanation.

CONCLUDING REMARKS

This completes my summary of life history obser-vations on chinook. Throughout, I have highligh-ted the degree of variation among chinookpopulations, and, in particular, the important dif-ferences between the stream- and ocean-type ra-ces. In my view, there is a sufficient basis forseparating the species into these two races. As wasoutlined in Figure 1, the races are distinguished byfundamental ecological differences in (1) the geo-graphic distribution of their spawning popula-tions; (2) the duration of their freshwater residenceas juveniles prior to seaward migration and asadults prior to spawning; (3) their oceanic distri-bution and dispersal; and (4) timing of theirspawning migrations. There is also evidence forgenetic segregation between these life historytypes (Healey 1983; Carl and Healey 1984), formorphological differences between juveniles (Carland Healey 1984), and for the inheritance of migra-tory timing (Rich and Holmes 1928).

Assuming that there are two races of chinooksalmon, one must then ask how these could ariseand how they could persist when the races aresympatric. If it were only the case that stream-typechinook occurred in the headwater tributaries of

the larger rivers, then one might presume thatstream-dwelling behaviour had arisen indepen-dently in each river and was maintained throughdisruptive selection (Merrell 1981). The environ-mental heterogeneity producing ocean- andstream-type juvenile life history patterns could berelated to a critical time of arrival in the ocean,such as that suggested by Walters et al. (1978) forpink and chum fry from the Fraser River. If such acritical time period existed, it could render sea-ward migration from the headwaters impracticalduring the first summer because the length of timerequired for the young fish to grow to smolt sizeand traverse the river would bring them to sea at abad time. The spring and summer redistributionmigrations of stream-type chinook within the riversystem correspond in timing with the spring andsummer seaward migrations of ocean-type chi-nook. These within-river dispersals of the stream-type race could be the remnants of the seawardmigratory behaviour of the ocean-type race, andsuggest that one type may be derived from theother. Some stream-type populations may, in fact,have arisen through disruptive selection in ocean-type populations. In many instances, the different

381


ocean distributions of stream- and ocean-type chi-nook, their different migratory behaviour while atsea, and the particular geographic distribution oftheir spawning populations argue against such anexplanation for the origin of stream-type chinook.Recent information on genetic differentiationamong populations of chinook, however, does notalways reveal patterns consistent with a segrega-tion into stream- and ocean-type races (Gharrett etal. 1987; Utter et al. 1989; Winans 1989).

An alternative explanation is that stream-typechinook developed (or persisted) in the Beringrefugium, or on the Asian coast, during the lastglaciation, whereas ocean-type chinook developed(or persisted) south of the glaciation on the NorthAmerican coast. With the retreat of the ice, bothraces could have expanded to occupy their presentdistributions. Gharrett et al. (1987) suggested thatthe genetic composition of chinook populationssupports this interpretation. For this explanationto hold, however, selection gradients or reproduc-tive isolating mechanisms must be sufficient toprevent introgression in the natural populations,at least where they are sympatric. It is also neces-sary to postulate some barrier preventing signifi-cant invasion of the ocean-type race north of 56°N.Conditions in the freshwater environment mightprovide such a barrier if, for example, riverine andestuarine productivity and temperature were toolow to permit chinook to reach a critical size forsmolting during their first summer. Riverine tem-peratures, at least during the summer growingseason, however, appear adequate for good growth(Healey 1983). Although I am confident that theexplanation for the absence of ocean-type chinooknorth of 56°N lies in the trade-off between survivaland growth for small chinook in the river versus inthe ocean, the precise mechanism preventingnorthward invasion of ocean- type chinook is notapparent.

These possible explanations, and others, for theexistence of stream- and ocean-type chinook needto be critically tested, and they offer an opportu-nity for fruitful research into the evolution of Pa-cific salmon. There is no doubt, however, thatPacific salmon have the capability to adapt quicklyto new opportunities. The recent appearance inthe Great Lakes of spring-spawning chinook,which must have developed from the fall-spawn-ing introduced race, attests to this (Kwain and

Thomas 1984), as does the development of even-year runs among pink salmon from a single odd-year release into the Great Lakes (Kwain andChappel 1978).

Segregating the species into stream- and ocean-type races accounts for only a small part of theinterpopulation variation in chinook. There is alsoconsiderable interpopulation (within race) and in-trapopulation variation that requires explanation(Healey and Heard 1984). Interpopulation varia-tion includes (1) differences in the proportion ofthe population migrating seaward at different ages(fry or fingerlings for ocean-type chinook, one- ortwo-year-old smolts for stream-type); (2) differen-ces in the proportion maturing at each reproduc-tive age; (3) differences in growth; and (4)differences in fecundity. Intrapopulation variationinvolves the same population attributes and leadsto a variety of life history tactics within any popu-lation. Variation in length of riverine and estuarineresidence has attracted considerable attention andhas been interpreted as reflecting life history adap-tation (Rich 1925; Reimers 1971; Healey 1980b,1982b; Levy and Northcote 1981). Differences inthe relative length of riverine and estuarine resi-dence by chinook in the Sixes River, Oregon, ledReimers (1971) to postulate five life history types ofchinook in that system. From analysis of adultscales, Reimers (1971) concluded that only one ofthese life history types contributed most to thespawning population in the Sixes River. Healey(1980b, 1982b) recognized two distinct behaviourpatterns among ocean-type chinook in VancouverIsland rivers: migration to the river estuary imme-diately after fry emergence in the spring, and mi-gration to the estuary four to six weeks after

" emergence.The existence of this degree of variation indi-

cates considerable plasticity in the species.Whether this plasticity is a consequence of genetic

. polymorphism, or is largely environmentally in-duced, remains to be demonstrated. Carl and Hea-ley (1984), however, found genetic differencesbetween fry and fingerling migrants in the Na-naimo River on Vancouver Island, suggesting agenetic basis for the two behaviour patterns ofocean-type chinook recognized by Healey (1980b,1982b).

Despite the apparent biological and ecologicalplasticity of the species, the possibility that many

382


populations may be specifically adapted to localconditions cannot be discounted on the basis ofpresent evidence. Such adaptation could have pro-duced unique gene pools in some instancesthrough the fixation of successful mutations.Given the intrapopulation variability of chinook,however, it seems more likely that local adapta-tions are due to the development of effective genecombinations rather than to fixation of mutations.Furthermore, it suggests that chinook should beable to adapt rapidly to new situations. The suc-cess of hatcheries in producing chinook on thePacific coast, and the success of chinook trans-plants to New Zealand, Chile, and the Great Lakes

testify to the adaptability of the species.The extent of intrapopulation variation further

suggests a degree of "bet hedging" in the life his-tory strategy of the chinook salmon (Stearns 1976).Such a tactic may be very appropriate for a specieslike chinook, which occurs in many small, possiblylocally adapted, spawning populations. Having avariety of life history tactics would serve to spreadthe risk of mortality across a number of habitatsand thus reduce the probability of complete failureof a year class. Indeed, this strategy may partlyexplain why chinook have been able to persist inthe face of continued heavy fishing pressure and,in some systems, significant habitat modification.

REFERENCES

Alderdice, D.F., and F.P.J. Velsen. 1978. Relationbetween temperature and incubation time foreggs of chinook salmon (Oncorhynchus tshawyt-scha). J. Fish. Res. Board Can. 35:69-75

Argue, A.W. 1970. A study of factors affecting ex-ploitation of Pacific salmon in the Canadiangauntlet fishery of Juan de Fuca Strait. Fish.Serv. (Can.) Pac. Reg. Tech. Rep. 1970-11:259 p.

Argue, A.W., and D.E. Marshall. 1976. Size and ageof chinook and coho salmon for subdivisions ofthe Strait of Georgia troll fishery, 1966. Fish. Mar.Serv. (Can.) Tech. Rep. PAC/T-76-18:175 p.

Aro, K.V. 1972. Recoveries of salmon tagged off-shore in the North Pacific Ocean by Japan andthe United States in 1970 and 1971, and addi-tional recoveries from earlier taggirigs by Can-ada, Japan, and the United States. Fish. Res.Board Can. MS Rep. Ser. 1186:31 p.

. Recoveries of salmon tagged offshore in theNorth Pacific Ocean by Japan and the UnitedStates in 1972 and 1973, and additional recover-ies from earlier laggings by Canada, Japan, andthe United States. Fish. Res. Board Can. MS Rep.Ser. 1328:20 p.

Aro, K.V., and M.P. Shepard. 1967. Pacific salmon inCanada, p. 225-327 In: Salmon of the North Pa-cific Ocean. Part iv. Spawning populations of

North Pacific salmon. Int. North Pac. Fish.Comm. Bull. 23

Aro, K.V., J.A. Thomson, and D.P. Giovando. 1971.Recoveries of salmon tagged offshore in theNorth Pacific Ocean by Canada, Japan, and theUnited States, 1956 to 1969. Fish. Res. BoardCan. MS Rep. Ser. 1147:493 p.

Atkinson, C.E., J.H. Rose, and T.O. Duncan. 1967.Pacific salmon in the United States, p. 43-223. In:Salmon of the North Pacific Ocean. Part iv.Spawning populations of North Pacific salmon.Int. North Pac. Fish. Comm. Bull. 23

Ball, E.A.R., and H. Godfrey. 1968a. Lengths andages of chinook salmon taken in the British Co-lumbia troll fishery and the Fraser River gillnetfishery in 1965. Fish. Res. Board Can. MS Rep.Ser. 952:130 p.

. 1968b. Lengths and ages of chinook salmontaken in the British Columbia troll fishery andthe Fraser River gill net fishery in 1966. Fish. Res.Board Can. MS Rep. Ser. 954:143 p.

-. 1969. Lengths and ages of chinook salmontaken in the British Columbia troll fishery in1968. Fish. Res. Board Can. MS Rep. Ser. 1073:41 p.

-. 1970. Lengths and ages of chinook salmontaken in the British Columbia troll fishery in

383


1969. Fish. Res. Board Can. MS Rep. Ser. 1121:101 p.

Baranski, C. 1979. Maturity rates for Puget Soundchinook stocks. Wash. Dep. Fish. Tech. Rep.43:13 p.

Becker, C.D. 1973. Food and growth parameters ofjuvenile chinook salmon, Oncorhynchus tshawyt-scha, in central Columbia River. Fish. Bull, (u.s.)71:387-400

Becker, C.D., D.A. Neitzel, and C.S. Abernethy.1983. Effects of dewatering on chinook salmonredds: tolerance of four development phases toone-time dewatering. N. Am. J. Fish. Manage.3:373-382

Becker, C.D., D.A. Neitzel, and D.H. Fickeisen.1982. Effects of dewatering on chinook salmonredds: tolerance of four development phases todaily dewaterings. Trans. Am. Fish. Soc. Ill:624-637

Bell, R. 1958. Time, size, and estimated numbers ofseaward migrants of chinook salmon and steel-head trout in the Brownlee-Oxbow section of themiddle Snake River. State of Idaho Departmentof Fish and Game, Boise, ID. 36 p.

Birtwell, I.K. 1978. Studies on the relationship be-tween juvenile chinook salmon and water qual-ity in the industrialized estuary of the SomassRiver, p. 58-78. In: B.C. Shepherd and R.M.J.Ginetz (rapps.). Proceedings of the 1977 North-east Pacific Chinook and Coho Salmon Work-shop. Fish. Mar. Serv. (Can.) Tech. Rep. 759:164 p.

Bjornn, T.C. 1971. Trout and salmon movements intwo Idaho streams as related to temperature,food, stream flow, cover and population density.Trans. Am. Fish. Soc. 100:423-438

Bodznick, D. 1975. The relationship of the olfactoryEEC evoked by naturally-occurring stream wa-ters to the homing behaviour of sockeye salmon(Oncorhynchus nerka Walbaum). Comp. Biochem.Physiol. 52 (3A):487-495

Borque, S.C., and K.R. Pitre. 1972. Size and matu-rity of troll chinook salmon (Oncorhynchus tsha-wytscha) caught off the west coast of VancouverIsland in 1969 and 1970. Fish. Mar. Serv. (Can.)Pac. Reg. Tech. Rep. 1972-7:44 p.

Bottom, D. 1984. Food habits of juvenile chinook inSixes estuary, p. 15-21. In: Research and devel-opment of Oregon's coastal chinook stocks: an-nual progress report. Oregon Department of

384

Fish and Wildlife, Portland, ORBovee, K.D. 1978. Probability-of-use criteria fo

the family Salmonidae. u.s. Fish Wildl. Servpws/OBS-78/07; Instream Flow Inf. Paper 480 p.

Brady, J.A. 1983. Lower Yukon River salmon tesland commercial fisheries, 1981. Alaska Dep. FisrGame ADF&G Tech. Data Rep. 89:91 p.

Briggs, J.C. 1953. The behaviour and reproductionof salmonid fishes in a small coastal stream,Calif. Dep. Fish Game Fish. Bull. 94:62 p.

Brown, R.L. 1986.1984 annual report of the Intera-gency Ecological Studies Program for the Sacra-mento-San Joaquin estuary, u.s. Fish WildlifeService, Stockton, CA. 133 p.

Burck, W.A. 1967. Mature stream-reared springchinook salmon. Res. Briefs Fish Comm. Oreg.13(1):128 p.

Burck, W.A., and P.E. Reimers. 1978. Temporal andspatial distribution of fall chinook spawning inElk River. Oreg. Dep. Fish Wildl. Info. Rep. Ser.Fish. 78-3:14 p.

Burner, C.J. 1951. Characteristics of spawningnests of Columbia River salmon. Fish. Bull. FishWildl. Serv. 61:97-110

Carl, L.M. 1982. Natural reproduction of cohosalmon and chinook salmon in some Michiganstreams. N. Am. J. Fish. Manage. 2:375-380

Carl, L.M., and M.C. Healey. 1984. Differences inenzyme frequency and body morphology amongthree juvenile life history types of chinooksalmon (Oncorhynchus tshawytscha) in the Na-naimo River, British Columbia. Can. J. Fish.Aquat. Sci. 41:1070-1077

Chapman, D.W., and T.C. Bjornn. 1969. Distribu-tion of salmonids in streams, with special refer-ence to food and feeding, p. 153-176. In: T.G.Northcote (ed.). Symposium on Salmon andTrout in Streams. H.R. MacMillan Lectures inFisheries. Institute of Fisheries, University ofBritish Columbia, Vancouver, BC. 388 p.

Chapman, D.W., D.E. Weitcamp, T.L. Welsh, M.B.Dell, and T.H. Schadt. 1986. Effects of river flowon the distribution of chinook salmon redds.Trans. Am. Fish. Soc. 115:537-547

Chapman, W.M. 1943. The spawning of chinooksalmon in the main Columbia River. Copeia1943:168-170

Chapman, W.M., and E. Quistdorff. 1938. The foodof certain fishes of north central Columbia River


drainage, in particular, young chinook salmonand steelhead trout. Wash. Dep. Fish. Biol. Rep.37-A:l-14

Clark, G.H. 1929. Sacramento-San Joaquin salmon(Oncorhynchus tschawytscha) [sic] fishery of Cali-fornia. Calif. Dep. Fish Game Fish. Bull. 17:73 p.

Clarke, W.C., and J.E. Shelbourn. 1985. Growth anddevelopment of seawater adaptability by juve-nile fall chinook salmon (Oncorhynchus tshawyt-scha) in relation to temperature. Aquaculture45:21-31

Clarke, W.C., J.E. Shelbourn, T. Ogasawara, and T.Hirano. 1989. Effect of initial day length ongrowth, sea water adaptability and plasmagrowth hormone levels in underyearling coho,chinook, and chum salmon. Aquaculture82:51-62

Cleaver, F.C. 1969. Effects of ocean fishing on the1961-brood fall chinook salmon from ColumbiaRiver hatcheries. Res. Rep. Fish Comm. Oreg.1:76 p.

Clemens, W.A. 1934. The food of young springsalmon in Shuswap Lake, B.C. Can. Field-Nat,48:142 p.

Clemens, W.A., and J.A. Munro. 1934. The food ofthe squawfish. Biol. Board Can. Prog. Rep. Pac.Biol. Stn. Nanaimo Fish. Exp. Stn. (Pac.) 19:3-4

Collings, M.R., R.W. Smith, and G.T. Higgins. 1972.Hydrology of four streams in western Washing-ton as related to several Pacific salmon species:Humptulips, Elchoman, Green and Wynoocheerivers: open file report. United States GeologicalSurvey, Tacoma, WA. 128 p.

Congleton, J.L., S.K. Davis, and S.R. Foley. 1981.Distribution, abundance, and outmigration tim-ing of chum and chinook salmon fry in theSkagit salt marsh, p. 153-163. In: E.L. Brannonand E.O. Salo (eds.). Proceedings of the Salmonand Trout Migratory Behaviour Symposium.School of Fisheries, University of Washington,Seattle, WA

Conrad, J.F. 1965. Observations on "bad eggs" inColumbia River fall chinook salmon. Prog. Fish-Cult. 27:42-44

Coombs, B.D., and R.E. Burrows. 1957. Thresholdtemperatures for the normal development of chi-nook salmon eggs. Prog. Fish-Cult. 19:3-6

Coots, M. 1957. The spawning efficiency of kingsalmon (Oncorhynchus tshawytscha) in Fall Creek,Siskiyou County: 1954-55 investigations. Calif.

Dep. Fish Game Inland Fish. Admin. Rep.57-1:15 p.

Craddock, D.R., T.H. Blahm, and W.D. Parente.1976. Occurrence and utilization of zooplanktonby juvenile chinook salmon in the lower Colum-bia River. Trans. Am. Fish. Soc. 105:72-76

Cramer, S.P., and J.A. Lichatowich. 1978. Factorsinfluencing the rate of downstream migration ofjuvenile chinook salmon in the Rogue River, p.43-48. In: B.C. Shepherd and R.M.J. Ginetz(rapps.). Proceedings of the 1977 Northeast Pa-cific Chinook and Coho Salmon Workshop. Fish.Mar. Serv. (Can.) Tech. Rep. 759:164 p.

Dahlberg, M.L. 1982. Report of incidence of coded-wire tagged salmonids in catches of foreign com-mercial and research vessels operating in theNorth Pacific Ocean and Bering Sea during 1980-1982. (Submitted to annual meeting InternationalNorth Pacific Fisheries Commission, Tokyo, Ja-pan, November 1982.) Auke Bay Laboratory,Northwest and Alaska Fisheries Center, Na-tional Marine Fisheries Service, Auke Bay, AK.lip.

Dahlberg, M.L., and S. Fowler. 1985. Report ofincidence of coded-wire-tagged salmonids incatches of foreign commercial and research ves-sels operating in the North Pacific Ocean andBering Sea during 1984-85. (Submitted to an-nual meeting International North Pacific Fisher-ies Commission, Tokyo, Japan, November 1985.)Auke Bay Laboratory, Northwest and AlaskaFisheries Center, National Marine Fisheries Ser-vice, Auke Bay, AK. 16 p.

Dahlberg, M.L., P.P. Thrower, and S. Fowler. 1986.Incidence of coded-wire-tagged salmonids incatches of foreign commercial and research ves-sels operating in the North Pacific Ocean andBering Sea in 1985-1986. (Submitted to annualmeeting International North Pacific FisheriesCommission, Anchorage, Alaska, November1986.) Auke Bay Laboratory, Northwest andAlaska Fisheries Center, National Marine Fisher-ies Service, Auke Bay, AK. 26 p.

Don Chapman Consultants. 1989. Summer andwinter ecology of juvenile chinook salmon andsteelhead trout in the Wenatchee River, Wash-ington. Chelan County Public Utility, Wenat-chee, WA. 301 p.

Dunford, W.E. 1975. Space and food utilization bysalmonids in marsh habitats of the Fraser River

385


estuary. M.Sc. thesis. University of British Co-lumbia, Vancouver, BC. 81 p.

Edmundson, E., F.E. Everest, and D.W. Chapman.1968. Permanence of station in juvenile chinooksalmon and steelhead trout. J. Fish. Res. BoardCan. 25:1453-1464

Eisler, R., and R.C. Simon. 1961. Destruction ofsalmon larvae by Hydra oligactis. Trans. Am. Fish.Soc. 90:329-332

Everest, F.H., and D.W. Chapman. 1972. Habitatselection and spatial interaction by juvenile chi-nook salmon and steelhead trout in two Idahostreams. J. Fish. Res. Board Can. 29:91-100

Fedorenko, A.Y., F.J. Fraser, and D.T. Lightly. 1979.A limnological and salmonid resource study ofNitinat Lake: 1975-1977. Fish. Mar. Serv. (Can.)Tech. Rep. 839:86 p.

Fisher, J.P., W.G. Pearcy, and A.W. Chung. 1983.Studies of juvenile salmonids off the Oregon andWashington coast, 1982. Oreg. State Univ. Coll.Oceanog. Cruise Rep. 83-2:41 p.

. 1984. Studies of juvenile salmonids off theOregon and Washington coast, 1983. Oreg. StateUniv. Coll. Oceanog. Cruise Rep. 83-2; Oreg.State Univ. Sea Grant Coll. Program ORESU-T-85-004:29 p.

Foerster, R.E., and W.E. Ricker. 1941. The effect ofreduction of predaceous fish on survival ofyoung sockeye salmon at Cultus Lake. J. Fish.Res. Board Can. 5:315-336

Fowler, L. 1972. Growth and mortality of fingerlingchinook salmon as affected by egg size. Prog.Fish-Cult. 34:66-69

Fraser, C. McL. 1921. Further studies on thegrowth rate in Pacific salmon. Contrib. Can. Biol.1918-1920:7-27

Fraser, F.J., P.J. Starr, and A.Y. Fedorenko. 1982. Areview of the chinook and coho salmon of theFraser River. Can. Tech. Rep. Fish. Aquat. Sci.1126:130 p.

Fry, D.H., Jr., and E.P. Hughes. 1951. The Californiasalmon troll fishery. Pac. Mar. Fish. Comm. Bull.2:7-42

Galbreath, J.L., and R.L. Ridenhour. 1964. Fecun-dity of Columbia River chinook salmon. Res.Briefs Fish Comm. Oreg. 10(l):16-27

Gangmark, H.A., and R.G. Bakkala. 1960. A com-parative study of unstable and stable (artificalchannel) spawning streams for incubating kingsalmon at Mill Creek. Calif. Fish Game

46:151-164Gangmark, H.A., and R.D. Broad. 1955. Experimen-

tal hatching of king salmon in Mill Creek> atributary of the Sacramento River. Calif. FishGame 41:233-242

Gebhards, S.V. 1961. Emergence and mortality ofchinook salmon fry in a natural redd. Prog. Fish-Cult. 23:91

Gharrett, A.J., S.M. Shirley, and G.R. Tromble. 1987.Genetic relationships among populations of Al-askan chinook salmon (Oncorhynchus tshawyt-scha). Can. J. Fish. Aquat. Sci. 44:765-774

Gilbert, C.H. 1913. Age at maturity of the Pacificcoast salmon of the genus Oncorhynchus. Bull.Bur. Fish, (u.s.) 32:1-22

Ginetz, R.M.J. 1976. Chinook salmon in the northcoastal division. Fish. Mar. Serv. (Can.) Pac. Reg.Tech. Rep. PAC/T-76-12.-96 p.

Godfrey, H. 1968a. Ages and physical characteris-tics of maturing chinook salmon of the Nass,Skeena, and Fraser Rivers in 1964, 1965, and1966. Fish. Res. Board Can. MS Rep. Ser. 967:38 p.

. 1968b. Review of information obtained fromthe tagging and marking of chinook and cohosalmon in coastal waters of Canada and theUnited States. Fish. Res. Board Can. MS Rep. Ser.953:172 p.

Godfrey, H.D., D. Worlund, and H.T. Bilton. 1968.Tests on the accuracy of ageing chinook salmon(Oncorhynchus tshawytscha) from their scales. J.Fish. Res. Board Can. 25:1971-1982

Goodman, D. 1975. A synthesis of the impacts ofproposed expansion of the Vancouver Interna-tional Airport and other developments on thefisheries resources of the Fraser River estuary.Vol. I and II, Section II. In: Fisheries resourcesand food web components of the Fraser Riverestuary and an assessment of the impacts ofproposed expansion of the Vancouver Interna-tional Airport and other developments on theseresources. Prepared by Department of Environ-ment, Fisheries and Marine Service. Environ-ment Canada, Vancouver, BC

Gordon, D.K., and C.D. Levings. 1984. Seasonalchanges of inshore fish populations on Sturgeonand Roberts Bank, Fraser River estuary, BritishColumbia. Can. Tech. Rep. Fish. Aquat. Sci.1240:81 p.

Grachev, L.E. 1967. Growth rate of Kamchatkachinook salmon. Izv. Tikhookean. Nauchno-

386


Issled. Inst. Rybn. Khoz. Okeanogr. 57:89-97.(Transl. from Russian; National Marine FisheriesService, Northwest Fisheries Center, Seattle, WA)

Groves, A.B., G.B. Collins, and P.S. Trefethen. 1968.Roles of olfaction and vision in choice of spawn-ing site by homing adult chinook salmon (Oncor-hynchus tshawytscha). J. Fish. Res. Board Can.25:867-876

Hallock, R.J., and D.H. Fry, Jr. 1967. The five spe-cies of salmon, Oncorhynchus, in the SacramentoRiver, California. Calif. Fish Game 53:5-22

Hallock, R.J., D.H. Fry, Jr., and D.A. LaFaunce.1957. The use of wire fyke traps to estimate theruns of adult salmon and steelhead in the Sacra-mento River. Calif. Fish Game 43:271-298

Hankin, D.G., and M.C. Healey. 1986. Dependenceof exploitation rates for maximum yield andstock collapse on age and sex structure of chi-nook salmon (Oncorhynchus tshawytscha) stocks.Can. J. Fish. Aquat. Sci. 43:1746-1759

Hara, T.J., K. Ueda, and A. Gorbman. 1965. Elec-troencephalographic studies of homing salmon.Science 149:884-885

Hart, J.L. 1973. Pacific fishes of Canada. Bull. Fish.Res. Board Can. 180:740 p.

Hartt, A.C. 1966. Migrations of salmon in the NorthPacific Ocean and Bering Sea as determined byseining and tagging, 1959-1960. Int. North Pac.Fish. Comm. Bull. 19:141 p.

. 1980. Juvenile salmonids in the oceanic eco-system; the critical first summer, p. 25-57. In:W.J. McNeil and D.C. Himsworth (eds.). Sal-monid ecosystems of the North Pacific. OregonState University Press, Corvallis, OR

Hartt, A.C., and M.B. Dell. 1986. Early oceanicmigrations and growth of juvenile Pacific salmonand steelhead trout. Int. North Pac. Fish. Comm.Bull. 46:105 p.

Healey, M.C. 1976. Herring in the diets of Pacificsalmon in Georgia Strait. Fish. Res. Board Can.MS Rep. Ser. 1382:38 p.

. 1980a. The ecology of juvenile salmon inGeorgia Strait, British Columbia, p. 203-229. In:W.J. McNeil and D.C. Himsworth (eds.). Sal-monid ecosystems of the North Pacific. OregonState University Press, Corvallis, OR

•. 1980b. Utilization of the Nanaimo Riverestuary by juvenile chinook salmon, Oncorhyn-chus tshawytscha. Fish. Bull, (u.s.) 77:653-668

~ - . 1982a. Catch, escapement and stock-re-

cruitment for British Columbia chinook salmonsince 1951. Can. Tech. Rep. Fish. Aquat. Sci.1107:77 p.

-. 1982b. Juvenile Pacific salmon in estuaries:the life support system, p. 315-341. In: V.S.Kennedy (ed.). Estuarine comparisons. Aca-demic Press, New York, NY

-. 1982c. The distribution and residency ofjuvenile Pacific salmon in the Strait of Georgia,British Columbia, in relation to foraging success,p. 61-69. In: B.R. Melteff and R.A. Neve (eds.).Proceedings of the North Pacific AquacultureSymposium. Alaska Sea Grant Rep. 82-2

. 1983. Coastwide distribution and ocean mi-gration patterns of stream- and ocean-type chi-nook salmon, Oncorhynchus tshawytscha. Can.Field-Nat. 97:427-433

-. 1986. Regional and seasonal attributes ofcatch in the British Columbia troll fishery. Can.Tech. Rep. Fish. Aquat. Sci. 1494:65 p.

Healey, M.C., and C. Groot. 1987. Marine migrationand orientation of ocean-type chinook and sock-eye salmon, p. 298-312. In: M.J. Dadswell, R.J.Klanda, C.M. Moffitt, R.L. Saunders, R.A. Rulif-son, and J.E. Cooper (eds.). Common strategiesof anadromous and catadromous fishes. Am.Fish. Soc. Symp. 1

Healey, M.C, and W.R. Heard. 1984. Inter- andintra-population variation in the fecundity ofchinook salmon (Oncorhynchus tshawytscha) andits relevance to life history theory. Can. J. Fish.Aquat. Sci. 41:476-483

Healey, M.C., and P.P. Jordan. 1982. Observationson juvenile chum and chinook and spawningchinook in the Nanaimo River, British Columbia,during 1975-1981. Can. MS Rep. Fish. Aquat. Sci.1659:31 p.

Healey, M.C., R.V. Schmidt, P.P. Jordan, and R.M.Hungar. 1977. Juvenile salmon in the Nanaimoarea 1975. 2: length, weight, and growth. Fish.Mar. Serv. (Can.) MS Rep. 1438:147 p.

Heg, R., and J. Van Hyning. 1951. Food of thechinook and silver salmon taken off the Oregoncoast. Fish Comm. (Oreg.) Res. Briefs 3(2):32-40

Henry, K.A. 1978. Estimating natural and fishingmortalities of chinook salmon, Oncorhynchustshawytscha, in the ocean, based on recoveries ofmarked fish. Fish. Bull, (u.s.) 76:45-57

Herrmann, R.B. 1970. Food of juvenile chinook andchum salmon in the lower Chehalis River and

387


upper Grays Harbor, p. 59-82. In: Grays Harborcooperative water quality study 1964-1966.Wash. Dep. Fish. Tech. Rep. 7

Hoover, E.E. 1936. Contributions to the life historyof the chinook and landlocked salmon in NewHampshire. Copeia 1936:193-198

Hunter, J.G. 1959. Survival and production of pinkand chum salmon in a coastal stream. J. Fish.Res. Board Can. 16:835-886

International North Pacific Fisheries Commission(INPFC). 1972-1982. Statistical yearbooks 1970-1979. International North Pacific Fisheries Com-mission, Vancouver, BC

Irvine, J.R. 1986. Effects of varying discharge on thedownstream movement of salmon fry, (Oncor-hynchus tshawytscha) Walbaum. J. Fish Biol.28:17-28

Ito, J., Y. Ishida, and S. Ito. 1985. Stock identifica-tion of chinook salmon in offshore waters in 1974based on scale pattern analysis. (Submitted tothe annual meeting International North PacificFisheries Commission, Tokyo, Japan, November1985.) Fisheries Agency of Japan, Tokyo, Japan.14 p.

. 1986. Further analysis of stock identificationof chinook salmon in offshore waters in 1974.(Submitted to the annual meeting InternationalNorth Pacific Fisheries Commission, Anchorage,Alaska, November 1986.) Fisheries Agency ofJapan, Tokyo, Japan. 9 p.

Ito, J., K. Takagi, and S. Ito. 1974. The identificationof maturing and immature chinook salmon, On-corhynchus tshawytscha (Walbaum), in the off-shore stage and some related information. Bull.Far Seas Fish. Res. Lab. 11:67-75

Jefferts, K.B., P.K. Bergman, and H.F. Fiscus. 1963.A coded wire identification system for macro-organisms. Nature 198:460-462

Kissner, P.D., Jr. 1973. Annual progress report for astudy of chinook salmon in southeast Alaska.Study AFs-41-1 Alaska Department of Fish andGame, Division of Sport Fish, Juneau, AK. 24 p.

Kjelson, M.A., P.P. Raquel, and F.W. Fisher. 1981.Influences of freshwater inflow on chinooksalmon (Oncorhynchus tshawytscha) in the Sacra-mento-San Joaquin estuary, p. 88-102. In: R.D.Cross and D.L. Williams (eds.). Proceedings ofthe National Symposium on Freshwater Inflowto Estuaries, u.s. Fish Wildl. Serv. Biol. Serv.Prog. Fws/OBS-81/04(2)

. 1982. Life history of fall-run juvenile chi-nook salmon, Oncorhynchus tshawytscha, in theSacramento-San Joaquin estuary, California, p.393-411. In: V.S. Kennedy (ed.). Estuarine com-parisons. Academic Press, New York, NY

Knudsen, C.M., C.K. Harris, and N.D. Davis. 1983.Origins of chinook salmon in the area of theJapanese mothership and landbased driftnetsalmon fisheries in 1980. Univ. Wash. Fish. Res.Inst. FRi-uw-8315:71 p.

Kondo, H., Y. Hirano, N. Nakayama, and M.Miyaki. 1965. Offshore distribution and migra-tion of Pacific salmon (genus Oncorhynchus)based on tagging studies (1958-1961). Int. NorthPac. Fish. Comm. Bull. 17:213 p.

Kwain, W., and J.A. Chappel. 1978. First evidencefor even-year spawning pink salmon, Oncorhyn-chus gorbuscha, in Lake Superior. J. Fish. Res.Board Can. 35:1373-1376

Kwain, W., and E. Thomas. 1984. The first evidenceof spring spawning by chinook salmon in LakeSuperior. N. Am. J. Fish. Manage. 4:227-228

Levings, C.D. 1982. Short term use of a low tiderefuge in a sandflat by juvenile chinook, Oncor-hynchus tshawytscha, Fraser River estuary. Can.Tech. Rep. Fish. Aquat. Sci. 1111:33 p.

Levings, C.D., C.D. McAllister, and B.D. Chang.1986. Differential use of the Campbell River estu-ary, British Columbia, by wild and hatchery-reared juvenile chinook salmon (Oncorhynchustshawytscha). Can. J. Fish. Aquat. Sci. 43:1386-1397

Levy, D.A., and T.G. Northcote. 1981. The distribu-tion and abundance of juvenile salmon in marshhabitats of the Fraser River estuary. WestwaterRes. Cent. Univ. Br. Col. Tech. Rep. 25:117 p.

— . 1982. Juvenile salmon residency in a marsharea of the Fraser River estuary. Can. J. Fish.Aquat. Sci. 39:270-276

Levy, D.A., T.G. Northcote, and G.J. Birch. 1979.Juvenile salmon utilization of tidal channels inthe Fraser River estuary, British Columbia. West-water Res. Cent. Univ. Br. Col. Tech. Rep. 23:70p.

Lindbergh, J.M. 1982. A successful transplant ofPacific salmon to Chile. Proc. Gulf Caribb. Fish.Inst. 34:81-87

Lister, D.B., and H.S. Genoe. 1970. Stream habitatutilization by cohabiting under yearlings of chi-nook (Oncorhynchus tshawytscha) and coho (O.

388


kisutch) salmon in the Big Qualicum River, Brit-ish Columbia. J. Fish. Res. Board Can. 27:1215-1224

Lister, D.B., and C.E. Walker. 1966. The effect offlow control on freshwater survival of chum,coho, and chinook salmon in the Big QualicumRiver. Can. Fish Cult. 37:3-25

Lister, D.B., C.E. Walker, and M.A. Giles. 1971.Cowichan River chinook salmon escapementsand juvenile production 1965-1967. Fish. Serv.(Can.) Pac. Reg. Tech. Rep. 1971-3: 8 p.

Loeffel, R.E., and H.O. Wendler. 1969. Review ofthe Pacific coast chinook and coho salmon re-sources with special emphasis on the trollfishery, p. 1-107. In: Informal Committee on Chi-nook and Coho. Reports by the United Statesand Canada on the status, ocean migrations, andexploitation of northeast Pacific stocks of chi-nook and coho salmon, to 1964. Vol. 1: report bythe United States Section.

Loftus, W.F., and H.L. Lenon. 1977. Food habits ofthe salmon smolts, Oncorhynchus tshawytscha andO. keta, from the Salcha River, Alaska. Trans. Am.Fish. Soc. 106:235-240

Mains, E.M., and J.M. Smith. 1964. The distribu-tion, size, time, and current preferences of sea-ward migrant chinook salmon in the Columbiaand Snake Rivers. Wash. Dep. Fish. Fish. Res.Pap. 2(3):5-43

Major, R.L., J. Ito, S. Ito, and H. Godfrey. 1978.Distribution and origin of chinook salmon (On-corhynchus tshawytscha) in offshore waters of theNorth Pacific Ocean. Int. North Pac. Fish.Comm. Bull. 38:54 p.

Major, R.L., and J.L. Mighell. 1969. Egg-to-migrantsurvival of spring chinook salmon (Oncorhynchustshawytscha) in the Yakima River, Washington.Fish. Bull. (u.s.) 67:347-359

Major, R.L., S. Murai, and J. Lyons. 1977. Scalestudies to identify Asian and western Alaskanchinook salmon. Int. North Pac. Fish. Comm.Annu. Rep. 1975:68-71

Manzer, J.I., T. Ishida, A.E. Peterson, and M.G.Hanavan. 1965. Salmon of the North PacificOcean. Part V. Offshore distribution of salmon.Int. North Pac. Fish. Comm. Bull. 15:452 p.

McBride, D.N., H.H. Hamner, and L.S. Buklis. 1983.Age, sex, and size of Yukon River salmon catchand escapement, 1982. Alaska Dep. Fish GameADF&G Tech. Data Rep. 90:141 p.

McGregor, E.A. 1922. Observations on the eggyield of Klamath River king salmon. Calif. FishGame 8:160-164

. 1923. A possible separation of the riverraces of king salmon in ocean-caught fish bymeans of anatomical characters. Calif. Fish Game9:138-150

Mclsaac, D.O., and T.P. Quinn. 1988. Evidence for ahereditary component in homing behavior ofchinook salmon (Oncorhynchus tshawytscha). Can.J. Fish. Aquat. Sci. 45:2201-2205

McLeod, C.L., and J.P. O'Neil. 1983. Major rangeextensions of anadromous salmonids and firstrecord of chinook salmon in the Mackenzie Riverdrainage. Can. J. Zool. 61:2183-2184

McPhail, J.D., and C.C. Lindsey. 1970. Freshwaterfishes of northwestern Canada and Alaska. Bull.Fish. Res. Board Can. 173:381 p.

Meehan, W.R., and D.B. Siniff. 1962. A study of thedownstream migrations of anadromous fishes inthe Taku River, Alaska. Trans. Am. Fish. Soc.91:399-407

Merkel, T.J. 1957. Food habits of the king salmon,Oncorhynchus tshawytscha (Walbaum), in the vi-cinity of San Francisco, California. Calif. FishGame 43:249-270

Merrell, D.J. 1981. Ecological genetics. University ofMinnesota Press, Minneapolis, MN. 500 p.

Miller, D.R., J.G. Williams, and C.W. Sims. 1983.Distribution, abundance, and growth of juvenilesalmonids off the coast of Oregon and Washing-ton, summer 1980. Fish. Res. 2:1-17

Milne, D.J. 1955. Selectivity of trolling lures. Fish.Res. Board Can. Prog. Rep. Pac. Coast Stn.103:3-5

. 1964. Sizes and ages of chinook, Oncorhyn-chus tshawytscha, and coho, O. kisutch, salmon inthe British Columbia troll fisheries (1952-1959)and the Fraser River gill-net fishery (1956-1959).Fish. Res. Board Can. MS Rep. Ser. 776:42 p.

Mundie, J.H. 1969. Ecological implications of thediet of juvenile coho in streams, p. 135-152. In:T.G. Northcote (ed.). Symposium on Salmon andTrout in streams. H.R. MacMillanJ-,ectures in—Fisheries. Institute of Fisheries, University ofBritish Columbia, Vancouver, BC

Murphy, M.L., J. Heifetz, J.F. Thedinga, S.W. John-son, and K.V. Koski. 1989. Habitat utilization byjuvenile Pacific salmon (Oncorhynchus) in theglacial Taku River, southeast Alaska. Can. J. Fish.

389


Aquat. Sci. 46:1677-1685Myers, K.W. 1986. The effect of altering propor-

tions of Asian chinook stocks on regional scalepattern analysis. (Submitted to the annual meet-ing International North Pacific Fisheries Com-mission, Anchorage, Alaska, November 1986.)Univ. Wash. Fish. Res. Inst. FRi-uw-8605:44 p.

Myers, K.W., C.K. Harris, C.M. Knudsen, R.V.Walker, N.D. Davis, and D.E. Rogers. 1987. Stockorigins of chinook salmon in the area of theJapanese mothership salmon fishery. N. Am. J.Fish. Manage. 7:459-474

Myers, K.W., and H.F. Horton. 1982. Temporal useof an Oregon estuary by hatchery and wild juve-nile salmon, p. 377-392. In: V.S. Kennedy (ed.).Estuarine comparisons. Academic Press, NewYork, NY

Myers, K.W., D.E. Rogers, C.K. Harris, C.M. Knud-sen, R.V. Walker, and N.D. Davis. 1984. Originsof chinook salmon in the area of the Japanesemothership and landbased driftnet salmon fish-eries in 1975-1981. (Submitted to annual meet-ing International North Pacific FisheriesCommission, Vancouver, British Columbia, No-vember 1984.) Fisheries Research Institute, Uni-versity of Washington, Seattle, WA. 204 p.

Myers, K.W.W. 1980. An investigation of the utili-zation of four study areas in Yaquina Bay, Ore-gon, by hatchery and wild juvenile salmonids.M.Sc. thesis. Oregon State University, Corvallis,OR. 233 p.

Neave, F. 1943. Diurnal fluctuations in the up-stream migration of coho and spring salmon. J.Fish. Res. Board Can. 6:158-163

. 1958. The origin and speciation of Oncorhyn-chus. Proc. Trans. R. Soc. Can. Ser. 3, 52(5):25-39

Neilson, J.D., and C.E. Banford. 1983. Chinooksalmon (Oncorhynchus tshawytscha) spawnercharacteristics in relation to redd physical fea-tures. Can. J. Zool. 61:1524-1531

Neilson, J.D., and G.H. Geen. 1981. Enumeration ofspawning salmon from spawner residence timeand aerial counts. Trans. Am. Fish. Soc.110:554-556

. 1986. First-year growth rate of Sixes Riverchinook salmon as inferred from otoliths: effectson mortality and age at maturity. Trans. Amer.Fish. Soc. 115:28-33

Neilson, J.D., G.H. Geen, and D. Bottom. 1985.Estuarine growth of juvenile chinook salmon

(Oncorhynchus tshawytscha) as inferred from oto-lith microstructure. Can. J. Fish. Aquat. Sci.42:899-908

Nicholas, J.W., and D.G. Hankin. 1988. Chinooksalmon populations in Oregon coastal river ba-sins: description of life histories and assessmentof recent trends in run strength. Oreg. Dep.Fish. Wildl. Info. Rep. 88-1:1-359

Northcote, T.G. 1976. Biology of the lower Fraserand ecological effects of pollution, p. 85-119. In:A.H.J. Dorcey, I.K. Fox, K.J. Hall, T.G. Northcote,K.G. Peterson, W.H. Sproule-Jones, and J.H.Weins (eds.). The uncertain future of the lowerFraser. Westwater Research Centre, Universityof British Columbia, Vancouver, BC

Northcote, T.G., N.T. Johnston, and K. Tsumura.1979. Feeding relationships and food web struc-ture of lower Fraser River fishes. Westwater Res.Cent. Univ. Br. Col. Tech. Rep. 16:73 p.

Novotny, A.J., and C.V.W. Mahnken. 1971. Preda-tion on juvenile Pacific salmon by a marineisopod, Rocinella belliceps pugettensis (Crustacea,Isopoda). Fish. Bull, (u.s.) 69:699-701

Paine, J.R., F.K. Sandercock, and B.A. Minaker.1975. Big Qualicum River project 1972-1973.Fish. Mar. Serv. (Can.) Tech. Rep. pAC/T-75-15:126 p.

Park, D.L. 1969. Seasonal changes in downstreammigration of age-group 0 chinook salmon in theupper Columbia River. Trans. Am. Fish. Soc.98:315-317

Parker, R.R. 1962. Estimations of ocean mortalityrates for Pacific salmon (Oncorhynchus). ]. Fish.Res. Board Can. 19:561-589

. 1965. Estimation of sea mortality rates forthe 1961 brood-year pink salmon of the BellaCoola area, British Columbia. J. Fish. Res. BoardCan. 22:1523-1554

Parker, R.R., and W. Kirkness. 1956. King salmonand the ocean troll fishery of southeasternAlaska. Alaska Dep. Fish. Res. Rep. 1:64 p.

Patten, B.C. 1971. Predation by sculpins on fallchinook salmon, Oncorhynchus tshawytscha, fry ofhatchery origin, u.s. Nat. Mar. Fish. Serv. Spec.Sci. Rep. Fish. 621:14 p.

Pauley, G.B. 1967. Prespawning adult salmon mor-tality associated with a fungus of the genusDermocystidium. J. Fish. Res. Board Can. 24:843-848

Prakash, A. 1962. Seasonal changes in feeding of

390


coho and chinook (spring) salmon in southernBritish Columbia waters. J. Fish. Res. Board Can.19:851-866

Pritchard, A.L., and A.L. Tester. 1944. Food ofspring and coho salmon in British Columbia.Bull. Fish. Res. Board Can. 65:23 p.

Quinn, T.P., and K. Fresh. 1984. Homing and stray-ing in chinook salmon (Oncorhynchus tshawyt-scha) from Cowlitz River hatchery, Washington.Can. J. Fish. Aquat. Sci. 41:1078-1082

Raymond, H.L. 1968. Migration rates of yearlingchinook salmon in relation to flows and im-poundments in the Columbia and Snake rivers.Trans. Am. Fish. Soc. 97:356-359

Real, L.A. 1980. Fitness, uncertainty, and the role ofdiversification in evolution and behavior. Amer.Nat. 155:623-638

Reid, G. 1961. Stomach content analysis of troll-caught king and coho salmon, southeasternAlaska, 1957-1958. u.s. Fish Wildl. Serv. Spec.Sci. Rep. Fish. 379:8 p.

Reimers, P.E. 1968. Social behaviour among juve-nile fall chinook salmon. J. Fish. Res. Board Can.25:2005-2008

. 1971. The length of residence of juvenile fallchinook salmon in Sixes River, Oregon. Ph.D.thesis. Oregon State University, Corvallis,OR. 99 p.

Reimers, P.E., and R.E. Loeffel. 1967. The length ofresidence of juvenile fall chinook salmon in se-lected Columbia River tributaries. Res. BriefsFish Comm. Oreg. 13:5-19

Reimers, P.E., J.W. Nicholas, D.L. Bottom, T.W.Downey, K.M. Maciolek, J.D. Rodgers, and B.A.Miller. 1979. Coastal salmon ecology project.Annual progress report, October 1,1978 to Sep-tember 30,1979; completion report, July 1,1976to September 30, 1979. Oregon Department ofFish and Wildlife, Portland, OR. 44 p.

Reimers, P.E., J.W. Nicholas, T.W. Downey, R.E.Halliburton, and J.D. Rodgers. 1978. Fall chinookecology project. Federal aid progress report fish-eries AFC-76-2. Oregon Department of Fish andWildlife, Portland, OR. 52 p.

Rich, W.H. 1920. Early history and seaward migra-tion of chinook salmon in the Columbia andSacramento Rivers. Bull. Bur. Fish, (u.s.) 37:74 p.

~ . 1925. Growth and degree of maturity ofchinook salmon in the ocean. Bull. Bur. Fish.(u.s.) 41:15-90

. 1942. The salmon runs of the ColumbiaRiver in 1938. Fish. Bull. Fish Wildl. Serv.50:101-147

Rich, W.H., and H.B. Holmes. 1928. Experiments inmarking young chinook salmon on the ColumbiaRiver 1916 to 1927. Bull. Bur. Fish, (u.s.)44:215-264

Ricker, W.E. 1976. Review of the rate of growth andmortality of Pacific salmon in salt water andnoncatch mortality caused by fishing. J. Fish.Res. Board Can. 33:1483-1524

Ricker, W.E., and K.H. Loftus. 1968. Pacific salmonmove east. Fish. Counc. Can. Annu. Rev.43:37-43.

Robinson, C.K., L.A. Lapi, and E.W. Carter. 1982.Stomach contents of spiny dogfish (Squalus acan-thias) caught near the Qualicum and Fraser Riv-ers, April-May, 1980-1981. Can. MS Rep. Fish.Aquat. Sci. 1656:21 p.

Rombough, P.J. 1985. Initial egg weight, time tomaximum alevin weight, and optimal pondingtimes for chinook salmon, Oncorhynchus tshawyt-scha. Can. J. Fish. Aquat. Sci. 42:287-291

Rounsefell, G.A. 1957. Fecundity of North Ameri-can Salmonidae. Fish. Bull. Fish Wildl. Serv.57:451-468

Russell, L.R., K.R. Conlin, O.K. Johansen, and U.Orr. 1983. Chinook salmon studies in the Ne-chako River: 1980,1981,1982. Can. MS Rep. Fish.Aquat. Sci. 1728:185 p.

Rutter, C. 1902. Natural history of the quinnatsalmon: a report of investigations in the Sacra-mento River 1896-1901. Bull. u.s. Fish. Comm.22:65-142

Sasaki, S. 1966. Distribution and food habits ofking salmon, Oncorhynchus tshawytscha, andsteelhead rainbow trout, Salmo gairdnerii, in theSacramento-San Joaquin delta, p. 108-114. In:J.L. Turner and D.W. Kelley (comp.). Ecologicalstudies of the Sacramento-San Joaquin Delta.Calif. Dep. Fish Game Fish. Bull. 136

Schluchter, M.D., and J.A. Lichatowich. 1977. Juve-nile life histories of Rogue River spring chinooksalmon, Oncorhynchus tshawytscha (Walbaum), asdetermined by scale analysis. Oreg. Dep. Fish.Wildl. Info. Rep. Ser. Fish. 77-5:24 p.

Shaw, P.A., and J.A. Maga. 1943. The effect of min-ing silt on yield of fry from salmon spawningbeds. Calif. Fish Game 29:29-41

Shelton, J.M. 1955. The hatching of chinook salmon

391


eggs under simulated stream conditions. Prog.Fish-Cult. 17:20-35

Shelton, J.M., and R.D. Pollock. 1966. Siltation andegg survival in incubation channels. Trans. Am.Fish. Soc. 95:183-187

Shepherd, B. 1975. Upper Skeena chinook stocks:evaluation of the Bear-Sustut, Morice, and lowerBabine stocks. Department of Environment,Fisheries and Marine Service, Pacific Region,North Operations Branch, Vancouver, BC. 41 p.

Sibert, J., and B. Kask. 1978. Do fish have diets? p.48-57. In: B.C. Shepherd and R.M. Ginetz(rapps.). Proceedings of the 1977 Northeast Pa-cific Chinook and Coho Salmon Workshop. Fish.Mar. Serv. (Can.) Tech. Rep. 759

Silliman, R.P. 1941. Fluctuations in the diet of thechinook and silver salmons (Oncorhynchus tscha-wytscsha [sic] and O. kisutch) off Washington, asrelated to troll catch of salmon. Copeia 1941(2):80-87

. 1950. Fluctuations in abundance of Colum-bia River chinook salmon (Oncorhynchus tscha-wytscha) [sic] 1935-45. Fish. Bull. Fish Wildl.Serv. 51:364-383

Silver, S.J., C.E. Warren, and P. Doudoroff. 1963.Dissolved oxygen requirements of developingsteelhead trout and chinook salmon embryos atdifferent water velocities. Trans. Am. Fish. Soc.92:327-343

Slater, D.W. 1963. Winter-run chinook salmon inthe Sacramento River, California, with notes onwater temperature requirements at spawning.u.s. Fish Wildl. Serv. Spec. Sci. Rep. Fish. 461:9 p.

Smith, A.K. 1973. Development and application ofspawning velocity and depth criteria for Oregonsalmonids. Trans. Am. Fish. Soc. 102:312-316

Smith-, G.R. 1975. Fishes of the Pliocene GlennsFerry formation, southwest Idaho. Mus. Paleon-tol. Pap. Paleontol. 14:1-68

Snyder, J.O. 1931. Salmon of the Klamath River,California. Calif. Fish Game Fish. Bull. 34:130 p.

Stearns, S.C. 1976. Life history tactics: a review ofthe ideas. Q. Rev. Biol. 51:3-47

Stein, R.A., P.E. Reimers, and J.D. Hall. 1972. Socialinteraction between juvenile coho (Oncorhynchuskisutch) and fall chinook salmon (O. tshawytscha)in Sixes River, Oregon. J. Fish. Res. Board Can.29:1737-1748

Taylor, E.B. 1988. Adaptive variation in rheotacticand agonistic behavior in newly emerged fry of

chinook salmon, Oncorhynchus tshawytscha, fromocean- and stream-type populations. Can. J.Fish. Aquat. Sci. 45:237-243

Taylor, E.B., and P.A. Larkin. 1986. Current re-sponse and agonistic behavior in newly emergedfry of chinook salmon, Oncorhynchus tshawytscha,from ocean- and stream-type populations. Can.J. Fish. Aquat. Sci. 43:565-573

Taylor, F.H.C. 1969. The British Columbia offshoreherring survey, 1968-69. Fish. Res. Board Can.Tech. Rep. 140:54 p.

Thomas, A.E., J.L. Banks, and D.C. Greenland. 1969.Effect of yolk sac absorption on the swimmingability of fall chinook salmon. Trans. Am. Fish.Soc. 98:406-410

Thomas, W.K., R.E. Withler, and A.T. Beckenbach.1986. Mitochondrial DNA analysis of Pacific sal-monid evolution. Can. J. Zool. 64:1058-1064

Thompson, R.B. 1959a. Food of the squawfish, Pty-chocheilus oregonensis (Richardson), of the lowerColumbia River. Fish. Bull. Fish Wildl. Serv.60:43-58

. 1959b. A study of localized predation onmarked chinook salmon fingerlings released atMcNary dam. Wash. Dep. Fish. Fish. Res. Pap.2:82-83

Tsuyuki, H., and E. Roberts. 1966. Inter-speciesrelationships within the genus Oncorhynchusbased on biochemical systematics. J. Fish. Res.Board Can. 23:101-107

Tsuyuki, H., E. Roberts, and W.E. Vanstone. 1965.Comparative zone electropherograms of musclemyogens and blood hemoglobins of marine andfreshwater vertebrates and their application tobiochemical systematics. J. Fish. Res. Board Can.22:203-213

Uremovich, B. 1977. Straying of fall chinook salmonfrom Elk River hatchery into Sixes River,1970-1976. Oreg. Dep. Fish Wildl. Info. Rep. Ser.Fish. 77-6:7 p.

Utter, F., G. Milner, G. Stahl, and D. Teel. 1989.Genetic population structure of chinook salmon(Oncorhynchus tshawytscha), in the Pacific north-west. Fish. Bull (u.s.) 87:239-264

Van Hyning, J.M. 1951. The ocean salmon trollfishery of Oregon. Pac. Mar. Fish. Comm. Bull.2:43-76

Vronskiy, B.B. 1972. Reproductive biology of theKamchatka River chinook salmon (Oncorhynchustshawytscha (Walbaum)). J. Ichthyol. 12:259-273

392


Wagner, H.H., F.P. Conte, and J.L. Fessler. 1969.Development of osmotic and ionic regulation intwo races of chinook salmon, Oncorhynchus tsha-wytscha. Comp. Biochem. Physiol. 29:325-341

Wahle, R.J., E. Chaney, and R.E. Pearson. 1981.Areal distribution of marked Columbia Riverbasin spring chinook salmon recovered in fisher-ies and at parent hatcheries. Mar. Fish. Rev.43(12):l-9

Wahle, R.J., and R.R. Vreeland. 1977. Bioeconomiccontribution of Columbia River hatchery fall chi-nook salmon, 1961 through 1964 broods, to thePacific salmon fisheries. Fish. Bull, (u.s.)76:179-208

Waite, D.C. 1979. Chinook enhancement on theKenai peninsula. Preliminary report. Study No.AFS-46. Alaska Department of Fish and Game,Juneau, AK. 51 p.

Wales, J.H., and M. Coots. 1954. Efficiency of chi-nook salmon spawning in Fall Creek, California.Trans. Am. Fish. Soc. 84:137-149

Walters, C.J., R. Hilborn, R.M. Peterman, and MJ.Staley. 1978. Model for examining early oceanlimitation of Pacific salmon production. J. Fish.Res. Board Can. 35:1303-1315

Waugh, G.D. 1980. Salmon in New Zealand, p.277-303. In: J.E. Thorpe (ed.). Salmon ranching.Academic Press, New York, NY

Weisbart, M. 1968. Osmotic and ionic regulation inembryos, alevins, and fry of the five species ofPacific salmon. Can. J. Zool. 46:385-397

Wertheimer, A.C., and M.L. Dahlberg. 1983. Reportof the incidence of coded-wire tagged salmonidsin catches of foreign commercial and research

vessels operating in the North Pacific Ocean andBering Sea during 1982-1983. (Submitted to an-nual meeting International North Pacific Fisher-ies Commission, Anchorage, Alaska, November,1983.) Auke Bay Laboratory, Northwest andAlaska Fisheries Center, National Marine Fisher-ies Service, Auke Bay, AK. 14 p.

-. 1984. Report of the incidence of coded-wiretagged salmonids in catches of foreign commer-cial and research vessels operating in the NorthPacific Ocean and Bering Sea during 1983-1984.(Submitted to annual meeting InternationalNorth Pacific Fisheries Commission, Vancouver,British Columbia, November, 1984.) Auke BayLaboratory, Northwest and Alaska FisheriesCenter, National Marine Fisheries Service, AukeBay, AK. 14 p.

Wickett, W.P. 1954. The oxygen supply to salmoneggs in spawning beds. J. Fish. Res. Board Can.11:933-953

Winans, G.A. 1989. Genetic variability in chinooksalmon stocks from the Columbia River basin. N.Am. J. Fish Manage. 9:47-52

Wright, S., and J. Bernhard. 1972. Maturity rates ofocean-caught chinook salmon. Pac. Mar. Fish.Comm. Bull. 8:49-59

Wright, S., R. Kolb, and R. Brix. 1972. Size and agecharacteristics of chinook salmon taken byWashington's commercial troll and ocean sportfisheries 1963-1969. Pac. Mar. Fish. Comm. Bull.8:37-47

Yancey, R.M., and F.V. Thorsteinson. 1963. Theking salmon of Cook Inlet Alaska, u.s. Fish Wildl.Serv. Spec. Sci. Rep. Fish. 440:18 p.

393

Pacific Salmon Life Histories - jeffersonco-treis.infojeffersonco-treis.info/s/3.04 Water... ·...

Documents

Transcript of Pacific Salmon Life Histories - jeffersonco-treis.infojeffersonco-treis.info/s/3.04 Water... ·...