Full text of "NOAA technical report NMFS SSRF"

See other formats

NOAA TR NMFS SSRF-678

NOAA Technical Report NMFS SSRF- 678

ina Biological Ls;

U.S. DEPARTMENT OF COMMFRr.F

National uceanic and Atmospheric Administration

National Marine Fisheries Service

OCi 9

Distribution, Abundance, and
Growth of Juvenile Sockeye Salmon,
Oncorhynchus nerka, and Associated
Species in the Naknek River System,
1961-64

ROBERT J. ELLIS

SEATTLE, WA
September 1974

NOAA TECHNICAL REPORTS
National Marine Fisheries Service, Special Scientific Report— Fisheries Series

The major responsibilities of the National Marine Fisheries Service (NMFS) are to monitor and assess t ht- abundance and geographic distribution of fisher) resources to

understand and predict fluctuations in the quantity and distribution of these resources, and to establish levels for optimum use of the resources NMFS i- also charged with
the development and implementation of policies for managing national fishing grounds, development and enforcement of domestic fisheries regulations, surveillance ot foreign
fishing off United States coastal waters, and the development and enforcement of international fishery agreements and policies. NMFS also assists the fishing industry through
marketing service and economic analysis programs, and mortgage insurance and vessel construction subsidies It collects, analyzes, and publishes statistics on various phases ol
the industry.

The Special Scientific Report — Fisheries series was established in 1949. The series carries reports on scientific investigations that document long-term continuing programs
of NMFS. or intensive scientih. reports on studies of restricted scope The reports mav deal with applied fishery problems. The series is also used as a medium tor the publica-
tion of bibliographies of a specialized scientific nature.

NOAA Technical Reports NMFS SSRF are available free in limited numbers to governmental agencies, both Federal and State They are also available in exchange for
other scientific and technical publications in the marine sciences Individual copies may be obtained (unless otherwise noted) from D83. Technical Information Division,
Environmental Science Information Center. NOAA. Washington; DC. 20235. Recent SSRF's are:

619. Macrozooptankton and small nekton in the coastal waters off Vancouver Island
(Canada i and Washington, spring and H\ Donald S Daw -January 1971. iii +

94 pp.. 19 figs.. 13 tables.

635. A bibliography of the blackfin tuna. Thunnus atlanticus (Lesson). B\ Grant L.
Beardsley and David C. Simmons. August 1971. 10 pp. For sale by the Superintendent of
Documents, U S Government Printing Office. Washington, D.C. 20402.

620. The Trade Wind Zone Oceanography Pilot Study Part IX: The sea-level wind field
and wind stress values. July 1963 to .June 1965. By Gunter R. Seckel. June 1970, iii + 66
pp . 5 tigs

636. Oil pollution on Wake Island from the tanker R C Stoner. By Reginald M.
Gooding May 1971, iii + 12 pp.. 8 figs., 2 tables. For sale by the Superintendent of
Documents. U.S. Government Printing Office. Washington, D.C. 20402.

621. Predatum b> sculpins on tall chinook salmon, Oncorkynchus tshawytscha, fry of
hatchery origin By Benjamin G Patten. February' 1971, iii + 14 pp.. 6 figs.. 9 tables

622. Number and lengths, by season, of fishes caught with an otter trawl near Woods
Hole. Massachusetts. September 1961 to December 1962. By F. E. Lux and F E Ni< h.
Februarv 1971, iii + 15 pp.. 3 figs.. 19 tables.

623- Apparent abundance, distribution, and migrations of albacore. Thunnus alaiunga,
on the North Pacific longline grounds. By Brian J. Rothschild and Marian Y. Y Yong
September 1970, v + 37 pp., 19 figs., 5 tables.

624. Influence of mechanical processing on the quality and yield of bay scallop meats. By
N B Webb and F. B. Thomas. April 1971, iii + 11 pp., 9 figs.. 3 tables.

625. Distribution of salmon and related oceanographic features in the North Pacific
Ocean, spring 1968. By Robert R. French, Richard G. Bakkala, Masanao Osako, and Jun
Ito. March 1971. iii + 22 pp.. 19 figs., 3 tables.

626. Commercial fishery and biology of the freshwater shrimp. Macrobrachium, in the
Lower St Paul River, Liberia. 1952-53. By George C. Miller. February 1971. iii + 13 pp., 8
tigs , 7 tables.

627. Calico scallops of the Southeastern United States. 1959-69, By Robert Cummins, Jr.
June 1971. iii + 22 pp., 23 figs.. 3 tables.

628. Fur Seal Investigations, 1969 By NMFS. Marine Mammal Biological Laboratory
August 1971. 82 pp., 20 figs., 44 tables, 23 appendix A tables, 10 appendix B tables.

629. Analysis of the operations of seven Hawaiian skipjack tuna fishing vessels, June-
August 1967. By Richard N. Uchida and Ray F. Sumida. March 1971, v + 25pp., 14 figs..
21 tables For sale bv the Superintendent of Documents, U.S. Government Printing Of-
fice, Washington, D.C. 20402.

630. Blue crab meat I Preservation by freezing July 1971. iii + 13 pp.. 5 figs.. 2 tables
II. Effect of chemical treatments on acceptability. By Jurgen H. Strasser, Jean S. Lennon,
and Frederick J King July 1971. iii + 12 pp.. 1 fig , 9 tables.

631. Occurrence of thiaminase in some common aquatic animals of the United States
and Canada. By R A Greig and R H Gnaedinger July 1971, iii + 7 pp., 2 tables.

632. An annotated bibliography of attempts to rear the larvae of marine fishes in the
laboratory By Robert C. May August 1971, iii + 24 pp.. 1 appendix I table, 1 appendix II
table. For sale bv the Superintendent of Documents. U.S. Government Printing Office.
Washington, D.C. 20402.

633. Blueing of processed crab meat. II. Identification of some factors involved in the
blue discoloration of canned crab meat Callinectes sapidus By Melvin E. W:aters. May
1971. in + 7 pp., 1 fig.. 3 tables.

634. Age composition, weight, length, and sex of herring, Clupea pallasii, used for reduc-
tion in Alaska. 1929-66. By Gerald M. Reid. July 1971. iii + 25 pp., 4 figs., 18 tables.

637. Occurrence of larval, juvenile, and mature crabs in the vicinity of Beaufort Inlet.
North Carolina B\ Donnie L. Dudley and Mayo H. Judy. August 1971, iii +■ 10 pp.. 1 fig .

iles For sale by the Superintendent of Documents, U.S. Government Printing Office,
Washington. I) C 20402.

638. Length-weight relations of haddock from commercial landings in New England.
55 B> Bradford E. Brown and Richard C. Hennemuth. August 1971, v + 13 pp., 16

figs.. 6 tables. 10 appendix A tables. For sale by the Superintendent of Documents. U.S.
Government Printing Office. Washington. D.C. 20402.

639. A hydrographic survey of the Galveston Bay system, Texas 1963-66. By E.J. Pullen,
W. L.Trent, and G B. Adams. October 1971, v + 13 pp.. 15figs., 12 tables. For sale by the
Superintendent of Documents, US Government Printing Office, Washington, D.C.
20402

640. Annotated bibliography on the fishing industry and biology of the blue crab.
Callinectes sapidus. Bv Marlin E. Tagatz and Ann Bowman Hall. August 1971. 94 pp. For
sale bv the Superintendent of Documents. U.S. Government Printing Office, Washington,
D.I 20402

641 Use ol threadfin shad. Dnrosoma petenense, as live bait during experimental pole-
and-line fishing for skipjack tuna. Katsuwonus pelamts, in Hawaii Bv Robert T B.
Iversen. August 1971. iii + 10 pp., 3 figs.. 7 tables. For --ale by the Superintendent of
Documents. U S Government Printing Office. Washington. D.C. 20402.

642. Atlantic menhaden Breuoortia tyrannus resource and fishery — analysis ol decline

By Kenneth A. Henry August 1971, v + 32 pp., 40 figs , 5 appendix figs . 3 tables. 2
appendix tables Fur sale by the Superintendent of Documents. U.S. Government Printing
Office Washington, D C 20402.

643 Surface winds ol the southeastern tropical Atlantic Ocean By John M. Steignerand
Merton C Ingham October 1971. iii + 20 pp., 17 figs For sale by the Superintendent of
Documents. US Government Printing Office. Washington. D.C. 20402.

644 Inhibition of flesh browning and skin color fading in frozen fillets of yelloweye
snapper [Lutzanus viuanus). Bv Harold C. Thompson. Jr., and Mary H. Thompson
February 1972. iii + 6 pp.. 3 tables For sale by the Superintendent of Documents. U S
Government Printing Office. Washington, D.C. 20402.

645 Traveling screen for removal of debris from rivers By Daniel W Bates, Ernest \\
Murphey, and Martin G Beam October 1971. iii + 6 pp., 6 figs., 1 table. For sale h\ the
Superintendent of Documents. U.S. Government Printing Office. Washington, D.C

21*402

646. Dissolved nitrogen concentrations in the Columbia and Snake Rivers in 1970 and
their effect on chinook salmon andsteelhead trout By Wesley J. Ehel August 1971, iii + 7
pp., 2 figs . 6 tables For sale by the Superintendent of Documents. I S Government
Printing Office, Washington, D.C. 20402.

647 Revised annotated list of parasites from sea mammals caught off the west coasl ol
North America. By L. Margolisand M. D. Dailey. March 1972. iii + 23 pp For sale by the
Superintendent of Documents. U.S. Government Printing Office, Washington, D.C.
20402.

Continued on inside back cover

„Q ATMOSp^

''Went of

U.S. DEPARTMENT OF COMMERCE
Frederick B. Dent, Secretary

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION
Robert M. White, Administrator

NATIONAL MARINE FISHERIES SERVICE
Robert W. Scheming, Director

NOAA Technical Report NMFS SSRF-678

Distribution, Abundance, and Growth of
Juvenile Sockeye Salmon, Oncorhynchus
nerka, and Associated Species in the
Naknek River System, 1961-64

ROBERT J. ELLIS

.0^T'0/V

SEATTLE, WA
September 1974

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington. DC. 20402

The National Marine Fisheries Service (NMFS) does not approve, rec-
ommend or endorse any proprietary product or proprietary material
mentioned in this publication. No reference shall be made to NMFS, or
to this publication furnished by NMFS, in any advertising or sales pro-
motion which would indicate or imply that NMFS approves, recommends
or endorses any proprietary product or proprietary material mentioned
herein, or which has as its purpose an intent to cause directly or indirectly
the advertised product to be used or purchased because of this NMFS
publication.

CONTENTS

Page

Introduction 1

The study area 2

Methods and equipment 4

Sampling units 4

Types of gear 4

Measurements of fish 5

General distribution and abundance of all fish species 6

Abundance of juvenile sockeye salmon 6

Trends in abundance for the entire system 11

Comparative abundance among lakes 14

Abundance in each lake of the system 16

Coville Lake 16

Grosvenor Lake 19

Iliuk Arm 20

South Bay 20

West End 20

North Arm 21

Northwest Basin 21

Brooks Lake 21

Abundance of associated species 21

Pond smelt 21

Threespine sticklebacks 24

Ninespine sticklebacks 24

Interlake migration of presmolt sockeye salmon 25

Migration from Coville Lake to Grosvenor Lake 26

1961 26

1962 26

1963 27

1964 27

Migration from Grosvenor Lake to Iliuk Arm 27

Significance of the summer outmigrations from Coville Lake 28

Diel timing of migrations 29

Behavior of schools of age 0 fish at outlet of Coville Lake 31

Early rearing areas of sockeye salmon fry from Grosvenor River and Hardscrabble Creek 31

Size, length frequency, and growth 32

Juvenile sockeye salmon 33

Coville Lake and Coville River 35

Grosvenor Lake and Grosvenor River 36

Iliuk Arm 36

South Bay 36

West End 39

North Arm 41

Northwest Basin 41

Brooks Lake 41

Causes of differences in size of juvenile sockeye salmon on 1 September 41

Real differences in rates of growth 41

Differences in time of recruitment of fry 43

Differences in rates of dispersion of large and small or fast- and slow-growing fish 43

Differences in size of fry at time of emergence 44

Species commonly associated with juvenile sockeye salmon 44

Threespine sticklebacks 44

Ninespine sticklebacks 44

Pond smelt 44

Pygmy whitefish and least cisco 46

Predation on juvenile sockeye salmon 46

Lake trout 46

H umpback whitefish 48

iii

Arctic char and Dolly Varden 49

Other species 50

General significance of predation 50

Summary and significance for resource development 50

Acknowledgments 52

Literature cited '*.

Figures

1 . Naknek River system, Bristol Bay, Alaska 2

2. Coville Lake, Naknek River system 5

3. Weighted daily mean number of age 0 and age I sockeye salmon per standard tow in Naknek River
system, 1 1 July to 29 August 1962 13

4. Weighted daily mean number of age 0 sockeye salmon per standard tow in Naknek River system
1961-64 H

5. Weighted mean number of age 0 sockeye salmon per standard tow, Naknek River system 1961-63. . 15

6. Weighted mean number of age I sockeye salmon per standard tow, Naknek River system 1961-63. . 15

7. Mean number of age 0 sockeye salmon per standard tow, Naknek River system 1961-63 17

8. Mean number of age 0 sockeye salmon per standard tow, Coville Lake 1961-64 18

9. Mean number of pond smelt per standard tow in Coville Lake 1961-63 24

10. Mean number of threespine sticklebacks per standard tow. West End 1961-63 25

1 1. Curves of apparent growth of age 0 sockeye salmon captured in Coville Lake and Coville River
1961-64 ■ ■ ■ 34

12. Length frequency distributions of juvenile sockeye salmon captured in Coville Lake and Coville
River, July and September 1961-63 34

13. Length frequency distributions of juvenile sockeye salmon captured in Coville Lake and Coville
River, July and September 1964 35

14. Curves of apparent growth of age 0 sockeye salmon captured in Grosvenor Lake and Grosvenor River
1961-63 36

15. Length frequency distributions of juvenile sockeye salmon captured in Grosvenor Lake and Gros-
venor River, July and September 1961-63 37

16. Curves of apparent growth of age 0 sockeye salmon captured in Uiuk Arm 1961-63 37

17. Curves of apparent growth of age I sockeye salmon captured in Iliuk Arm 1961-63 37

18. Length frequency distributions of juvenile sockeye salmon captured in Iliuk Arm, July and August
1961-63 38

19. Curves of apparent growth of age 0 sockeye salmon captured in South Bay 1961-63 39

20. Curves of apparent growth of age I sockeye salmon captured in South Bay 1961-63 39

21. Length frequency distributions of juvenile sockeye salmon captured in South Bay, July and August
1961-63 40

22. Curves of apparent growth of age 0 sockeye salmon captured in West End 1962-63 41

23. Length frequency distributions of juvenile sockeye salmon captured in West End, July and August
1961-63 42

24. Curves of apparent growth of age 0 sockeye salmon captured in Brooks Lake 1961-63 43

25. Length frequency distributions of threespine sticklebacks captured in the Naknek River system
1961-64 45

26. Length frequency distributions of ninespine sticklebacks captured in the Naknek River system, 1961,
1963, and 1964 46

27. Length frequency distributions of pond smelt captured in the Naknek River system, 1961, 1963, and
1964 47

Tables

Area of lakes of the Naknek River system 3

i v

2. Data on spawning for lakes of the Naknek River system 1959-63 4

3. Percent frequency of occurrence and percent total number offish captured in lakes of the Naknek
River system 1962 7

4. Numbers of age 0 sockeye salmon taken in Coville Lake and Iliuk Arm 1964 10

5. Subplot portion of split-plot analysis of variance of catch of juvenile sockeye salmon in three lake
basins of the Naknek system 1962-64 11

6. Two-way analysis of variance of abundance of juvenile sockeye salmon in selected lakes of the
Naknek system 1961-64 12

7. Mean number of age 0 and age I sockeye salmon taken in Naknek River system, August 1961-64. . 14

8. Relative abundance of spawning grounds and average catch per unit of effort of age 0 sockeye salmon

in July 1961-63 in lakes of the Naknek River system 16

9. Mean fork length and standard deviation of age 0 sockeye salmon taken in Coville Lake and Coville
River, 1 1 July to 1 Sept. 1963 19

10. Mean fork length and standard deviation of age 0 sockeye salmon taken in Coville Lake and Coville
River, 4 July to 5 Sept. 1964 19

11. Split-plot analysis of variance of abundance of pond smelt, threespine stickleback, and ninespine
stickleback 1962-64 22

12. Estimated number of age 0 sockeye salmon migrating from Coville Lake to Grosvenor Lake, 22 July to

10 Sept. 1961 26

13. Estimated numbers of age 0 and age I sockeye salmon migrating from Coville Lake to Grosvenor
Lake, 29 May to 15 Sept. 1962 27

14. Estimated numbers of age 0 and age I sockeye salmon migrating from Coville Lake to Grosvenor
Lake, 20 June to 17 Sept. 1963 27

15. Estimated numbers of age 0 and ~~= I sockeye salmon migrating from Coville Lake to Grosvenor
Lake, 1 1 July to 7 Sept. 1964 28

16. Estimated numbers of age 0 sockeye salmon migrating from Grosvenor Lake to Iliuk Arm, 15 July to

17 Sept. 1962 28

17. Number of age 0 sockeye salmon in Coville Lake at the end of summer and number that migrated from
lake during summer 1961-64 28

18. General magnitude of age 0 sockeye salmon in interlake migrations and of lake populations in July and
August 1961-63, Coville River-Iliuk Arm area 29

19. Rate of catch and mean size of age 0 sockeye salmon migrating down Coville and Grosvenor Rivers
between July and September 1961-62 30

20. Numbers of recently emerged sockeye salmon captured on shores of Grosvenor River in May and
June 1962 32

21. Mean fork lengths of age 0 and age I sockeye salmon in each lake of Naknek River system and Coville
and Grosvenor Rivers on 20 August and 1 September 1961-64 33

22. Mean surface water temperatures, mean number of age 0 and age I sockeye salmon, pond smelt, and
threespine and ninespine sticklebacks, and mean fork lengths of age 0 sockeye salmon in lakes of the
Naknek River system 1961-63 43

23. Stomach contents of lake trout captured in 1963 48

24. Length frequencies of lake trout captured in Grosvenor Lake 1963 48

25. Length frequencies of lake trout captured in Grosvenor Lake 1964 49

26. Length frequencies of humpback whitefish captured in Coville Lake 1963 49

Distribution, Abundance, and Growth of Juvenile Sockeye Salmon,

Oncorhynchus nerka, and Associated Species in

the Naknek River System, 1961-64

ROBERT J. ELLIS'

ABSTRACT

The Naknelt River system contains eight interconnected and generally biologically discrete basins,
each with a different ratio of spawning grounds to rearing area for sockeye salmon, Oncorhynchus nerka,
and different densities of juvenile sockeye salmon and associated species offish. Juvenile sockeye salmon
and other pelagic species were sampled with tow nets at night. Sockeye salmon were the most common and
abundant species in all basins, followed by threespine sticklebacks, ninespine sticklebacks, and pond
smelt. Eighteen other species of potential competitor or predator fish were present.

In the summers of 1961 to 1963, juvenile sockeye salmon in the pelagic areas had a characteristic
pattern of abundance for the entire system: abundance (catch per tow) of age 0 increased from early
summer to midsummer and then declined to late August. The abundance in late August varied about
threefold and, in genera!, was independent of variations in the number of parents from 1960 to 1963.

In July the abundance of age 0 fish in each basin was proportional to the amount of known contiguous
spawning ground, but by late August this relation no longer existed. This change was at least partly due to
migration of the age 0 fish — generally from basins of greater abundance of fish to those of lesser abun-
dance. The larger and faster growing fish were the first to migrate. Not all basins were involved in these
migrations.

The production of sockeye salmon smolts in the Naknek system is relatively stable. At least three
major factors probably contribute to this stability: (1) the presence of several major spawning units or
races in widely separated spawning grounds of different types, (2) the presence of several connected lakes,
and (3) the migratory behavior of juvenile sockeye salmon during their first summer.

A mechanism which prevents the population of juvenile sockeye salmon from exceeding some upper
limit is not apparent in the Naknek system. A reduction in growth in areas of high density was not
apparent in the Naknek system in 1961-64 and apparently did not occur in 1957-65. Many kinds of
predators on juvenile salmon are present but probably are not limiting production of smolts.

The data on abundance and growth of juvenile sockeye salmon and the distribution of the escapement
and spawning grounds indicate that it should be possible to increase the production of sockeye salmon in
the Naknek system. Two of the major basins. North Arm and Brooks Lake, which constitute about 35% of
the system, are now producing juveniles at very low levels. North Arm appears to suffer from too little
spawning area, whereas Brooks Lake appears to have adequate spawning area but too few spawners.

Three factors in the biology of juvenile sockeye salmon of the Naknek system are of special signifi-
cance to the managers of the resource and should be investigated in any effort to enhance the production
of sockeye salmon in the Naknek system: ( 1 1 the abundance of smolts each spring is fairly constant for the
system as a whole and not closely related to the abundance of the parents or, from 1961-64, even to the
original abundance of age 0 fish; (2) the apparent growth of juvenile sockeye salmon and potential
competitor species is not related to the abundance of these fish in any lake of the Naknek system; and (3)
two major lakes, constituting about iS% of the rearing waters, do not receive age 0 sockeye salmon from
other basins and are supporting relatively few sockeye salmon.

The question of what escapement of adult sockeye salmon is needed to ensure full production of
juveniles is considered. The present study indicates that escapements in the range of 600,000 to 1,000,000
fish, as recommended by other studies, would probably fully use the present combination of spawning and
rearing areas without danger of overburdening the food supply.

The Naknek River system — the Naknek River and
tributary lakes — is one of several major producers of
sockeye salmon, Oncorhynchus nerka, in Bristol Bay,
Alaska. The annual commercial value of the catch of
sockeye salmon from the Naknek system has varied in
recent years from a few hundred thousand to more
than a million dollars, and the ultimate goal of fishery
research here is to stabilize the production at the

'Auke Bay Fisheries Laboratory, National Marine Fisheries Ser-
vice, NOAA, Auke Bay, AK 99821.

higher or even increased levels. As biologists learn
more of the life history of sockeye salmon, it becomes
increasingly evident that although most stocks (races)
have the same general life history, each stock has
unique characteristics that are determined by the
biological and physical environments in which each
stock evolved. It is the interaction between these
characteristics and the environment that makes some
stocks more productive than other stocks in the same
year and some years more productive than other years
for the same stock.

The sockeye salmon of the Naknek system have the
general freshwater life history common to most stocks
of the species. Adults return to fresh water in early
summer, ascend the system through rivers and lakes,
and spawn in gravel of streams or lake beaches. The
embryos overwinter in the gravel, and young salmon
emerge and enter the littoral areas of the lakes in
spring. The juvenile salmon soon move out into the
pelagic areas where they feed on zooplankton for 1 or
2 summers before going to the ocean as smolts in the
spring. In the Naknek system, smolts are yearlings
(age I); 2-year-olds (age II); or, rarely, 3-year-olds
(age III). Each lake in the Naknek system has its own
unique combination of physical features and assem-
blage of other species offish associated with the young
sockeye salmon.

The National Marine Fisheries Service (formerly
the Bureau of Commercial Fisheries) has conducted
research on the Naknek system since about 1940, but
intensive work on juvenile sockeye salmon and as-
sociated species offish began in 1961. A principal ob-
jective of this research has been to define some of the
details of the life history of the juvenile sockeye salm-
on in the system. The results of the research on
juvenile sockeye salmon through 1962 were presented
in a report that summarized all available information
on the major sockeye salmon systems of southwestern
Alaska (Burgner et al., 1969).

I continued the work on juvenile sockeye salmon
and associated species in the Naknek system, and in
this report I analyze the data collected from 1961
through 1964. First is a description of the general dis-
tribution and relative abundance of all species of fish

in the system, based on sampling with several types of
gear. This is followed by a discussion of the abun-
dance of juvenile sockeye salmon and a few associated
species in the habitats where these fish are most
abundant — the pelagic areas. Next is the account of
the migrations of young-of-the-year (age 0) sockeye
salmon from lake to lake in two areas. Changes in
average lengths and length-frequency distributions are
then used to determine relative growth in the lakes of
the system. The significance of predators in control-
ling the numbers of juvenile sockeye salmon in the
Naknek system is considered next. Finally, all of the
available information is marshaled and summarized to
consider for the fishery manager what factors seem to
be limiting the production of sockeye salmon in the
Naknek system and what might be done to increase
production.

THE STUDY AREA

The freshwater environment of sockeye salmon in-
cludes the spawning grounds of streams or lake
beaches, followed briefly by the open waters of the
spawning streams or beaches, and then the littoral
areas of the lakes for a few days or weeks and the
pelagic areas of the lakes for several months, followed,
again briefly, by the outlet river as the juveniles go to
the ocean as smolts.

The Naknek system (Fig. 1) consists of four major
connected lakes — Coville, Grosvenor, Naknek, and
Brooks — and the outlet stream, Naknek River, which
connects the lakes to the ocean. Naknek Lake con-
tains four distinct basins and a large shallow outlet

HAMMERSLY

V"lv. HARDSCRABBLE
CREEK ■

HEADWATER
CREEK

Figure 1.— Naknek River system, Bristol Bay, Alaska, showing sampling units where juvenile sockeye salmon were studied from

1961 to 1964.

portion, each of which I treat as an entity — Iliuk Arm,
South Bay, West End, North Arm, and Northwest
Basin. Two small lakes at relatively high elevations,
Hammersly and Murray, receive small numbers of
adult sockeye salmon, but were not part of this study.
The basic bathymetry and limnology of the lakes of
the Naknek system have been determined. The total
surface area and the areas within the 5-m contour,
selected as the arbitrary limit of the pelagic area, for
each lake or basin and the sampling units within each
lake or basin are itemized in Table 1 . The limnology of
these lakes was intensively studied in 1961 and 1962,
and details of the chemistry and productivity were
summarized and compared with other western Alaska
lakes (Burgneretal., 1969). In general, the lakes of the
Naknek system are deep and oligotrophic and have a
pH of about 7.2 and alkalinity of about 26 ppm. Max-
imum summer surface temperatures reach 12° to 16°C.

and although thermoclines occasionally occur, they
usually last only a few days.

Each lake of the Naknek system has several spawn-
ing grounds that are used by sockeye salmon, but
neither the extent of the spawning grounds nor the
numbers of spawners in the escapements (the adult
salmon that escape the fishery and enter fresh water to
spawn) are uniformly proportional to the size of the
lakes (Table 2). In Table 2 the various types of stream
spawning grounds have been combined for each lake
(the few known beach spawning areas are not signifi-
cant). The distribution of spawners among the several
lakes each year is variable and only occasionally pro-
portional to the amount of spawning ground contigu-
ous to each lake. For example, American Creek
(Coville Lake) has about one-third of the system's
spawning ground, but from 1959 to 1963 it received
from 10% to 60% of the escapement.

Table 1. --Total surface area, area within 5-m contour, percent of each lake deeper than 5 m, and
percent of system total deeper than 5 m for lakes of the Naknek River system.

Lake and
sampling area

Total surface
area (km2)

Area within

5-m contour

(birj

Percent of each lake

deeper than 5-m
Contribution Total
of each area lake

Percent of

system total

deeper than

5 m

Coville Lake1
C-l
C-2
Total
Grosvenor Lake
G-l and 2
G-3 and 4
Total
Iliuk Arm
.VI 5
N-14
N-13
Total
Sou tli Bay
N-ll
N-9
N-6
Total
West End
N-4
N-2
N-l
Total
North Arm (all
units combined)
Nor times t Basin

N-3
Brooks Lake
B-l and 2
System total

9.3
24.1

1.1
19.2

5.4
94.6

—

0.2
2.9

53.4

20.5

100.0

60.8

5.1

50.9

42.5

27.9
58.1

42.3

5"."

4.2

5.7

75.2

66.0

100.0

90.0

9.9

19.2

55.5
41.1

18.2

52.4
59.1

20.5
56.1
45.6

2."
4.8
5.S

95.6

89.7

100.0

95.8

15.5

15.6
16.2
42.6

11.9

15.2
59.8

17.8
22.8
59.4

1.8

2.5
5.9

74.4

66.9

100.0

89.9

10.0

56.0
81.0

81.4

50.8
74. S
5b. 1

31.4

46.5

22.5

7.6
11.2

5.4

21S.4

161.7

100.0

74.0

24 . 2

181.5

40.8

74.9

162.1

74.6

89.

70.

99.6

4.5

11.1

790.2

670.0

84. S

!In 1963 and 1964 Coville Lake was divided into more sample areas; the percent of the surface
area in water deeper than 5 m in each sampling area was: 1965--C-1 = 5.50, C-2U = 55.86, C-2M =
28.58, C-2L = 10.26; 1964--C-1 = 10.67, C-2 = 51.50, C-5 = 55.57, C-4 = 11.77, C-5 = 10.66.

Table 2. --Area of potential spawning grounds, numbers of spawners in escapements, and numbers of
smolts produced by each brood year for lakes of the Naknek River system, 19S9-63.

Lake or basin

Surface
area

Got,2)

Area of

potential

spawning

grounds (ha)

Area per unit
lake area
(ha/km2)

Spawners in escapement (thousands)
1959 1960 1961 1962 1963

Coville Lake
Grosvenor Lake
Iliuk Arm
South Bay
West End
North Arm
Northwest Basin
Brooks Lake
System total6
Total smolts
produced6 by
brood year
(millions)

JO.

73.

93.

74.
218.
181.

40.

74.

790.;

111.0

!29.6

34.5

35.5

147.8

7.5

0.7

18.0

55T7T

0.40
0.57
0.07
0.68
0.04
0.02
0.24

1,000

ISO
22
10

218

85
235

4 72
(4) 75
6 8

0.45

!, 251.8

54
200

'10

52S.4 551.1 725.1 905.4

13.0 16.7 11.1 12.1 20.8

1 Includes Hardscrabble Creek; does not include beach spawning areas.

2Hardscrabble Creek weir count.

3Includes Brooks River, which commonly has three w-aves of spawning activity.

^Salmon were observed spawning in the West tnd in 1961, but the number is not known.

5Field Reports, 1962 and 1965, Brooks Lake Field Station, Natl. Mar. Fish. Serv. Auke Bay
Fish. Lab., Auke Bay, AK 99821.

6Stewart, Donald M. (editor). 1969. 1967 Bristol Bay red salmon smolt studies, Appendix
D, Table 2, p. 64. Alaska Uep. Fish Game, Inform. Leafl. 154.

METHODS AND EQUIPMENT

Sampling Units

For sampling, the lakes were divided into units,
generally on the basis of surface area. Each unit was
designated by a system of letters and numbers (N-l,
N-2, C-l, C-2, etc.— Fig. 1). Coville Lake was further
divided in 1963 and 1964, and the designations of the
sampling units were changed (see Fig. 2). The original
objectives were to establish units of about equal size
that were small enough to reveal possible gradients in
biological attributes and few enough to permit sam-
pling with a limited effort. As the study progressed
some units were further divided and others combined.

Types of Gear

Several types of gear were used to sample fish and
many revisions were made throughout the 4 yr of the
study.

Pelagic areas were sampled with tow nets similar to
those used by Johnson (1956) and Burgner ( 1958). Two
types of tow nets were used. The first, which was used
in all 4 yr, had a round metal hoop 3 m (10 ft) in
diameter with an attached cone-shaped mesh bag
about 7.6 m (25 ft) long. It was connected to two boats
by bridles and steel cables retrieved by a gasoline-
powered winch (1961 and 1962) or by ropes retrieved
by hand (1963 [in part] and 1964). The second net.

which was used only for some collecting in 1964, had a
2.7-m-square (9 ft) opening and was towed by ropes
and retrieved by hand.

Tow netting was usually done between 2200 and
0200, or in general from sunset to sunrise. Two kinds
of tows were made: (1) surface tows (0 to 3 m) with the
center of the net 1.4 or 1.5 m from the surface; and (2)
deep tows (3 to 6 m) with the center of the net 4.1 or
4.5 m from the surface. To produce a "standard" tow,
the net was pulled through the water over a 457 m
( 1 ,500 ft) course in about 6 min 15 sec for a surface tow
and 6 min 45 sec for a deep tow. Most tows were of the
surface type in 1961, but in 1962, 1963, and 1964, a
sequence of tows — one surface, two deep, and one
surface — was used.

Field crews selected the specific track to be towed
on any night within an area; the general objective was
to tow near the middle of a sampling area. When one
considers that the crews depended on outlines of hills
and mountains and running time for orientation, the
selection of specific sampling tracks must be con-
sidered as random, with bias toward the center of the
sampling area.

Littoral areas were sampled with beach seines, gen-
erally in water less than 3 m deep. Two types of nylon
seines were used. One was 31 m (100 ft) long; the
center 6 m was 1 m high and had four meshes per inch
(2.5 cm), and the balance was 1.2 m high and had two

meshes per inch. The other seine was 40 m (130 ft)
long and 3 m high; the center 9 m of the web had four
meshes per inch and the balance had two.

Pelagic and littoral areas were also sampled with
floating box traps in 1962 and 1963. The box portion of
the trap was about 1.2 m (4 ft) square in cross section
by 1 .8 m (6 ft) long; wings extended 4.5 m from the box
and the lead was 15 m long. The box and wings had
four meshes per inch and the lead had two. To sepa-
rate fish entering from each side, the box had a
lengthwise partition connected to the lead.

A small otter trawl (gulf-type shrimp try-net) was
used sporadically throughout the system. The wings
had a spread of about 9 m and were about 1 m high.
The net was cotton and had two meshes per inch in the
wings and body and four meshes per inch in the tail.

Gill nets were also fished sporadically. The sizes
varied from a 1.3-cm ('/i-inch) bar to a 10-cm (4-inch)
bar. Small nets were nylon and large ones were cotton
or linen.

Rivers and streams were sampled with small and
large fyke nets. The small nets were 1 m (3 ft) square
with 1.2-m (4-ft) wings and were made of nylon web
with eight meshes per inch. The large nets were 1.2 m
square or 1.2 m wide by 1.5 m high and had 1.8-m
wings. The large nets were nylon web with two
meshes per inch in the wings and body and four
meshes per inch in the tail and cod end. The cod end of
the net was often replaced with a 20.2-cm (8-inch)

diameter flexible hose connected to a floating livebox.
With this arrangement several thousand juveniles
could be collected without many being killed.

Angling with sport fishing gear was used to supple-
ment other sampling methods.

Measurements of Fish

Sockeye salmon juveniles and associated species
were usually measured for fork length (tip of snout to
fork of tail) to the nearest millimeter and weighed
(drained weight) to the nearest higher gram. The fish
were usually preserved in 10% Formalin for at least
48 h, but less than 1 wk, before being measured or
weighed. Sockeye salmon smolts and recently
emerged fry were measured alive, but anesthetized;
the fry were measured for total length (tip of snout to
tip of tail in normal extension).

The preserved juvenile sockeye salmon were also
routinely weighed by 3-mm size groups on a triple
beam balance. Length and weight data were combined
to yield "condition factors." These condition factors
were somewhat variable but usually well above 1.0000
for all fish from all lakes. No utility was seen in the
condition data and the weight data will not be consid-
ered in this report.

AMERICAN CREEK

Figure 2. — Coville Lake, Naknek River system, showing units where juvenile sockeye salmon were sampled with tow nets in 1963

(lower) and 1964 (upper).

GENERAL DISTRIBUTION AND ABUNDANCE
OF ALL FISH SPECIES

Although the principal subject of this study was
juvenile sockeye salmon, data were collected on all
species of fish encountered because of probable in-
teractions among the species. Earlier work (Johnson,
1956) had indicated that juvenile sockeye salmon were
readily available to tow nets in summer in the pelagic
portion of the freshwater rearing areas and our effort
was concentrated on this gear and habitat. We sam-
pled with other gear in other habitats, however, to
learn more of the biology of all the species present.

The greatest effort with all types of gear was in 1962;
the results for that year are summarized in Table 3 to
give a general picture of the distribution and relative
abundance of all species. The table shows the percent
frequency of occurrence of each species in collections
made with each type of gear and its contribution to the
total catch as percent of the total number of fish cap-
tured by each gear in each lake. The data are known to
be biased in at least three ways: (1) most of the sam-
pling was done from 15 July to 1 September, and
marked seasonal changes in distribution are known to
occur for many species; (2) each type of gear has its
peculiar abilities to catch the various species; and (3)
the distribution of fishing effort varied between areas
in regard to type of gear, amount of effort, and season.
Because of these biases, detailed discussion of the dis-
tribution of all species is not warranted and the abun-
dance in relation to juvenile sockeye salmon will be
treated in detail only for those species consistently and
abundantly captured in the pelagic areas in tow nets —
threesprine and ninespine sticklebacks and pond
smelt.

Five species offish were clearly predominant in the
collections — sockeye salmon, threespine and nine-
spine sticklebacks, pygmy whitefish, and pond smelt
(Table 3). The most widely distributed and, in general,
the most abundant species was the sockeye salmon.
Juvenile sockeye salmon were taken with all appropri-
ate gear and in all major lakes of the system. The
distribution of threespine and ninespine sticklebacks
approximated that of the sockeye salmon and in a few
areas the sticklebacks were more abundant than
juvenile salmon (e.g.. West End and Northwest
Basin). The other two species that occurred abun-
dantly in some collections, pygmy whitefish and pond
smelt, were each abundant in some basins, but were
never abundant together. The pond smelt was abun-
dant only in tow net catches in Coville Lake and the
pygmy whitefish only in trawl and seine catches in
Brooks Lake and parts of South Bay.

None of the many other species were ever abundant
in the collections. Some, such as the coho salmon
(most gear) and the Arctic lamprey (tow nets and fyke
nets), were collected in many locations, whereas
others, such as the burbot and least cisco, were col-
lected in only a few locations. Local concentrations of

some predators coincide in time and place with migra-
tions of juvenile sockeye salmon, for instance the lake
trout and Arctic char at the outlet of Coville Lake and
Arctic char and northern pike in parts of Grosvenor
River. Intensive study of each species is needed to
determine its abundance and role in the ecology of the
system.

ABUNDANCE OF JUVENILE
SOCKEYE SALMON

The tow netting to determine abundance of juvenile
sockeye salmon was largely exploratory in 1961 when
some areas and depths were sampled frequently and
others not at all. From 1962 to 1964, however, the
sampling was done systematically by season, area, and
depth.

Although it has never been firmly established, the
assumption that changes in the abundance of juvenile
sockeye salmon in tow net catches reflect actual
changes in their abundance has proved to be a work-
able hypothesis. The work of Pella ( 1968), who used a
recording echo sounder in conjunction with tow net-
ting, showed that tow netting is at least a good index of
relative abundance of sockeye salmon in the area
being sampled. The validity of tow net sampling for
measuring the abundance of juvenile sockeye salmon
was further substantiated in the present study: catches
declined in the lake from which fish were migrating
(Coville Lake) and increased in the lake to which they
migrated (Iliuk Arm).

Assessment of the abundance of fish in the pelagic
areas of the lakes is based on tow net data from sam-
pling mainly at night. Night sampling with tow nets
proved to be successful in western Alaska (Burgner et
al., 1969), although workers in British Columbia found
it best to sample with tow nets only during the transi-
tional period from dusk to darkness (Johnson, 1956;
Ruggles, 1966). Echograms and the results of concur-
rent tow netting by Pella (1968) in a lake in western
Alaska demonstrate that juvenile sockeye salmon re-
main dispersed at night near the surface in pelagic
areas of lakes. I found no consistent differences in the
rate of catch of juvenile sockeye salmon in tow nets
during different parts of the night in the Naknek sys-
tem. Some of the tow net data from Iliuk Arm were
collected in daylight because the water was so opaque
that sampling was apparently as effective in daylight as
in darkness.

The average catch per tow for four tows — two sur-
face and two deep — was used as the standard unit of
measure of abundance. The relative abundance in the
two depths frequently varied between lakes within a
year and between years within a lake. When unequal
numbers of tows were made at the two depths, the
averages for the two depths were averaged to give
equal significance to each depth. The only exception
to the use of this standard was for sampling in area
N-l, which was too shallow for deep tows.

In 1964 to compare the fishing capabilities of the two

■s

. fi

i a

o c p

u o >-.

w S -G

H ■=! h

U CB o

2-5'

-> a.

■6

p lo I u

: o

lo r-

i . i

I— ( (N

i— (

CD tO

i i i

cn <*

i— 1 H

r-~

r-j to

r*-

-3-

i O -=f

to to

CN O

r-. r-.

i oo cn

•3- ««*■

i— f i-H

CO CNJ

, ,

CD i

. ,

cn .

i— i

u <
H

1 CTi

CN 1

*3"

*3-

CN i
CD

r- i

■—I

i r-

^ rH

«*

r- «*

i CD

o o

•<*

CN
r-H

LO .

CD i

CT>
rH

O CO

cn rsi

rH tO

, ^H

--H *^t

LO tO

1 !-H

i-H tO

[

t--

J [

r- i

. E ?-n^

if) nj

"■ c

o H col

^•3:3^

< u

5-§

s h « .

"4H O t

■a -.cu jd
- &J(-> c |

3 ro
o. o4 3 «

u CD

O ro

DO r-t

• a ro

to -H

: ^h uj

>. 3

3 M ^H ,

i . o of> ro
■ O. (-• E S
WIJ u ii l

uj< a, <

w p. ro

-r

CNJ

i— 1

CNJ

r-.

CD i— I rH

cn cn
t lo

...

rolr-H CT3 OJ

-i -h 3 'J

tfl

4-»

^-^

,«

■ H

-1

■

r-C

o <u

""3

3 +->

H U OT)

P< • +-> <+h

h p,njifi oj

3 lo cj ai c

O -H

LO C i-H rH rH

tL,

■h a) ni a

£ rH O O 3

h g^^co

tO LO

O LO

iO o

CO LO

*^J- vO

CD O

Ol H-)

r-» o

CO Li,

rH rH
r-t IO

o r -

O rH
LO -3"

LO to

rH CD

CO -o
o r-j

«-o vO
r-- en

CTi

•H

t -

CO T-J rH

o o o

to rH r--

O LO O

G 'vi m

o 'H -a

C^ .H

CO OJ DVLJ I

tfl

u o
w 6 -

tJ 3

V) L)

f- p «)

O O trt

5)1.^1-*:
in - .

3 2,'S

: c c a
o o

a r4

a, o <a -.

E tfl ^ in

CO .Q O 4-1

— JIPhU m -<; I

—' g ^ CO c v

X -J a.

lii

11 01

£'

...

U) -*J

.5 "3

< u

•H

42 o

r^ '-»

(fl <J

'(/) 05

■M -H

J-< 'H

■H P<

Jq cT

-S r>l

Is to

S U

^> ",

1)^

c &

X-3

I'S"

■SI

O.
CO LO I

Uj|.r

i ^ ai+JH

i .e c3 'm
i u -i <

tn .c

^H t/1

Oj-r-1

in p.

T) XI1"

rH U
CO J3

P. in

- j j ._

CO rH CO | U
rH|.rt 3 'J

P.4H
in in

QJ CO
11 U

■fj

tfl

+->

afl

sw.s~

rH +J

c«

3 3

*-» >.

t/1

U CL>

U OT.

ITi rH

.— 1

r--

• +-> m

— 1

P.CUH O0

DO CO

ifl U H) C

3 +-»

G rH rH rH

rH +J

H CTj Ctf fi,

+J o

CU4-» P g

I/) 'J

rH O O C3

CTj ^— '

•H

pfHHU)

+->

CD T3
C-. CD

3 g
*J c

P.-H
CO +-J

o cri

■ ;

+-»

,-j

l-c

•H

*-■ a

*-< ^<J

CD CD

i'S

xcd

3
a-

-a

«-o oo

en o

r-j o

ro r-o

rH o

t~-- rH

O rH

rH "3-
tH r--

^-^

■6

,—N 0

^ s

■ 1

to 3

rH C

+-> co

ca Q

JC o

OjtrH

co £

co co

■j

^:u

•H -3

0) 03

—

4-1 .H

■. '.>^ to

CD P.

(Jl

O 0 1 CD c

^— '"— '.^ <D O

■H l/l

u ^-> >. e

J3 3

o C

c c o ca

01 -H

5 C

u o

O O CO ~4

^ ca

+->

co S

6 E >-

L> to

-^ DC

•H i — 1

i-H H 0) <U

U CD

CJ ca

sun n

co ~6

co t-.

co co -h o

.O O

+-»

G •. <-< O

-5

ECU

to M

ca ti

g O ID OH c

3 jd :> t— ' 3 -h

— i

l_) CJ

-3

< CJ

LO \D

r—i

-3-

rH rH

r-- lo

rH t--

bti

5^3

p.
ta col

£•

01" a

t+H O

a, +j c

CO I-H

■3.3

3 ■-" t-H

H ^ ai +JH
-rH M ^ U *-■
Q X cO h

ch u

at ^ w

r< OJ

r^ U 3

o a

•h 3 E

+-> CO

CO P, CO O

rH 1)

rH CO

-P

<d g

J3 -H

c a> w

H tV) O

e o

3 73

1) O P

TJI «Q

■M bflU

a =- co >-< 1.

i— i

o o

■s

nl i
o|u

r-tLrH

CO to

LO rH

rH ■

rH ^r
tH O

^3- ^f

LO rH

r- rH

-H

Cs| Kl

O rH

rH CD

O O

vD -3-

• r-

O LO rH

t-- CO

r>-

vO rH

Csl CO

rH CH

a> r- ^3-

rH O CO

OOH

;&,

■H

' )

+-»

D. 3

i-j

3 h;

rfl DO t/1

rH r*

-<! «

U CO (

H CO v

(J o

CO u

LJl

c
p.

H fljCJi

3 to O
U

CO C rH .

■H CO

E^r-(

c3 o -a

'3 3

CO CO

shapes and sizes of the tow nets used — a 3-m-diameter
round net and a 2.7-m-square net — I made 22 paired
tows using the two nets alternately. The sampling con-
sisted of 44 standard tows at the surface on four dates
in three sampling units (two in Coville Lake and one in
Iliuk Arm). The catch of age 0 sockeye salmon for
each tow and the figures used to test the hypothesis of
no difference in the rate of catch of the two nets with
the Wilcoxon matched-pairs signed-ranks test (Siegel,
1956) are presented in Table 4. The derived T = 53
with N = 19 is greater than the theoretical T = 46 at
the 5% level of significance, so the hypothesis of no
difference in the effectiveness of the two nets is ac-
cepted. The results of sampling with the two types of
nets are combined in my subsequent analyses.

Most of the 1961 tows were made only at the surface
and their comparability to the data collected at both
depths in later years was uncertain. A split-plot
analysis of variance (Snedecor, 1956) was made to de-
termine if the average catches of juvenile sockeye

salmon in surface tows and deep tows were signifi-
cantly different. This analysis was restricted to Coville
Lake, Iliuk Arm, and South Bay, 1962 to 1964, be-
cause of lack of sufficient data from other areas. The
model is

■ijk = ^+ Bi +

+ eU

+ T + (MT)n. +o

>jk ^°ijk.

where X

ijk

(MT)jk

°ijk

average catch at the Ath depth in the

jth time period in the z'th area,

block effect, i.e., area,

fixed main treatment effect, i.e., time,

whole plot error, i.e., area-time

interaction,

fixed subplot treatment, i.e., depth,

time-depth interaction, and

subplot error, i.e., area-depth plus

area-time depth interactions.

The split-plot analysis involved only surface and
deep paired tows (the subplot treatment), three time

Table 4. --Numbers of age 0 sockeye salmon taken in 22 paired tows in 3-m-diameter round and 2.7-
m-square nets in two sampling units in Coville Lake and one in Iliuk Arm, 1964, and calculations
used in Wilcoxon matched-pairs signed-ranks test (Siegel, 19S6) .

Age 0

date

sockeye salmon in--

Rank of sign

Sampling unit,

Round Square

of least

and sample number

net net

Difference

Rank

frequency

Coville Lake C-

August 8

43 503

-460

13 12

1.5

14 10

3.5

4 41

-37

12 12

8 8

Coville Lake C-

•4

August 18

1 2

-1

1.5

3 3

7 21

-14

9 18

-9

9 14

-5

8 4

3.5

Coville Lake C

-5

August 21

14 49

-35

27 46

-19

10.5

56 37

10.5

49 117

-68

118 49

17 25

-6

Iliuk Arm N-13

September 1

288 97

191

83 148

-65

19 97

-78

60 68

-8

Total (N = 1

862 1,579

—

53.0

Table S. --Subplot portion of a split-plot analysis of variance of catch of juvenile sockeye
salmon (age 0 and I combined) in tow nets in three lake basins of the Naknek system, 1962-64.
Depth of tow (shallow and deep) and depth- time interaction are tested.

Basin and

year

Source

Coville Lake

1962

Depth

3,879.54

2.15(NS)

Depth- time

4,353.78

2.41 (NS)

Error

1,804.51

1963

Depth

55,472.78

3.93*

Depth- time

43,788.31

3.10*

Error

14,127.66

1964

Depth

57,483.14

2.32(NS)

Depth- time

38,489.71

1.55(NS)

Error

24,830.54

Iliuk Arm

19621

Depth

649.74

<1(NS)

Depth- time

991.26

1.175(NS)

Error

844.23

1963

Depth

718.84

<1(NS)

Depth- time

564.23

<1(NS)

Error

2,686.76

South Bay

19622

Depth

63.43

<1(NS)

Depth- time

4,830.58

3.41 (NS)

Error

1,415.72

1963

Depth

2,007.46

1.081 (NS)

Depth- time

7,833.65

4.22*

Error

1,857.66

!No 10- to 20-ft tows made in midperiod in 1962.
2No samples from area N-6 in late time period.
*lndicates 101 level of significance.

periods each season (pre-26July, 27 July to 10 August,
after 10 August — the main treatment effect), and the
various number of areas within each lake (the block
effect). Because of unequal numbers of observations
per cell, the analysis was done with untransformed
data consisting of one observation per cell — the mean
for the area-time period-depth. In only two instances
did a significant difference appear in the subplot
treatments involving depth of tow (Table 5), i.e., there
were no consistent significant differences in catches of
juvenile sockeye salmon in surface versus deep tows.
Because of the indicated lack of difference between
surface and deep tows in the lakes with the largest
catches and the most sampling. I have assumed that
the surface catches in 1961 reasonably represent the
abundance in the surface to 20-ft depth. Pella (1968)
did not find a significant difference in abundance of
juvenile salmon with depth in Lake Aleknagik.

A two-way analysis of variance among areas and
times within lakes of average catches of juvenile sock-
eye salmon in tow nets was made for those lakes with
the most useful data — Coville Lake, Iliuk Arm, and
South Bay (Table 6). These lakes had the most sam-
ples and usually had the largest catches and the
greatest changes in abundance. Only averaged paired

tows (one shallow and one deep for the same night and
area) were used in the analysis. The analysis was done
with the same untransformed data as in the split-plot
analysis. However, the error terms used for the F tests
were obtained by using the individual catches in each
area-time cell (resulting in more degrees of freedom
than in the split-plot analysis) as suggested by Scheffe
(1959).

Statistically significant effects of areas, time, and
area-time interaction on the abundance of juvenile
sockeye salmon occurred in less than half the tests (17
of 39). Although the effects of areas and times were
frequently not statistically significant, the differences
observed were usually consistent from year to year
and agreed with the observed changes (such as inter-
lake migrations) and with the observations that num-
bers of age 0 fish increase during the first part of each
season and decrease later each season. I have, there-
fore, presented the quantitative results of the tow net
sampling in general summaries consisting of bar and
line graphs.

Trends in Abundance for the Entire System

Some stocks of juvenile sockeye salmon in the
Naknek system begin to migrate oceanward as soon as

Table 6. --Two-way analysis of variance of abundance of juvenile sockeye salmon in tow net catches
in selected lakes of the Naknek system, 1961-64. The analysis involves effects of areas and time
where one pair of surface and deep tows for each area and time was treated as one sample except
for Coville Lake in 1961, when each tow was a sample.

Year

1961
(age 0)

1962
(age 0)

1963
(age 0)

1964
(age 0)

1961
(age 0 and I)

(age 0)

1962
(age 0 and I)

1963
(age 0 and I)

(age 0)

1961
(age 0 and I)

(age 0)

Source

Coville Lake

Area

6,088.739

1.985(NS)

Time

469.205

<1(NS)

Time -area

109.192

<1(NS)

Error

3,067.286

Area

173.544

<1(NS)

Time

41,369.197

6.954**

Time -area

443.344

<1(NS)

Error

5,948.760

Area

44,960.561

1.007(NS)

Time

71,657.853

1 . 605 (NS)

Time-area

19,975.607

<1(NS)

Error

44,637.880

Area

75,014.098

5.470*

Time

51,733.660

2.393(NS)

Time- area

55,299.560

2.558**

Error

21,615.91

Iliuk Arm

Area

96.510

<1(NS)

Time

435.171

<1 (NS)

Time -area

670.576

1.245(NS)

Error

538.62

Area

61.450

<1 (NS)

Time

467.351

1.147(NS)

Time -area

505.180

1.240(NS)

Error

407.438

Area

3,311.066

55.375**

Time

3,255.427

54.760**

Time -area

6,139.228

65.691**

Error

93.598

Area

7,316.271

7.266**

Time

5,149.771

5.128*

Time -area

7,256.416

7.207**

Error

1,006.865

Area

1,370.424

4.118**

Time

299.361

<1(NS)

Time -area

1,221.070

5.670**

Error

352.756

Sou tli Bay

Area

494.114

3.668*

Time

187.391

1.391 (NS)

Time -area

741.177

5.502**

Error

134.715

Area

223.296

10.401**

Time

406.916

5.708**

Time -area

243.122

6.219**

Error

Year

Source

South Bav--Cont.

1962

(age 0 and I)

1963
(age 0 and I)

Area
Time

Time- area
Error

Area
Time

Time -area
Error

5.445
5,066.558
7,429.308
2,595.589

<1(NS)

1.952(NS)

2.862*

2,386.507
7,099.841
1,226.840
2,606.070

<1(NS)

2.724(NS)

<1(NS)

indicates 101 level of significance.
**Indicates 5?<> level of significance.

they leave the spawning grounds, although they do not
actually enter the ocean until the spring or early sum-
mer of their second or third year. As a result, the
numbers of age 0 and age I sockeye salmon increase in
the basins closer to the outlet river, while the number
of juveniles in the system is declining gradually.

I, therefore, evaluate mortality of juvenile sockeye
salmon in the Naknek system by examining the abun-
dance data for the system as a whole. For 1962 I was
able to calculate an average catch per tow by age class
each day for the system from 10 July to 29 August.
The sampling was done quite regularly and, in general,
each sampling unit shown in Figure 1 was sampled
once every 2 wk. By assigning the catch per tow found
by averaging the most recent preceding and following
sampling in each unit to those days on which no data
were collected, the weighted (by the surface area of
each sampling unit) rate of catch was calculated for
each day for the entire Naknek system. The data were
smoothed by a moving average of 3 (giving the middate
a double weight) for age 0 and age 1 fish (Fig. 3). Three
general time periods of abundance for age 0 fish appear
in these data: (1) the early period, when catches were
increasing — before 26 July; (2) the middle period,
when catches were generally stable — from 26 July to
about 10 August; and (3) the late period, when catches
were decreasing rapidly — after 10 August. The rate of
catch of age I fish decreased gradually during the sea-
son.

The mean catch per tow by lake and the contribu-
tion of each lake to the catch for the entire system for
the early, middle, and late time periods were calcu-
lated for each year from 1961 to 1964 for age 0 and age
I sockeye salmon. In 1961 and 1962 tow netting was
done some place in the system on most nights from
early July to early September, so that the averages for
individual time periods may represent a period of as
many as 20 days. In contrast, the data for the respec-

tive periods in 1963 and 1964 were collected within 2
days of 10 July, 1 August, and 29 August. Therefore,
the figures for the early and late time periods are the
results of shorter periods of mortality in 1963 and 1964
than in 1961 and 1962.

The weighted daily mean number of age 0 sockeye
salmon caught per tow for the entire Naknek system in
1962 and the means for the early, middle, and late
periods in 1963 and for the late period only in 1961 and
1964 are shown in Figure 4.

Only a general relation exists between the abun-
dance of age 0 sockeye salmon in tow net catches in
late August and the number of resulting smolts (Table
7). In 1961-63 the mean number of age 0 fish in the
catches ranged from 8.8 to 13.2, and the number of
resulting smolts (ages I and II) ranged from 11 to 16.7
million. Age 0 fish were about 1 .5 times as abundant in
1964 as in the other years and produced an unusually
large number of smolts — 14.7 million age I (about 25%

2°-

£ xlO

5 4

o AGE 0

°oo

o° o0 o0oooo • AGE 1
o°° • .

°Oo O ° o

•

0
0

••

1 1

JULY

25 |

AUG.

25 5

SEPT.

Figure 3. — Weighted daily mean number of age 0 and age 1 sockeye
salmon per standard tow in Naknek River system (all lakes combined),
11 July to 29 August 1962. The mean catch for the system was weighted
by the surface area of each sampling unit and the daily estimates were
smoothed by a moving average of three — (A + 2B + C) ■*■ 4.

?22

w 20

u. 16
o

rr 14

z 10

uj 8
o 6

£ 4
o
uj 2

■

▲ 1961

■

o 1962
• 1963
■ 1964

o" o0 o°ooo
o°o o „

o o °°« » °
oo o

0
0

°"°"\ A

oo •

•

°<x>„ •

1 , 1 I 1 I ■ 1

15
JULY

15
AUG.

SEPT.

Figure 4. — Weighted daily mean number of age 0 sockeye salmon per
standard tow in Naknek River system (all lakes combined) for 1962 and
means for early, middle, and late periods for 1963 and for late period
only in 1961 and 1964. (See Fig. 3 for explanation of weighting proce-
dure.)

more than in the previous high year) and 20.8 million
age I and age II combined.

The systemwide average catch per tow of age I
sockeye salmon generally declined slowly from July
through August each year and the abundance after 10
August ranged from 2.0 to 5.4 from 1961 to 1964 (Table
7). A decrease in abundance of age I fish was expected
because of a continuing outmigration of smolts during
the summer and natural mortalities. The range of
abundance of age I fish after 10 August each year was
similar to the range in abundance of age II smolts the
next spring (age III smolts are rare in the Naknek
system).

Comparative Abundance among Lakes

Although age 0 and age I sockeye salmon were
commonly captured together, the abundance of each
age class is considered separately because of differ-
ences in behavior and distribution among the lakes.

The general picture of the relative seasonal abun-
dance of age 0 and age I sockeye salmon in tow net
catches in each lake for 1961-63 are shown in Figures 5
and 6. Because it was not always feasible to maintain
the sampling schedule, the data in Figures 5 and 6 are
not complete for all years and all time periods. The
great difference in the scale of the ordinates of Figures
5 and 6 should be noted: only general sampling periods
are indicated in the graphs because I wish to consider
only the seasonal trends in abundance.

The most marked changes in the abundance of age 0
fish (and changes involving the greatest numbers of
fish) are the decreases in Coville Lake and concurrent
increases in Iliuk Arm for each time period (Fig. 5);
similar but much smaller increases for age 0 fish ap-
pear in Grosvenor Lake and South Bay. The decrease
in abundance of age 0 fish in Coville Lake and the
increase in the other lakes are due in part to an ob-
served downsystem migration of age 0 fish. This sum-
mer outmigration is probably significant in Coville
Lake, but the significance of these fish to the rest of
the system is uncertain (discussed in more detail later).
Part of the increase in numbers of age 0 fish downlake
from Coville Lake was due to a continuing recruitment
of fry from spawning grounds directly tributary to the
lakes, as indicated by the occurrence of the small fish
in samples taken in late July and most of August in all
years.

The abundance of age 0 fish in each lake in July is

Table 7. --Mean number of age 0 and age I sockeye salmon taken in tow nets in the Naknek River
system (all lakes combined) in August 1961-64 and resulting numbers of smolts produced. Age 0
fish in August can become age I or age II smolts, but age I fish in August can become only age
II smolts (rarely age III).

Fish in

parent

Mean number

Age of fish

and
)ling

escapement
(thousands)

tOW

fish per
net catch

Smolts

produced1 (m:

11 ions)

year of samj

Age I

Age II

Total

Age 0

1961

828.4

11.9

8.0

8.7

16.7

1962

551.1

15.2

6.0

5.0

11.0

1963

725.1

8.8

2 . 2

9.9

12.1

1964

905.4

25.0

14.7

6.1

20.8

Age I

1961

2,251.8

1.9

8.5

1962

828.4

4.7

8.7

1963

351.1

5.2

5.0

1964

725.1

5.4

—

9.9

Stewart, Donald M. (editor). 1969.
Dep. Fish Game, Inform. Leafl. 154.

1967 Bristol Bay red salmon smolt studies. Alaska

MIDDLE

160

EARLY j , LATE

EARLY, , LATE

EARLY LATE

EARLY, , LATE

1961

100

19 62

5
o
i- 60

- '' \ \ \

40
30

i
. / /

40
30

IT
bj

\ \

/ /

"•40

X
10

\ \

"- 20

— /

y/?

<r 0

COVILLE LAKE

GR0SVEN0R LAKE

ILIUK ARM

SOUTH BAY

z 50

a 40

I 30

/ \

/ i

/ ^

- ,/V

/ I96l\

" / ""

1 ^S 1961

^ ■"

WEST END

NORTH ARM

NORTHWEST BASIN

BROOKS LAKE

Figure 5. — Weighted mean number of age 0 sockeye salmon per standard tow by early, middle, and
late time periods in each lake of the Naknek River system 1961-63.

MIDDLE

LATE

MIDDLE
EARLY, , LATE

MIDDLE
EARLY, .LATE

MIDDLE
EARLY, , LATE

WEST END

NORTH ARM

NORTHWEST BASIN

BROOKS LAKE

Figure 6. — Weighted mean number of age 1 sockeye salmon per standard tow by early, middle, and
late time periods in each lake of the Naknek River system 1961-63.

related only in a general way to the abundance of po-
tential spawning grounds per unit lake area (Table 8).
The largest catches of age 0 fish came from the lake
with the greatest amount of probable spawning
grounds per unit of lake area — Coville Lake yielded
about 96 fish per tow and has 3.32 ha of spawning area
per square kilometer of lake. The lowest densities of
age 0 fish were generally in basins that had the lowest
ratios of spawning grounds to lake area — Northwest
Basin has 0.02 ha of spawning ground per square
kilometer of lake and North Arm has 0.04. The excep-
tion to this is Brooks Lake which has an intermediate
abundance of spawning area (0.24 ha), but a low abun-
dance of young sockeye salmon. The other lakes had
variable catches of age 0 sockeye salmon, seemingly
independent of the abundance of their spawning
grounds.

The lakes fall into three groups in terms of abun-
dance of age I sockeye salmon (Fig. 6): (1) lakes that
never have many age I fish and usually none after
July — Coville and Grosvenor Lakes and Northwest
Basin; (2) lakes that usually have a few age I fish all
summer — Brooks Lake, North Arm, and West End;
and (3) lakes that have many age I fish through the
summer — Iliuk Arm and South Bay. The last two ba-
sins constitute only about 25% of the system's surface
area, but contain about 90% of the age I sockeye salm-
on in the July to September period. The decline in
abundance of age I fish in Iliuk Arm and South Bay
each summer is concurrent with the downsystem mi-
gration of age I fish into these lakes from Grosvenor
Lake and the continued outmigration of smolts from
the system via the Naknek River.2

Abundance in Each Lake of the System

The preceding section described in general terms
the abundance of juvenile sockeye salmon in the
pelagic areas of the system as a whole. This section
will discuss the abundance of juvenile sockeye salmon
in the lakes and connecting rivers in detail and some of
the factors affecting it. To facilitate comparisons
among the lakes. Figure 7 shows the number of age 0
fish for the early, middle, and late time periods by
sampling unit in each lake of the Naknek system for
1961-63. The 1964 data are not shown in Figure 7 be-
cause they are complete only for Coville Lake.

Coville Lake. — Studies of juvenile sockeye salmon
in the Naknek system were gradually concentrated in
Coville Lake because it seemed to have special fea-
tures which would facilitate understanding the
dynamics of the population. These features are: (1) an

Table 8. --Relative abundance of spawning grounds
and average catch per unit of effort of age 0
sockeye salmon in early July 1961-63 in lakes of
the Naknek River system.

Area of poten-

tial spawning

Age 0

grounds per

sockeye

unit lake area

salmon

Lake or basin

(ha/km2)

per tow

Coville Lake

3.32

West End

0.68

Grosvenor Lake

0.40

Iliuk Arm

0.37

Brooks Lake

0.24

South Bay

0.07

North Arm

0.04

Northwest Basin

0.02

2The main smolt migration is complete and sampling was usually
ended by late July, but the migration was sampled intermittently in
August 1956 and 1958 and to 7 September 1962. The smolt migration
extended through August, but involved relatively few fish. (H. W.
Jaenicke. National Marine Fisheries Service. Auke Bay Fisheries
Laboratory, Auke Bay, AK 99821, pers. comm.l

abundant population of fast-growing juvenile sockeye
salmon and associated species; (2) a single major
spawning stock of sockeye salmon; and (3) a narrow
lake basin with the major source of sockeye salmon fry
at the end opposite the outlet. This combination of
characteristics simplified sampling and offered a better
opportunity for detecting gradients in biological condi-
tions.

The mean rate of catch of age 0 sockeye salmon in
tow nets in sampling units of Coville Lake is shown in
Figure 7 for the standard time periods for 1961-63 and
in Figure 8 for several time periods for 1961-64. In
1961 the abundance differed markedly from the other
years in that an early-season (about mid-July) high was
not observed. Although it may be that sampling in
1961 began after the early-season maximum of abun-
dance, the pattern of recruitment of fry from American
Creek, the major source of fry to the lake, and mortal-
ity in the lake may have been quite different in 1961
than in 1962-64, as indicated by the greater abundance
of age 0 fish at the end of the summer in 1961 (Fig. 7).
The catches of age 0 sockeye salmon in the lake de-
clined markedly through the summer in 1962, 1963,
and 1964 and were similar at the end of August each
year. The analysis of variance showed significant dif-
ferences in abundance due to time only in 1963 (Table
6).

Because the major source of juvenile sockeye salm-
on in Coville Lake is at the end opposite the outlet, a
gradient in abundance and possibly in size of juveniles
might be expected. To increase the chance of detect-
ing such a gradient, the downlake sampling area — unit
C-2 in 1961 and 1962 (Fig. 1) — was divided into three
units in 1963 and into four in 1964 (Fig. 2). In Figure 8,
C-l is the sampling unit closest to American Creek in
all 4 yr and C-2, C-2L, and C-5 are the units closest to
the outlet of the lake in 1961-62, 1963, and 1964 respec-
tively.

Neither abundance nor size of age 0 sockeye salmon
in Coville Lake showed a gradient from the source to

sndmws on

I 1

SldWVS ON

i i —

1 Oh

(71

■

, 1 1

SOh

33diNVS ON

_l I I 1-

g£

9 °

5 Z

tt CO

< m

^ I

31dlNVS ON

dj 5
o <

■Z O
D 2

aidwvs on

■
1

31dWVS ON

■

1
i i i l _

I <M

33dWVS ON

31dWVS

^^^■■H

^^B^BHI^^^^H

i i I i

I en

31dWVS ON

8 0
£0

.■r o

I «

M01 d3d HSU dO M38WnN NV3W

c —
{ u

■ 3

— **

Of =

o. -

(/}

C-

£-

—

•=

—

-
a

jr.

•-

5
-

E 2

"3 -

/■

■a

0
<u
u
a

—

■-I

—

-C

4- .-

eZ U £

CVJ

■

^^^

<-.

/■~ ■

_r "

■

^ ■

■

c\j
. 1 1 1 1 1 1

-m

1 1 1 1 1 > 1 1 1 1

2£
3 o

1961

—

— ■

—

^™

i » i

o *■

■o 3

3 2

■8 1

[/)

_£

- E

E S

■a =
S .Sf

O T5

J i

s -

O .3
- 3

0* S

o o o o

~ o <r> cd

N01 H3d HSId dO a39W0N NV3W

C **

c o» ■*■>

S 5 s

I £ a

«u J

«» " —

u X u

3 C .=

J. Q. £

Table 9. --Mean fork length (X) and standard deviation (SD) of age 0 sockeye salmon taken in tow
nets in sampling units of Coville Lake and in fyke nets in Coville River by time periods, July 11
to September 1, 1963.

Sampling

July 7-11

July 19-21

July 51 -Aug. 1

Aug. 15-17 Aug. 31-Sept. 1

unit

X length

length

C-1

C-2U
C-2M
C-2L
Coville
River

34.8
47.0
44.2
41.2

+3.8
+ 3.0
+3.4
+3.9

41.6

47.4
45.2

49.8

+ 5.2

+4.8
+4.5

+4.6

51.0

54.0
51.3
53.4

51.7

+5.6
+6.3
+6.1
+4.8

+5.3

53.6
57.1
51.1

56.6

+5.3
+5.3
+6.9

+5.1

57.0
57.8
62.0

57.4

61.5

+9.1
+6.5
+3.4
+8.3

+5.4

the outlet. In 1963 and 1964 the second sampling units
downlake (C-2U and C-2 respectively) yielded fewer
fish per tow than did the adjacent units uplake (C-1) or
downlake (C-2M and C-3) in July and the first part of
August. In mid- and late-August the rate of catch of
juveniles was greatly reduced in C-1 and it appears
that in 1963 many of the fish that had been in C-1 had
moved downlake into C-2U. Analysis of variance
showed significant differences in abundance among
areas of Coville Lake only in 1964 (the year when five
areas were sampled). In early July of both years the
average length of fish was smaller close to the" upper
end of the lake, but later, fish with the smallest aver-
age lengths were from the area closest to the outlet.
No dines or gradients among the areas appear in the
length data (Tables 9 and 10).

The migration of age 0 sockeye salmon from Coville
Lake to Grosvenor Lake began in mid-July, but did
not involve many fish until about the first of August.
In 1963 and 1964 juvenile salmon captured in fyke nets
in Coville River as they were leaving the lake were
generally larger than the average size of those remain-
ing in the lake (Tables 9 and 10). In 1964, when the
lake was sampled in most detail, the smallest average
size was found in the area adjacent to the outlet (Table
10). It appears that fish leaving the lake were the larger
members of the "normal" length frequency which re-

sulted in a smaller average size for those remaining in
the area adjacent to the outlet.

Grosvenor Lake. — The seasonal change in abun-
dance of age 0 sockeye salmon in tow net catches
from Grosvenor Lake was generally similar to that of
the system as a whole (Fig. 3) in 1961 , 1962, and 1963;
i.e., the abundance increased during July and de-
creased in the latter part of August (Figs. 5 and 7). The
only exception was the unusually large (for Grosvenor
Lake) average catch for one series of tows made in
areas G-l and G-2 in the sampling in the late period in
1963 (Fig. 7). There was no marked concentration of
young salmon early in the summer in the pelagic areas
near the major spawning grounds (sampling units G-3
and G-4 — Fig. 1).

The decline in abundance of juveniles in Grosvenor
Lake in August occurred during the time of an immi-
gration offish from Coville Lake via Coville River and
an emigration from Grosvenor Lake via Grosvenor
River. Several large catches of juveniles were made
with tow nets very close to the outlet of Grosvenor
Lake indicating that migrants accumulated here; these
data are not included in Figure 7 because this area was
not part of a regular sampling area.

Sockeye salmon in the 27- to 39-mm size groups
were recruited to Grosvenor Lake in middle and late

Table 10. --Mean fork length (T) and standard deviation (SD) of age 0 sockeye salmon taken in tow
nets in sampling units of Coville Lake and in fyke nets in Coville River by time periods, July 4
to September 5, 1964.

Sampling

July 4-

■6

July 16-

-17

Aug. 2-

-4

Aug. 15-

-18

Sept. 2-

■5

unit

X length

X" length

X~ length

X length

X~ length

C-1

54.2

+2.8

46.5

+2.9

50.5

+4.0

57.1

+4.9

59.3

+4.7

C-2

46.2

+2.7

52.6

+6.5

57.6

+6.0

59.4

+5.0

C-3

43.6

+ 2.4

51.3

+4.2

57.5

+4.7

58.0

+ 7.2

C-4

42.6

+5.7

50.2

+4.4

57.1

75.9

C-5

36.4

+5.4

42.5

+5.2

48.4

+4.2

49.7

+4.8

54.8

+5.4

Coville

River

54.9

+5.4

60.7

+3.3

61.1

+ 5.2

August 1961 and 1963, but relatively few fish in these
size groups were collected during this time in 1962 (see
section on length frequencies). These small and pre-
sumably late-emerging fish must originate within the
Grosvenor Lake basin because they have never ap-
peared in samples from Coville Lake or Coville River.
They may be progeny of adults that spawn in the shal-
low beach areas — the development of these fish could
be slower than normal because of low temperature and
oxygen levels accompanying severe winter conditions,
as has been reported for progeny of beach spawners at
Lake Kitoi (Smoker, 1957).

Iliuk Arm. — Iliuk Arm has only one known intraba-
sin spawning ground of significant size, Margot Creek,
but it has enough extrabasin spawning ground (includ-
ing one tributary to the Savonoski River above Gros-
venor River) to yield a spawning-area-to-lake-area
ratio intermediate for the Naknek system — 0.37 ha per
square kilometer (Table 2). In addition to the known
grounds, it is suspected that spawning occurs in beach
areas within Iliuk Arm and in streams tributary to the
Savonoski River. A variable recruitment to Iliuk Arm
from Hardscrabble Creek in Grosvenor Lake is
known to occur, but the potential spawning ground of
Hardscrabble Creek is not assigned to Iliuk Arm. The
greatest number of recently emerged fry and interlake
migrants probably enter Iliuk Arm through Savonoski
River in sampling unit N-15 (Fig. 1).

The abundance of age 0 sockeye salmon increased
in Iliuk Arm in the summers of 1961, 1962, and 1963
(Fig. 5), but the relative abundance in the three sam-
pling areas varied from year to year (Fig. 7). This
variation may be due to year-to-year changes in the
relative number of recruits from different sources. The
trends in abundance in each of the sampling areas dur-
ing the summer in 1961 and 1963 are similar — mortality
or migration from the basin exceeded recruitment in
August in N-15 and N-14 and resulted in a shift in the
center of abundance to the downlake end of the basin
by the end of August (Fig. 7). This did not happen in
1962 when a relatively intensive summer migration of
age 0 sockeye salmon down Grosvenor River was ob-
served. A similar migration occurred in Grosvenor
River during the summers of 1961 and 1963, but was
not well sampled. The analysis of variance showed
significant differences in abundance of juvenile salmon
(age 0 and I combined) among sampling areas, times,
and in time-area interaction for 1962 and 1963 (Table
6). The differences between areas and the time-area
interaction were also significant in 1963, but time alone
was not when only age 0 are considered.

In 1964 Iliuk Arm was sampled only at the end of
August, and at this time (as in 1963), age 0 sockeye
salmon were least abundant in the uplake unit, N-15.
and most abundant in the midlake unit. N-14. The
weighted average catch per tow for the entire basin at
this time was the greatest observed in Iliuk Arm from
1961 to 1964.

South Bay. — South Bay receives sockeye salmon fry
from both outside and inside its basin. The only sig-
nificant source within the basin is Brooks River in
unit N-ll, although some fry may result from beach
spawning along the south shore of N-6. Because
downsystem migrants are recruited from Iliuk Arm,
the major recruitment from outside the basin is also
into N-ll.

There was generally a gradient in abundance of age
0 fish in South Bay — the largest catches were made in
N-ll and the smallest in N-6 (Fig. 7). The greatest
abundance of age 0 sockeye salmon recorded for
South Bay between 1961 and 1964 was at the end of
August 1964; at that time (as in 1963) the catches were
largest in N-ll and smallest in N-6. Analysis of vari-
ance for age 0 indicated significant differences in
abundance among sampling areas, times, and time-
area interactions in 1961 and only in time-area interac-
tions in 1962 (Table 6). During the 4 yr of this study,
South Bay always had its greatest number of age 0 fish
at the end of summer.

West End. — The West End basin is the shallow out-
let end of Naknek Lake and is mostly less than 10 m
deep. It contains a few small spawning streams and the
extensive but essentially unknown spawning areas of
the upper end of Naknek River, i.e., between the
rapids and the lake. It is assumed that sockeye salmon
that hatch in the Naknek River move upstream into
West End shortly after they emerge. The evidence for
this is all negative — very few zero freshwater-age-type
adults return to the system,3 and the Naknek River
and adjacent lagoons are probably not extensive
enough to rear enough smolts to produce the large
number of adults that spawn in the river in some years.
However, sampling with tow nets in the early period in
unit N-l never yielded many age 0 sockeye salmon
(Fig. 7).

The abundance of age 0 sockeye salmon in the West
End in 1962 and 1963 (Fig. 5) was low in the early
period, high in the middle period, and low in the late
period. In 1964 this basin was sampled only on 3 Sep-
tember. Comparison of length frequencies of age Ofish
(discussed later) and their seasonal abundance in Iliuk
Arm and South Bay with similar data from West End
(Fig. 7) indicates that the movement of interlake
migrants during the summer does not continue on into
West End. This is also shown by the marked reduction
in abundance of age 0 fish through South Bay from
sampling units N-ll to N-9 to the unit adjacent to the
West End, N-6. Although it seems that the movement
of age 0 fish downlake does not extend through South
Bay by early September, the situation in Coville Lake
thoughout July 1963 should be recalled, i.e., areas of
greater abundance occurred both uplake and downlake
from an area of low abundance (Fig. 8) comparable to

3Unpublished data on file at National Marine Fisheries Service.
Auke Bay Fisheries Laboratory, Auke Bay. AK 99821.

N-l 1 (high abundance), N-9 and N-6 (low abundance),
and N-4 (high abundance).

The source of the relatively larger number of age 0
fish in N-4 in 1963 (Fig. 7) may have been the beach
spawning areas a'ong the south shore of N-6 (the base
of the north slope of Dumpling Mountain). But again
the situation may be comparable to the situation in
Coville Lake where juveniles were relatively scarce
over the deep water of the central basin but were
abundant over the shallower water at both ends. The
age 0 fish may have passed through or around the deep
water of N-9 and N-6 without being sampled. The data
for 1962 and 1963 in Figure 7 indicate this may have
happened. If it did, this movement involved more of
the larger fish, for the average lengths of the age 0 fish
were generally greater in the West End.

North Arm. — The North Arm basin is the largest in
the system in surface area (over water deeper than 5
m) and volume, but has next to the lowest ratio of
potential spawning area to lake area (Table 2). In addi-
tion to having little intrabasin spawning grounds, re-
cruitment of juvenile sockeye salmon to North Arm
from other basins is limited by the drainage pattern
— the flow of water into North Arm is surface runoff
via several small streams and the flow out is through
narrow channels and over shoals. The two factors
— little intrabasin spawning and little recruitment from
other lakes — resulted in the lowest abundance of
juvenile sockeye salmon observed in the Naknek sys-
tem (Fig. 7).

Northwest Basin. — The Northwest Basin appears to
be as much an entity as North Arm. The Northwest
Basin is small and comparatively deep and has only a
shallow connection to the rest of the Naknek system
via the West End. This basin has several small lateral
spawning streams along its north shore, but the ratio of
spawning area to lake area is the lowest in the system
(Table 2). The general abundance of age 0 sockeye
salmon is also quite low — only North Arm and Brooks
Lake produced lower rates of catch.

The type of spawning ground in Northwest Basin
— small lateral streams — is generally more intensively
utilized by sockeye salmon than the larger
intermediate-sized streams such as Bay of Islands
Creek in North Arm and Headwater Creek in Brooks
Lake (Burgner et al., 1969). The greater intensity of
use of the spawning grounds in the Northwest Basin
could account for the greater abundance of age 0 sock-
eye salmon there (Fig. 7) than in North Arm or Brooks
Lake, in spite of their larger ratios of spawning area to
lake area.

Brooks Lake. — Brooks Lake is similar to North Arm
and Northwest Basin in terms of abundance of
juvenile sockeye salmon (relatively low — Fig. 7), in
lacking recruitment from other basins, and in not hav-
ing a recruitment of fry in midsummer (based on shape

of the 1962 catch curve [Fig. 5] and length frequency
data). In late summer and early fall it is usual for few
to several thousand age 0 sockeye salmon to migrate
from Brooks Lake to South Bay.4

ABUNDANCE OF ASSOCIATED SPECIES

In general, the catches of fish other than sockeye
salmon in tow nets were not consistent within or be-
tween years either for species or basin. The effects of
such factors as spawning migrations and recruitment
of age 0 fish cannot be analyzed because there are not
enough data on age composition or length frequency.
Hence, only general comments can be made on the
abundance of associated species.

The three species most commonly associated with
sockeye salmon in tow nets were pond smelt, three-
spine sticklebacks, and ninespine sticklebacks. In the
sections that follow, the catch-per-tow data for these
three species were summarized by semimonthly
periods for 1961-63 for areas and lakes. The lake aver-
ages were derived by the procedure used with the
salmon data. Although most sampling was in the sur-
face to 3-m zone in 1961, the data for that year are
treated here as equivalent to those of 1962-64 because,
as with juvenile sockeye salmon, the split-plot analysis
of variance tests did not indicate consistent differences
between average catches at the two depths (Table 11).

Pygmy whitefish and least cisco were captured with
some type of gear in most basins of the system (Table
3), but because neither species consistently occurred
in significant numbers in the tow net samples, they will
not be discussed in this section.

Pond Smelt

Pond smelt occur in all basins of the Naknek
system5 and have been taken by all suitable gear, but
they occurred in large numbers only in samples col-
lected with tow nets in Coville Lake. In Coville Lake,
tow net catches of pond smelt fluctuated greatly and
erratically during the season (Fig. 9) — much more than
the catches of juvenile salmon, which were charac-
terized by a relatively steady seasonal decline (Fig. 5).
Age 0 pond smelt, which first occurred in late August,
were never the most numerous age group in the
catches.

The comparatively large catches of pond smelt in
Coville Lake in 1963 may reflect good survival of the

1962 year class. Generally favorable growing condi-
tions for fish in 1962 were indicated by the greater
growth of juvenile sockeye salmon that year. Signifi-
cant differences in abundance were indicated in the
split-plot analysis of variance for depth in 1962 and

1963 and for interaction of depth-time in 1963 (Table

4From Brooks Lake Field Station Reports. 1961-65, on file Na-
tional Marine Fisheries Service, Auke Bay Fisheries Laboratory,
Auke Bay, AK 99821.

5 Pond smelt have not been reported from Brooks Lake, but
have been seen in a tributary to Brooks Lake and in Brooks River
near the outlet of Brooks Lake (Heard, Wallace, and Hartman.
1969).

Table 11. --Split-plot analysis of variance of abundance in tow net catches of pond smelt (Coville
Lake), threespine stickleback (Coville Lake and West End) and ninespine stickleback (Coville Lake
and West End), 1962-64. Analysis involves only paired surface and deep tows and considers vari-
ation due to sampling areas, time (July 1-15, July 16-51, August 1-15, August 16-51), and depth
(surface versus deep) of tow.

Species and year

Source

Coville

Lake

Pond smelt

1962

Main plot

Area

5,160.7

2.84 (NS)

Time

864.88

<1(NS)

Error

1,819.94

Subplot

Depth

9,469.72

4.95*

Depth- time

1,733.01

<1(NS)

Error

1,911.32

1965

Main plot

Area

55,508.54

<1(NS)

Time

7,181.71

<1(NS)

Error

87,479.80

Subplot

Depth

56,921.61

6.18*

Depth- time

70,008.62

7.63**

Error

9,207.64

1964

Main plot

Area

1,578.45

<1(NS)

Time

1,845.53

<1(NS)

Error

1,974.71

Subplot

Depth

792.90

<1(NS)

Depth- time

1,535.92

1.07(NS)

Error

1,439.89

Threespine stickleback

1962

Main plot

Area

19,320.31

<1(KS)

Time

21,865.52

<1(NS)

Error

26,681.50

Subplot

Depth

50,989.20

1 . 39 (NS)

Depth- time

21,231.25

<1(NS)

Error

23,664.14

1963

Main plot

Area

152.52

<1(NS)

Time

425.19

<1(NS)

Error

683.69

Subplot

Depth

971.26

1 . 79 (NS)

Depth- time

410.51

<1(NS)

Error

543.64

1964

Main plot

Area

127.11

1 . 06 (NS)

Time

259.55

2.16(NS)

Error

120.10

Subplot

Depth

9.45

<1(NS)

Depth- time

2.93

<1(NS)

Error

43.03

Table 11. --Split-plot analysis of variance of abundance in tow net catches of pond smelt (Coville
Lake), threespine stickleback (Coville Lake and West End) and ninespine stickleback (Coville Lake
and West End), 1962-64. Analysis involves only paired surface and deep tows and considers vari-
ation due to sampling areas, time (July 1-15, July 16-31, August 1-15, August 16-31), and depth
(surface versus deep) of tow. --Continued

Species and year

Source

Coville Lake-

-Con1

Ninespine stickleback

1962

Main plot

Area

114.08

3.68(NS)

Time

,209.52

71.22**

Error

31.02

Subplot

Depth

,430.08

21.23**

Depth- time

31.02

<1(NS)

Error

67.37

1963

Main plot

Area

1.05

<1(NS)

Time

924.80

4 . 04 (NS)

Error

229.05

Subplot

Depth

81.45

<1(NS)

Depth- time

224.87

<1(NS)

Error

288.01

1964

Main plot

Area

12.96

2.84*

Time

23.75

5.20**

Error

4.56

Subplot

Depth

2.85

3.43*

Depth -time

0.50

<1(NS)

Error

0.83

West End (N4-

-N2 onD

Threespine stickleback

1963

Main plot

Area

568

,327.69

<1(NS)

Time

686

,189.58

<1 (NS)

Error

881

,226.75

Subplot

Depth

611

,782.52

<1(NS)

Depth- time

760

,305.58

1 . 06 (NS)

Error

716

,034.73

Ninespine stickleback

1963

Main plot

Area

,064.08

2.14 (NS)

Time

681.02

1.37(NS)

Error

497.90

Subplot

Depth

18.75

<1(NS)

Depth- time

422.69

10.5**

Error

42.04

* Indicates 10% level of significance.
**Indicates 5% level of significance.

11). The significant results in 1963 were due to a few
very large catches in the middle and late time
periods in surface tows only. No consistent differ-
ences are apparent.

Threespine Sticklebacks

Threespine sticklebacks were captured with tow
nets and all other suitable gear in all basins of the
Naknek system (Table 3). In general, the areas that
yielded only few juvenile sockeye salmon — North
Arm, Grosvenor Lake, and Brooks Lake — also
yielded only few threespine sticklebacks.

The outstanding feature of the abundance of three-
spine sticklebacks in the tow net catches is the varia-
tion from one sampling period to the next. The abun-
dance of threespine sticklebacks during each summer
from 1961 to 1963 in the West End (a region of great
abundance) by sampling area (Fig. 10) illustrates this
point. No significant differences in abundance with
time, depth, or area appeared in the split-plot analysis
of variance of data collected in 1962-64 in Coville
Lake, the lake for which most data are available
(Table 1 1). Catches of threespine sticklebacks resem-
ble those of the pond smelt (Fig. 9) in that the abun-
dance in tow nets fluctuated independently in adjacent
sampling areas.

The catches of threespine sticklebacks increased
during the summer in some lakes and were fairly uni-
form through the summer in others. Only a few
threespine sticklebacks were captured with tow nets in
the first half of July in Coville Lake, Iliuk Arm. and
South Bay, but in August they were taken in moderate
numbers in these basins. A similar increase in catches
during the summer occurred in the lakes where they
were never taken abundantly, i.e., Grosvenor Lake.
North Arm, and Brooks Lake. In the areas of rela-
tively great abundance, Northwest Basin and West
End, this species was about as numerous in catches
the first half of July as in late August. At Karluk Lake
on Kodiak Island, in 1961 and 1962, threespine
sticklebacks were abundant in the littoral areas and
virtually absent in the pelagic areas in early July, but
by summer they were mostly in the pelagic areas." A
similar shift to pelagic areas was found in Lake Nerka
of the Wood River system (Burgner, 1962).

Age 0 threespine sticklebacks were rare in tow net
catches until late August and even then they were so
small that they could pass easily through the smallest
mesh of the net unless their spines were erect.

Ninespine Sticklebacks

Ninespine sticklebacks and threespine sticklebacks
commonly occurred in the same catches and, in gen-
eral, the observations on threespine sticklebacks apply
to ninespine sticklebacks. The average abundance of

6B. Drucker. National Marine Fisheries Service. Auke Bay
Fisheries Laboratory. Auke Bay, AK 99821, pers. comm.

377

4 37

340
320
300
280
260
240
2 20

IE
UJ
O-

I 200
tn

o 180

2 160

140

120

100

40 -

20 -

•CI
• C -2

I I I

234 1234 1234

1961 1962 1963

SEMIMONTHLY TIME PERIOD

Figure 9. — Mean number of pond smelt per standard tow in Coville
Lake (units C-l and C-2) by semimonthly time periods, 1961-63. Time
periods are: 1 — July 1-15; 2 — July 16-31; 3 — August 1-15; 4 — August
16-31.

ninespine sticklebacks in the tow nets was markedly
lower in July than in August both in Coville Lake,
where moderate numbers were captured, and in Iliuk
Arm, South Bay, and Northwest Basin, where only a
few were captured. Four of the five significant differ-
ences in abundance shown in the split-plot analysis for

Coville Lake (Table 11) involve both time and depth,
which may reflect the offshore movement of adults
after the early summer spawning and recruitment of
yearlings to catchable size. The seasonal change in
abundance was not as evident in West End, where this
species occurred in greatest numbers. No consistent
year-to-year trends in abundance were observed.

The abundance of ninespine sticklebacks in tow net
catches exceeded that of threespine sticklebacks only
in area C-l of Coville Lake — this is the uplake end
adjacent to large areas of submerged aquatic plants,
mostly Potamogeton spp. The catches of the two
species were about equal in the rest of Coville Lake,
but in the other lakes ninespine sticklebacks were gen-
erally much less abundant than threespine stickle-
backs.

INTERLAKE MIGRATION OF
PRESMOLT SOCKEYE SALMON

Although juvenile sockeye salmon normally trans-
form to smolts and migrate to salt water at age I or
older, some oceanward migration of presmolts (age 0
fish) has been reported7 (Narver, 1968). Outmigrations
of presmolt sockeye salmon amounting to as much as
21% of the subsequent smolt production for the brood
year had occurred in Brooks River in 1958 and 1960
and again in 1961. In the summer of 1961 a similar
migration of age 0 fish from Coville Lake to Gros-
venor Lake via Coville River was sampled intermit-
tently. From these data, I estimated that several mil-
lion age 0 fish had left Coville Lake.

The results of the sampling at Coville Lake in 1961
prompted further studies to answer the following ques-
tions: (1) Do significant numbers of age 0 sockeye
salmon usually migrate from Coville Lake? (2) Do sig-
nificant numbers of juvenile sockeye salmon over-
winter in Coville Lake and migrate as age 1 smolts in
May and June? (3) Do the age 0 sockeye salmon leav-
ing Coville Lake during the summer remain in Gros-
venor Lake until they become smolts, or do they con-
tinue downsystem to Naknek Lake their first summer?
(4) What is the cause of the presmolt migration? (5) Do
the behavior patterns of these fish resemble those of
smolts or fry, or are they unique to summer migrants?
Information pertaining to these questions was
gathered by the routine sampling of the Naknek sys-
tem and by special studies in Coville and Grosvenor
Rivers in 1962, 1963. and 1964, in addition to the sam-
pling in Coville River in 1961.

Large and small fyke nets were both used to sample
migrating fish in the rivers. Although current ve-
locities were not measured, the small nets (1 m) were
fished in waters of about 0.3 meters per second (mps).
The large nets (1.2 x 1.2 m or 1.2 x 1.5 m) were
generally fished only in currents greater than 0.3 mps.

Newly emerged fry could pass through the wings and
body of the large nets, but would be retained in the cod
end and by all parts of the small nets. The small nets
were usually fished from stakes driven into the
streambed and the large nets were fished from a cable
strung across the river. The cod end of the large fyke
net was often connected to a floating box (Craddock,
1961) that held the fish so that they could be released
alive and uninjured.

In Coville River the estimate of the outmigration of
juvenile salmon is based on sampling with fyke nets
fished near the mouth where the river is about 24 to 46
m wide and 0.3 to 2 m deep; the current velocity is

7Wilbur L. Hartman. William R. Heard, and Charles W. Strick-
land. 1962. Red salmon studies at Brooks Lake Biological Field
Station, 1961. On file. National Marine Fisheries Service, Auke
Bay Fisheries Laboratory, Auke Bay. AK 99821, 53 p.

n — I — I 1 — l — I — r

234 1234 1234

1961 1962 1963

SEMIMONTHLY TIME PERIOD

Figure 10. — Mean number of threespine sticklebacks per standard
tow in West End (units N-4, N-2, and N-l) by semimonthly time
periods, 1961-63. Time periods are: 1 — July 1-15; 2 — July 16-31;
3_August 1-15; 4— August 16-31.

about 0.3 to 1.2 mps. The width of the stream was
divided into four equal segments and the middle 1.2 m
of each segment was sampled with a 1.2-m-wide fyke
net. Two sampling schemes were used: (1) regularly,
the site that passed the most water (and caught the
most fish) was fished as an index; and (2) at intervals,
based on observed changes in the character of the mi-
gration, nets were fished at the four sites following a
modified Latin-square design (Cochran and Cox,
1957) so that the number offish migrating in the entire
stream during the period could be estimated. In the
Latin-square scheme catches were classified accord-
ing to site, time of day, and days — factors considered
to have the greatest influence on variability of the in-
dividual catches. Estimates of the numbers offish that
migrated through the sites were obtained by fitting a
multiplicative model8 to the observed catches, esti-
mating numbers migrating through unsampled site-
time of day-day strata from parameter estimates of the
model, and then summing over all strata (sampled and
unsampled), and finally expanding this total to account
for the proportion of the river sampled by the nets.

The estimated outmigration of juvenile sockeye
salmon during the Latin-square and the number caught
in the index site during the same period (the index site
was fished continuously) were used to estimate the
portion of the total migrants captured at the index site.
This figure, the index catch expansion factor, is used
to estimate the number migrating when only the index
net was fished.

The estimate of the numbers of juvenile salmon mi-
grating out of Coville Lake during the period sampled
each year is based on a combination of the Latin-
square estimates and the index catches. For periods
when migration estimates from Latin-square sampling
were made, the daily migration was estimated by di-
viding the expanded Latin-square estimate by the
number of days involved; when only index sampling
was done, the daily migration was estimated by ex-
panding the catch in the index net by the index catch
expansion factor. The index catch expansion factor
was used up to the halfway date toward the next
Latin-square period and then the factor for the next
period was used. When no sampling was done for a
day or days, the average of the preceding and the fol-
lowing estimates was used. The estimated total migra-
tion for the season is the sum of the estimates for each
day.

In Grosvenor River juvenile salmon were sampled
with fyke nets at two general locations. Recently
emerged fry were sampled with the 1-m fyke nets in
the shallow water along shore near the outlet of the
lake. Older fish were sampled with the 1.2-m nets
which were attached to a cable at a point about 2 miles
below Grosvenor Lake, just above the island in Gros-

"The model was developed by Jerome J. Pella of the National
Marine Fisheries Service, Auke Bay Fisheries Laboratory, and a
full description and analysis of the model and its application will be
published soon.

venor River. The river was about 78 m wide and 0.6 to
2 m deep where the cable crossed and the water veloc-
ity was from 0.6 to 1.2 mps where the fyke net was
fished.

For purposes of analysis, I have summarized the
data by 10-day intervals; 1 August was arbitrarily
selected as the starting date.

Migration from Coville Lake to
Grosvenor Lake

1961. — In 1961 the sampling of juvenile sockeye
salmon migrating down Coville River was exploratory
and intended mainly to determine the timing and the
age classes involved. The sampling was done in two
periods — early (18 May to 1 1 June) and late (27 July to
10 September). During the early period the small
(1-m-square) fyke nets and seines were used and both
age 0 and age I fish were caught. During the late
period the fishing was mainly with the 1-m-square fyke
net and mostly age 0 fish were caught.

Because so many age 0 fish appeared to be involved
in the summer migration, I have made an order of
magnitude approximation of the number that migrated
from Coville Lake to Grosvenor Lake from 27 July to
10 September. Knowledge gained in subsequent years
makes the following assumptions reasonable: (1) the
fyke nets caught 4% of the juvenile sockeye salmon
migrating down Coville River during the periods
fished (based on portion of river sampled); (2) the rate
of catch during the time fished each day was typical of
the whole day; and (3) the catch per day can be aver-
aged for 10-day periods. Using these assumptions, I
estimated that in 1961 about 5.6 million age 0 sockeye
salmon migrated from Coville Lake to Grosvenor
Lake between 22 July and 10 September (Table 12).

1962. — In 1962 the migration of sockeye salmon
down Coville River was sampled from 29 May to 15
September and more systematically than in 1961. A
cable was installed across the river near Grosvenor
Lake where the river was about 24 m wide. The four
6-m sites were established on the cable and the large
fyke nets were fished in the middle of each site. Nets

Table 12. --Estimated number of age 0 sockeye
salmon migrating from Coville Lake to Grosvenor
Lake by 10-day periods between July 22 and
September 10, 1961.

Fish migrating
Period each period

July 22-51
August 1-10
August 11-20
August 21-30
August 31-September 10
Total

871,000

1,058,000

3,459,000

96,000

71,000

5,555,000

were fished on Latin-square schedules as follows: four
1.5-h periods (2100 to 0300) each sampling day from 31
May to 2 August and sixteen 1.5-h periods each sam-
pling day from 16 to 23 August. Seven Latin-square
sampling schemes were completed, three of 4 days
length and four of 1 day. The estimated outmigration
for each of the Latin-square schemes was: (1) from
2100. 31 May to 2100, 5 June (sampled every other
day). 43,700 age I and older; (2) 2100, 8 June to 2100,
15 June (sampled every other day), 3,210 age I and
older; (3) from 2100, 26 July to 2100, 2 August (sam-
pled every other day), 160,703 age 0; (4) 2100, 16 Au-
gust to 2100, 17 August, 151,240 age 0; (5) 2100, 18
August to 2100, 19 August, 50,075 age 0; (6) 2100, 20
August to 2100, 21 August, 13,120 age 0; and (7) 2100,
22 August to 2100, 23 August, 28,940 age 0. The num-
bers of each age of sockeye salmon migrating from
Coville Lake to Grosvenor Lake from 21 May to 15
September 1962 (based on the sampling with fyke nets)
were about 2,237,000 age 0 and 60,500 age 1
(Table 13).

1963. — In 1963 the outmigration of sockeye salmon
from Coville Lake was sampled from 20 June to 17
September. Fyke nets were fished in Coville River
from a cable as in 1962, but the location was about 15
m downstream where the river is 30 m wide and the
depth more uniform. Nets were fished on Latin-square
schedules 5 to 12 August and 13 to 17 September with
four sites and four 6-h fishing periods each day. The
estimated outmigration for each of the Latin-square
schemes was: (1) from 2100, 5 August to 2100, 12 Au-

Table 13. --Estimated numbers of age 0 and age
I sockeye salmon migrating from Coville Lake
to Grosvenor Lake (by 10-day periods) , May 29
to September 15, 1962, based on results of
fishing with 4-ft fyke nets in Coville River.

Table 14. - -Estimated numbers of age 0 and age
I sockeye salmon migrating from Coville Lake
to Grosvenor Lake (by 10-day periods), June 20
to September 17, 1963, based on results of
fishing with 4-ft fyke nets in Coville River.

Age 0

Age I

Period

fish

May 29-June 1

18,344

June 2-11

36,875

June 12-21

2,389

June 22-July 1

240

July 2-11

899

Julv 12-21

4,109

1,583

July 22-31

184,468

268

August 1-10

92,699

104

August 11-20

414,702

August 21-30

774,079

August 51- September 9

516,036

September 10-15

250,819

Total

2,236,912

60,502

Age 0

Age I

Period

fish

June 20-21

762

289

June 22-July 1

2,539

2,537

Julv 2-11

5,598

516

July 12-21

40,756

3,189

July 22-31

148,318

1,650

August 1-10

393,619

August 11-20

132,672

August 21-50

55,330

August 31 -September 9

69,688

September 10-17

70,448

Total

917,750

8,264

Several thousand age 0 fish were captured
in 1-m fyke nets fished intermittently along
shore. These fish are assumed to have origi-
nated from spawning in Coville River.

gust (sampling every other day), 120,100 age 0 and (2)
2100, 13 September to 2100. 17 September, 28,275 age
0. The estimates of the juvenile sockeye salmon migra-
ting by 10-day periods from 20 June to 17 September
are 918,000 age 0 and 8,300 age I (Table 14). The rela-
tively few age 0 fish that migrated before 12 July were
probably not interlake migrants, but were progeny of
females that spawned in Coville River.

1964. — The migration of juvenile sockeye salmon
from Coville Lake was sampled with the same
techniques and at the same cable site in 1964 as in 1963.
Sampling was done intermittently from 1 1 July to 8
September. The index net was fished on 31 days and
two Latin-square schedules were completed — one
from 31 July to 4 August and the other from 20 to 25
August. The estimated outmigration for each of the
Latin-square schemes was: (1) from 1800, 31 July to
1800, 4 August, 122,569 age 0 and (2) from 1800, 20
August to 1800, 25 August (22-23 August not fished).
715,719 age 0. The estimates of the juvenile sockeye
salmon that migrated in 10-day periods from 1 1 July to
7 September 1964 are about 3,036,000 age 0 and 3,900
age I (Table 15).

Migration from Grosvenor Lake to Hiuk Arm

The numbers of presmolt sockeye salmon migrating
from Grosvenor Lake to lliuk Arm, the next basin
downstream, was estimated from the results of fyke
netting in Grosvenor River. The nets were fished in-
termittently on 29 days between 30 May and 17 Sep-
tember 1962 and on 4 days between 10 August and 10
September 1963. On the basis of the seasonal variation
in the rate of catch of age 0 fish in Coville and Gros-
venor Rivers, I assumed that the summer interlake
migration of this age group began about 15 July. Some
age I fish were usually found in the fyke net catches in
Grosvenor River and I assumed that these fish had

Table 15. --Estimated numbers of age 0 and age
I sockeye salmon migrating from Coville Lake
to Grosvenor Lake (by 10-day periods) , July 11
to September 7, 1964, based on results of
fishing with 4-ft fyke nets in Coville River.

Age 0

Age I

Period

fish

fisli

July 11

302

July 12-21

2,723

1,883

July 22-31

189,595

1,351

August 1-10

183,921

292

August 11-20

1,288,903

116

August 21-30

1,139,396

218

August 31-September 7

251,155

Total

3,035,975

3,946

originated in Grosvenor Lake or had spent at least one
winter there because age I fish were virtually absent
from the Coville River summer migrations.

My estimate of the number of age 0 sockeye salmon
that migrated from Grosvenor Lake to Iliuk Arm from
15 July to 17 September 1962 is 3.9 million (Table 16).
This estimate is made by expanding the daily estimates
by a factor of 20. The factor of 20, though subjective,
is believed to be conservative and was selected after
considering the width of the channel at the fishing site
(about 76 m), other physical conditions (such as water
depth and current velocity), and the behavior of these
migrating fish in relation to the fyke nets at Grosvenor
and Coville Rivers.

Significance of the Summer
Outmigrations from Coville Lake

The significance of the summer outmigrations of age
0 sockeye salmon from Coville Lake can now be con-
sidered. The best estimates of the number of age 0
sockeye salmon in Coville Lake about 1 September
and estimates of the number that migrated from the
lake during the summer each year from 1961 to 1964
are shown in Table 17. The number that migrated in

Table 16. --Estimated numbers of age 0 sockeye
salmon migrating from Grosvenor Lake to Iliuk
Arm (by 10-day periods) , July 15 to September
17, 1962, based on fyke net catches in Grosve-
nor River.

Period

Age 0 fish

July 15-21
July 22-31
August 1-10
August 11-20
August 21-30
August 31-September 9
September 10-17
Total

860

67,400

126,180

842,960

941,280

499,500

1,590,940

3,870,900

1961, 1962, and 1964 greatly exceeded the number that
remained in the lake and in 1963, the number that mi-
grated was equal to the number that remained in the
lake. Furthermore, in none of the years did the migra-
tion appear to be over when the sampling was ended.
Although none of these data are precise, the summer
outmigration of age 0 fish from Coville Lake appears
to be significant to that lake.

The question of whether significant numbers of age
0 sockeye salmon remain in Coville Lake through the
winter to migrate as age I smolts cannot be answered
directly. Because ice frequently persists in Coville and
Grosvenor Lakes until early June, it is difficult to
reach Coville River and sample the spring migration.
In 1961 and 1962 the migration apparently started
while ice covered the lakes and was well underway
when sampling began because the rate of migration of
age I fish (assumed to be smolts) generally declined
from the first sampling. It is possible that the migration
of age 0 fish usually continues into the fall and only
relatively few fish remain to migrate as age I. Ruggles
(1966) reported such an overwinter shift in distribution
(seaward) of presmolts between basins of Owikeno
Lake, British Columbia.

Table 17. --Number of age 0 sockeye salmon in
Coville Lake at the end of summer (September
1) and number that migrated from the lake dur-
ing the summer, 1961-64.

Age 0 sockeye

Number

salmon1 in

migrating

Coville Lake

during

Year

on Sept. 1

summer

1961

3.8

5.5

1962

0.6

2.2

1963

0.9

1964

0.4

3.0

Product of average catch per standard
tow and number of standard tow volumes to a
depth of 10 m; there are about 61,000 such
standard tow volumes in Coville Lake,

Evidence on the immediate fate of age 0 sockeye
salmon that leave Coville Lake indicates that these
fish continue downsystem through Grosvenor Lake
and into Iliuk Arm the same summer. This evidence,
which is circumstantial and pertains to numbers and
size of the fish, comes from fyke netting in Coville and
Grosvenor Rivers and tow netting in Grosvenor Lake
and Iliuk Arm. The data indicate that age 0 sockeye
salmon migrating from Coville Lake during the sum-
mer continue downsystem into Iliuk Arm within a few
weeks.

The immediate fate of age 0 sockeye salmon that
migrated from Coville Lake to Grosvenor Lake can be
inferred from the number that enter and the number

that leave Grosvenor Lake and from the trends in
abundance of the populations in Grosvenor Lake and
lliuk Arm. Order of magnitude estimates of the
number of age 0 fish at these points — Coville River.
Grosvenor Lake, Grosvenor River, and lliuk Arm
— in July and August 1961-63 are summarized in Table
18. The estimate for Grosvenor River in 1963 is based
on the relation of the catches in August and September
of 1962 (Table 16) and 1963 and the estimated total
migration of 1962. There is no evidence that the
number of age 0 fish in Grosvenor Lake increased in
August, even in 1961 when the migration from Coville
Lake was largest. The number of age 0 fish in lliuk
Arm increased during the summer each year, how-
ever, and the increase was greatest in the year of mi-
gration of greatest numbers offish to Grosvenor Lake
from Coville Lake — 1961. Observations of even the
general magnitude of the migration out of Grosvenor
Lake into Grosvenor River are available only for 1962
and 1963 (fyke nets were fished in Grosvenor River 30
days in 1962 and 14 days in 1963). These observations
indicate that more age 0 fish left Grosvenor Lake in
1962 — the year when more entered from Coville Lake.
All indications are that most of the age 0 fish entering
Grosvenor Lake from Coville Lake in July and August
continue downsystem into lliuk Arm the same sum-
mer.

The sizes (length frequencies) of the age 0 sockeye
salmon that left Coville Lake and entered Grosvenor
Lake are similar to those for age 0 fish taken by tow
nets in Coville Lake and those leaving Grosvenor
Lake. All three differ, however, from the samples col-
lected with tow nets in Grosvenor Lake (see length
frequency graphs in later section of this paper). The
fish migrating from Coville Lake were either not pres-
ent in the parts of the lake sampled by tow nets in
Grosvenor Lake, or the number present in these areas
at the time of sampling (the "instantaneous" number)
was too small to be significant in the catches. The
latter is likely because visual observations, beach sein-

ing, and trap netting along the shores of Grosvenor
Lake all indicated very few age 0 sockeye salmon in
the littoral areas — the area not sampled by tow nets.

Diel Timing of Migrations

Although juvenile sockeye salmon usually migrate
downriver only during dusk or darkness (Hartman,
Heard, and Drucker, 1967), the interlake migrants did
not always follow this pattern. A restriction of down-
stream migration to the dark period of each day was
clearly the case for presmolts in Grosvenor River and
just as clearly not the case for similar fish in Coville
River. Results of sampling in Coville River in 1961 and
1962 and in Grosvenor River in 1962 show the diel
timing of this migration (Table 19). In Coville River no
consistent differences in intensity of movement
occurred — age 0 sockeye salmon migrated in great
numbers in both daylight and darkness. In Grosvenor
River, however, relatively few migrants were captured
during daylight, but large catches were made during
darkness.

Differences in the abundance and size of juveniles in
tow net catches near the outlets of Coville and Gros-
venor Lakes probably resulted from the differences in
their diel migratory behavior. Unusually large catches
of age 0 sockeye salmon were made with tow nets near
the outlet of Grosvenor Lake on 3 nights during the
period when large catches of migrants were made with
fyke nets in Grosvenor River. The rate of catch in tow
nets decreased progressively as fishing was done
farther from the outlet of the lake. The length fre-
quency distributions of fish from these large catches
were similar to those in samples of fish from Gros-
venor River and unlike those from Grosvenor Lake.
Although juvenile sockeye salmon were abundant in
the sampling area near the outlet end of Coville Lake,
tow netting in the immediate vicinity of the outlet did
not produce unusually large catches. It appears that
migrants accumulated at the outlet end of Coville

Table 18. --General magnitude of age 0 sockeye salmon in interlake migrations and of lake
populations in July and August 1961-63, Covile River-Iliuk Arm area.

Fish migrating

down Coville

River into

Grosvenor Lake

in July and

August

Fish in
Lake:

Fish migrating
Grosvenor from Grosvenor
on-- Lake in July

Fish

in lliuk Arm1
on--

Year

Aug. 1

Sept. 1 and August

Aug. 1

Sept. 1

5.0
1.5
0.9

1
2

1961
1962
1963

1 2

2 <2

Product of average catch per standard tow and number of standard tow volumes to a depth of
10 m. There are about 270,000 such standard tow volumes in lliuk Arm and 200,000 in Grosvenor
Lake.

Table 19. --Rate of catch in fyke nets and mean size of age 0 sockeye salmon migrating down
Coville and Grosvenor Rivers during dark and light periods1 between July and September 1961-62.

Period of day fished

Partly

dark and

Mean
fork

Dark

parti)

' light

Light

Mean number

length

Hours

of fish per

Hours

of fish per

Hours

fish per

Area and period

(mm)

fished

hour

fished

hour

fished

hour

Coville River2

July 27-31, 1961

49.1

1.0

27.0

7.0

1.7

5.0

240.0

Aug. 1-15, 1961

52.1

0.0

1.5

333.3

17.8

73.4

Aug. 16-31, 1961

56.5

0.0

22.0

6.6

7.7

103.9

Sept. 1-8, 1961

56.2

0.8

242.7

39.8

7.8

0.0

July 3-13, 1962

18.0

0.0

3.0

0.0

July 16-51, 1962

50.3

43.0

54.0

21.0

322.5

49.0

83.9

Aug. 1-16, 1962

56.2

24.0

60.6

13.0

19.5

45.0

37.9

Aug. 17-31, 1962

58.6

76.5

112.6

85.5

157.5

90.5

308.6

Sept. 15-15, 1962

63.1

25.0

54.6

59.0

37.0

27.0

274.6

Grosvenor River3

July 6-8, 1962

7.0

0.7

0.0

July 15-26, 1962

42.2

13.0

0.7

17.0

2.7

4.0

0.0

Aug. 9-12, 1962

57.2

6.0

88.7

29.0

65.4

22.0

0.0

Aug. 18-51, 1962

60.8

38.0

592.6

22.5

129.0

21.0

0.5

Sept. 1-9, 1962

62.3

19.5

271.6

35.8

61.8

28.5

453.6

Sept. 15-17, 1962

65.9

22.8

102.2

0.0

—

12.5

11.2

1 Average sunrise and sunset times were determined for each semimonthly period from pyrheli-
ograph records from Coville Lake outlet. Dark = sunset to sunrise when pyrheliograph reading was
0. Light = sunrise to sunset.

2In 1961, a 1-m-square fyke net was fished from steel posts driven into stream bottom. In
1962, a 1.2-m- square fyke net was fished from a cable strung across the stream; the wings were
spread to 1.3m.

3A 1.2-m- square fyke net was fished from a cable strung across the stream; the wings were
spread to 1.8 m.

40ne fishing period of 4 hours duration produced an exceptional catch of 1,500 juvenile
sockeye salmon.

Lake, but were not concentrated near the river as at
Grosvenor Lake. The accumulation of fish near the
outlet of Grosvenor Lake probably resulted from their
reluctance to migrate down the river during daylight.
Juvenile sockeye salmon have been studied in sev-
eral multibasin systems similar to the Naknek system
and oceanward interlake migrations of significant
numbers of age 0 sockeye salmon during the summer
are apparently rare. The several basins of the Babine
River system have markedly dissimilar densities of fry
early in the summer as the result of the unequal dis-
tribution of spawning adults (much as in the Naknek
system). Unlike the fry of the Naknek system, the fry
of the Babine system do not disperse over the lakes
during summer (Johnson. 1958). (The greatest number
of spawners per unit lake area is in the most upsystem
lake of the Naknek system, but in the lower end of the
Babine system.) In the Wood River system there is a
minor migration of fry from small lakes to a larger lake
(Burgner, 1962). In the Chignik River system there is
little downsystem movement of age 0 fry between
lakes, but here (similar to the Babine system) the

downsystem lake usually has the greater density of
spawners. There is a migration of fry from the lower
lake (Chignik Lake) to the lagoon-like estuary (Burg-
ner et al., 1969). A recent study of growth patterns on
scales of adult sockeye salmon from the Chignik sys-
tem indicates that age 0 fish did migrate to a downsys-
tem lake in 19569 (Narver, 1968).

A migration unusual because of its direction has
been reported for Owikeno Lake, British Columbia.
Ruggles (1966) found a movement of age 0 sockeye
salmon from one lake basin to another away from the
direction of the outlet to the ocean. The time of the
migration and relative density offish in the two basins
before and after the migration were not reported. Dur-
ing the winter another migration occurred, but this
time it was oceanward.

"At 1700 on 30 July 1962 the speed of movement was estimated
for 10 schools of age 0 sockeye salmon moving downstream at the
outlet of Coville Lake. The current speed, gaged by observing a
floating wood chip, was about 0.2 feet per second (fps). The speed of
the schools averaged about 1.9 fps, indicating a swimming speed
downstream of about 1.7 fps.

Behavior of Schools of Age 0 Fish
at Outlet of Coville Lake

Although the behavior of the juvenile sockeye
salmon in the interlake migrations in the Naknek sys-
tem were not studied in detail, incidental observations
of the fish involved are presented here because the
phenomenon of large-scale interlake migrations of
these presmolt fish is unique.

Interlake migrants first appeared in the shoal waters
(less than 3 m) at the outlet of Coville Lake (i.e., the
origin of Coville River). Here the basin of the lake
becomes so narrow and shallow that the current is
readily visible. In mid-June 1962, groups of several
hundred age 0 fish were frequently seen moving down-
stream about the same speed as the current and appar-
ently feeding at or near the surface — the fish were
breaking the surface so frequently that the movement
of the groups could be followed by an observer on
shore. These groups were not concentrated along the
shore or over the deepest water, but were seen at one
time or another over the entire outlet area. Individual
fish were most often facing downstream. When the
water depth decreased to about 1 m and the current
velocity had noticeably increased (and possibly when
the fry first made visual contact with the bottom — i.e.,
they first "realized" they were moving downstream),
the fry abruptly changed their orientation.

The visual cue as to direction or perhaps simply the
existence of movement seemed to halt the downstream
migration. All the fish of a group would suddenly turn,
face upstream, and move laterally across the current
until they were in water about 15 to 20 cm deep. They
then moved upstream in a narrow band until they
reached slower water and disappeared into deeper
water — that is, they appeared to return to the lake.
From mid-June to mid-July, during daylight, schools
of age 0 sockeye salmon were frequently seen feeding
in shallow (1 to 2 m) water along the lake shores and
islands over most of Coville Lake.

About mid-July the behavior of the age 0 sockeye
salmon at the outlet of Coville Lake had noticeably
changed and the schools now appeared to be actively
migrating. They were still close to the surface, but
moved faster than the current (see footnote 10) and no
longer changed orientation when the water became
shoaled to about 1 m deep or moved toward shore
when the currently velocity increased. The orientation
of individuals and ultimately the entire school was
suddenly reversed (Hartman, Heard, and Drucker,
1967) as the school passed over the edge of the shoal
water into deeper water of the stream proper. This
orientation was soon reversed and the fish again
moved actively downstream into Grosvenor Lake and
deeper water.

On one occasion after the summer outmigration of
age 0 sockeye salmon was in progress, a reversal of the
migration was noted. At 0900 on 19 August 1962 fish
were seen moving upstream near the outlet of Coville
Lake. The characteristics of the movement, i.e..

speed, school compactness, size, etc., were the same
as for the downstream movement. In the afternoon of
the same day the migration had resumed its normal
(for that time of the summer) direction. This was the
only reversal of the direction of migration observed
here, but reversed migration has been commonly ob-
served in smolts in the Babine system (Groot, 1965).

EARLY REARING AREAS OF

SOCKEYE SALMON FRY FROM

GROSVENOR RIVER AND HARDSCRABBLE

CREEK

Some stocks of sockeye salmon spawn in rivers that
connect lakes or connect a lake to the ocean; their
progeny may migrate either upstream (Andrew and
Geen, 1960) or downstream (most commonly) to reach
freshwater pelagic rearing areas. A choice of migration
direction is possible in three major connecting rivers in
the Naknek system — Brooks, Naknek. and Gros-
venor (Fig. 1). Fry from Brooks River move down-
stream into South Bay (Merrell, 1964); we assume fry
from Naknek River move upstream into Naknek Lake
rather than going directly to the ocean because adults
of freshwater-age 0 are rare in the escapement. The
immediate destination of fry migrating from Gros-
venor River was unknown until 1962.

In the spring of 1962 I studied the fry originating in
Grosvenor River and Hardscrabble Creek to deter-
mine the basin to which they first migrated.
Hardscrabble Creek was studied because it is close to
Grosvenor River and the work in the two streams
could be done from a single camp. Moreover, I felt
that information on the timing and other characteris-
tics of the outmigration from Hardscrabble Creek
might corroborate the work in Grosvenor River. Ul-
timately the two streams were found to be closely re-
lated. This work was exploratory and the results are
qualitative. Descriptions of the upstream migration of
fry in other areas indicated the upstream migration is
obvious — for example, Johnson ( 1956) described these
fry as "... a massed living band moving
upstream. . ."; McCart (1967) stated ". . . (upstream)
migrants moved in tightly knit schools at the surface,
close to shore, often in water only a few centimeters
deep."

In Grosvenor River small fyke nets were fished
along each side (east and west shores) of the river
(usually with one wing extended to shore) near Gros-
venor Lake where the river first becomes less than 50
m wide. Initially, nets were fished to sample both the
upstream and downstream migrations, but most sam-
pling was done to catch downstream migrants. In
Hardscrabble Creek a fyke net was fished in fast water
about 0.6 m deep on the first gravelly riffle above
Grosvenor Lake (about 200 m from the lake at low
lake water level). One fyke net set was made in the
Savonoski River to learn if fry were produced in that
system above its confluence with Grosvenor River.

Visual observations were made during daylight and
darkness while walking along Grosvenor Lake from
Hardscrabble Creek to Grosvenor River and along
Grosvenor River on the shore or in shallow water.
Hand-held lights were used at night.

Most of the sockeye salmon fry captured in the fyke
nets (Table 20) or seen migrating were moving down-
stream on the east shore of Grosvenor River at night.
A few fry were seen moving upstream near shore from
Grosvenor River to Grosvenor Lake from May to
June; the only other indication of an upstream move-
ment from Grosvenor River was the capture of a few
fry in fyke nets open downstream (Table 20). Fyke
nets fished in shallow water near the lower end of
Grosvenor River on 11, 16, 17, and 19 May indicated
that recently emerged fry were moving downstream,
but the origin of these fry is uncertain. Presumably
they were a mixture of fry originating in Grosvenor
River and Hardscrabble Creek.

Migrating fry were also sampled intermittently with
fyke nets in Hardscrabble Creek between 11 May and
25 June. During this time the water level and velocity
changed so that the rate of catch of fry in Hardscrab-
ble Creek is the result of straining greatly different
proportions of the total flow and, presumably, of the
nightly migration. Therefore, only one general conclu-
sion can be made about the migration — some sockeye
salmon fry were moving downstream in Hardscrabble
Creek between 11 May and 25 June.

Visual observations along the edge of Hardscrabble
Creek at night below the fyke net collecting site
showed that the fry usually swam downstream. Some
fry stayed in shallow water in the delta of the stream
and could be seen along the lake shore and on into
Grosvenor River. It appeared that at least some fry

from Hardscrabble Creek never entered the pelagic
area of Grosvenor Lake but stayed in water from
Hardscrabble Creek well downstream in Grosvenor
River. Hardscrabble Creek water was not mixed with
water from Grosvenor Lake until about 180 m below
the lake. The two waters were initially quite
distinct — the water of Hardscrabble Creek was murky
from erosion products and glacial melt and the water
of Grosvenor Lake was clear. Further evidence that at
least some fry moved directly from Hardscrabble
Creek to Grosvenor River was found by comparing
fry from Hardscrabble Creek and from the head of
Grosvenor River. Samples collected on the same or
adjacent nights in the two areas were nearly identical
in regard to length frequencies and the proportion of
fry containing visible yolk. If the fry leaving Gros-
venor Lake had been in the lake very long, they would
have absorbed more yolk and increased in length.

I concluded that Uiuk Arm is the basin of first resi-
dence of practically all sockeye salmon fry originating
in Grosvenor River and of an unknown portion of
those originating in Hardscrabble Creek. Iliuk Arm
also receives some fry from streams tributary to the
Savonoski River above Grosvenor River.

SIZE, LENGTH FREQUENCY,
AND GROWTH

Intimately associated with the abundance of ani-
mals are the growth and size of individuals. In the
present study, the sizes of individual fish in the
catches were measured so that the effects of biological
and physical factors on size could be determined and
groups of fish could be identified. Although both
length and weight were measured, only the length

Table 20. --Numbers of recently emerged sockeye salmon fry captured in fyke nets set on the east
and west shores of Grosvenor River near Grosvenor Lake in May and June 1962 to determine whether
fry were migrating upstream or downstream.

Date

Downstream migrants

East shore West shore

Hours
fished

Fry caught
per hour

Hours
fished

Fry caught
per hour

Upstream migrants

East shore West shore

Hours
fished

Fry caught
per hour

Hours
fished

Fry caught
per hour

May 17
May 18
May 19
May 20
May 21
May 22
May 23
May 24
May 25
June 10
June 11
June 16
June 17
June 19
June 25

1.0

0.5
2.5
4.8
1.7
4.2
7.5
3.5

21.0
7.5
1.8
2.8

11.4

12.1

51.0

44.0

190.0

1-05.0

95.3

94.2

12.7

194.5

36.8

40.1

70.6

70.7

1.6

1.7
1.0

1.7
4.2

8.2
20.0

2.2
158.3

5.5
0.5
2.0
2.8

0.4
0.0
2.0
1.1

measurements proved to be useful in final analysis.

The most extensive data on size of juvenile sockeye
salmon came from collections made with tow nets.
These data, in the form of average lengths and length
frequencies, have been used to relate changes in aver-
age size with time (apparent growth) to abundance of
sockeye salmon and other species and in some in-
stances to investigate the effects of migrations.

Average lengths of juvenile sockeye salmon in
catches were used to calculate "growth" curve equa-
tions which describe the average size by age class each
day. After trying several mathematical models and
visually examining the fit of the curves to the actual
data, I selected a second-degree polynomial
(Snedecor, 1956), in which length is related to time in
days since 30 May (i.e. 1 June = day 1 ; 1 July = day 3 1 ;
and 1 September = day 93). The equations describing
the average length have been used to: (1) calculate the
average size on other than dates of sampling by extrapo-
lation or interpolation; (2) make estimates of size from
combined data for sampling areas within a basin; and (3 )
plot graphs (apparent-growth curves) describing the
changes in length during a season.

Juvenile Sockeye Salmon

Curves depicting the average lengths of juvenile
sockeye salmon in the summers from 1961 to 1963 for
all basins and 1961 to 1964 for Coville Lake are pre-

sented in this section as each lake is discussed. Be-
cause of known differences between the average size
of migrating and nonmigrating fish, probable size-
related differences in mortality, and known variations
in time of recruitment of fry from the spawning
grounds, the curves represent only "apparent
growth."

A comparison of the average size of the fish from
different areas supplies part of the knowledge needed
to understand differences in apparent growth, but for a
more complete understanding knowledge of the length
frequency composition of the population is also
needed. Graphs of the percent frequency of juvenile
sockeye salmon by 3-mm size groups by time periods
have been prepared for 1961 to 1964. The length fre-
quencies offish sampled will be discussed and related
to their average lengths for each lake.

The mean fork length of age 0 and age I sockeye
salmon on 20 August and 1 September 1961-64 by
sampling area and lake are summarized in Table 21 for
each basin. The sizes used are those estimated from
the calculated growth curves rather than the empirical
data even when collections were made on 20 August or
1 September. The dates 20 August and 1 September
were selected for comparison for different reasons — 20
August is late enough in the season to indicate growth
conditions for the summer and early enough to avoid
most sampling problems caused by the early fall
storms; 1 September is the date used in much of the

Table 21. --Mean fork lengths of age 0 and age I sockeye salmon in each lake of the Naknek River
system and in Coville and Grosvenor Rivers on August 20 and September 1, 1961-64. (Weighted by
abundance and average size in each sampling area.)

Mean fork length (mm) on--

Age of fish and

August 20

Sept ember 1

sampling area

1961

1962 1963

1964 1951 1962 1963

1964

Age 0

Coville Lake
Coville River
Grosvenor Lake
Grosvenor River
Iliuk Arm
South Bay
West End
North Arm
Northwest Basin
Brooks Lake

Age I

Coville Lake
Coville River
Grosvenor Lake
Grosvenor River
Iliuk Arm
South Bay
West End
North Arm
Northwest Basin
Brooks Lake

51.1

55.8

56.3

56.9

54.2

59.6

57.8

58.2

58.0

58.6

59.2

60.6

62.7

61.1

46.5

51.0

45.8

46.3

51.7

51.0

47.3

60.1

62.0

62.8

64.9

46.9

51.0

53.2

57.4

61.3

58.3

47.4

42.8

54.8

50.7

46.8

63.4

55.3

52.3

59.8

62.9

56.2

64.7

69.0

60.8

57.7

56.0

53.7

45.7

55.3

49.7

51.3

60.2

53.7

84.3
84.5

83.4
83.8

79.9
81.9

78.4

91.0

86.0

77.9

86.5

85.0

84.9

85.2

89.3

89.4

94.0

1962

V^T

/£

• C- 1

» C-2

» COVILLE RIVER

III!

1963

C - 1
C -2U
C - 2M
C -2L

COVILLE RIVER

I , ■

— 1 1 1 1

0 30 40 50 60 70 80 90 100 0 30 40 50 60 70 80 90 100

DAYS SINCE MAY 31

Figure 11. — Curves of apparent growth of age 0 sockeye salmon captured in
tow nets in sampling units of Coville Lake and in fyke nets in Coville River
1961-64.

' I ' I ' | i I i 1 i M | ' I ' I ' 1

C-l, JULY 11-21, N=667
C-2, JULY II -21,14=470
COVILLE RIVER, JULY IB,N=I55

1 I ' 1 ' I I I ' I ' 1 ' I ' I ' I ' 1 ' I ' I ' I

* C-i, AUG I, N= i9T

> C2U.M.8 L JULY 31, AuC I, N = 920

' COVILLE RIVER, JULY 30,N = 24|

'IMMHTMI'I'

30 -

; 25 ~

z 10

• C-l. AUG 4-i3,N = i59
o C-2, AUG 4-13, N = 744

I i | ■ I M ' I ' I ' I ' 1 ■ I ' I ■ ITT

• C-l, NO SAMPLE

° C-2 , AUG 5,N=63

* COVILLE RIVER. AUG. 5,N:40

f I ' T ' I i I '

• C-l, NO SAMPLE

° C-2U.M, 6L.AUG IS.N = 4I3

• COVILLE RIVER. AUG I5.N=2I6

• C-l, AUG. 24.N=26
o C-2, AUG 2B,N=66

» COVILLE RIVER
AUG 29-30, N = I70

I ' I I I ' I I I I

• C-l, NO SAMPLE

» C-2U.M.8L.AUG 3i-SEPT,i.N = IO

* COVILLE RIVER , SEPT I. N = 2 6 2

I ' I ■ I ' I U ■ I ' M I ■ | I

30 42 54 66 78 90 102 30 42 54 66 78 90 102

Figure 12. — Length frequency distributions of juvenile sockeye salmon (age 0 and age I combined) captured
in tow nets in sampling units of Coville Lake and in fyke nets in Coville River for several time periods between
July and September 1961-63. (See Figs. 1 and 2 for designations of sampling units.)

existing literature on size of juvenile sockeye salmon.
When sampling was not done on or after 1 September
or the apparent growth was negative or otherwise
anomalous, the length on 1 September was estimated
as follows: for age 0 fish, the increase in length in
Coville Lake and for age I fish, the increase in length
in Iliuk Arm between the last date of sampling in the
lake in question and 1 September was added to the
calculated size on the last date of acceptable sampling
in the lake in question.

Coville Lake and Coville River. — Curves depicting
the apparent growth of age 0 sockeye salmon collected
in tow nets in Coville Lake and in fyke nets in Coville
River are presented by area for 1961-64 in Figure 11.
In general, average lengths increased rapidly from
early July to mid-August (days 3 1 to 80) and somewhat
slower thereafter. The decrease in rate is most appar-
ent in 1963 and 1964 when more areas were sampled at
shorter intervals. The average size of the emigrants in
Coville River was clearly greater than that of the
juveniles taken in tow nets in the area adjacent to the

river— C-2L in 1963 and C-5 in 1964. There was little
difference among the other four areas (C-l to C-4) in
1964.

The length frequency distributions of the juvenile
sockeye salmon from Coville Lake and Coville River
in 1961-64 are presented in Figures 12 and 13. The
frequencies are generally unimodal and the observed
differences in average length (Table 21) are due to the
greater abundance of larger sizes rather than to differ-
ences in ranges in lengths in the samples.

The greater average lengths of age 0 sockeye salmon
in the lakes of the Naknek system on 1 September
1962 (Table 21) than in the other years was true for fish
from Coville Lake, but not for those from Coville
River. No explanation is offered for the lack of a larger
average size for fish in Coville River.

Age I and older fish were rare in catches from tow
nets in Coville Lake and fyke nets in Coville River
during the summer. These older fish appear in the
length frequencies in sufficient numbers to cause a
bimodal curve in the samples from tow net catches
only in early July 1964 (Fig. 13).

I ' I ' I ' I ' I ' I ' I M ' I i I i I i I '

436JH • C-l T04, JULY 16-17,
N=l,256
° C-5. JULY 16 , N=679

I i I ' I U i I ' I ' I ' I ' I ' I I I i I I

• C-l T04, AUG 2-3

N=737
o C-5. AUG. 4

N = 62l
. COVILLE RIVER

AUG 4

N=277

J \ • C-IT04, AUG 16

J 1 N = 1,076

/N 1 o C-5, AUG 15-16

V \ N'393

k I \ \ » COVILLE RIVER
l\ \\ AUG. 18

A \ """

i J

'W J

o a a

X A »

0 •

1 ' 1 ' 1 ' 1 '

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

C-l TO 4, SEPT 2-5
N = I26

C-5, SEPT. 2
N = 27

. COVILLE RIVER
SEPT 1-5
N = 27l

1 I ' I I I I |

I ' I ' I ' I '

' I ' I l 1 I I I I ' I ' I U I I I I ' I I I I
30 42 54 66 78 90 30

1964

FORK LENGTH (MM )

I I ' I M ' I ' | I I I I I I I

42 54 66 78 90

1964

Figure 13. — Length frequency distributions of juvenile sockeye salmon (age 0 and age I
combined) captured in tow nets in sampling units of Coville Lake and in fyke nets in
Coville River for several time periods between July and September 1964. (See Figs. 1 and
2 for designations of sampling units.)

Grosvenor Lake and Grosvenor River. — The appar-
ent growth of age 0 sockeye salmon each summer in
1961-63 in Grosvenor Lake and Grosvenor River is
shown in Figure 14. Grosvenor Lake was divided into
four areas for tow net sampling, but only in 1961 were
enough samples obtained to describe the growth for
each area. For 1962 and 1963 data for the four areas
were combined to calculate a single growth curve. Col-
lections were made in Grosvenor River only in 1962
and 1963.

The apparent growth of juvenile sockeye salmon
from Grosvenor Lake is unique in two regards — the
average length offish in particular sampling areas fre-
quently decreased during the summer, and the fish
here were generally the smallest in the system on any
date. The size of the outmigrating fish captured in fyke
nets in Grosvenor River increased during the summer
and these fish were generally the largest in the system
on any date.

G-l
G-2
G-3
G-4

All units
combined
grosvenor river

1962

■<

. 1

1 I

1963

+ +

O 30 40 50 60 70 80 90 100
DAYS SINCE MAT 31

Figure 14. — Curves of apparent
growth of age 0 sockeye salmon cap-
tured in tow nets in sampling units of
Grosvenor Lake and in fyke nets in
Grosvenor River 1961-63.

The decrease in average length of age 0 sockeye
salmon in Grosvenor Lake in August was at least
partly due to a late recruitment of fry that had recently
emerged from the spawning gravels. This late recruit-
ment appeared each year from 1961 to 1963 and caused
the marked bimodality of length frequency curves
— these late recruits are represented in the peak on the
left in the 30- to 45-mm size range in Figure 15. The
spawning grounds and circumstances that produce
these fry in Grosvenor Lake are unknown, but a simi-
lar late recruitment has been observed for sockeye
salmon in Karluk Lake where spawning occurs over a
period of 4 to 5 mo (Burgner et al., 1969).

Age I and older sockeye salmon were rarely taken in
tow nets in Grosvenor Lake and were relatively scarce
in fyke nets in Grosvenor River.

Iliuk Arm. — Because there were no consistent dif-
ferences in size of age 0 or of age I fish among the
three sampling units of Iliuk Arm in 1961-63 (for 1962,
growth curves for each unit are shown for compari-
son), the data from all the units were combined in
calculating the growth curves (Fig. 16). The apparent
growth of age 0 fish showed little or no evidence of
slowing by 1 September and the average size of the age
0 fish in Iliuk Arm (Table 21) was generally inter-
mediate among the lakes of the system. This good ap-
parent growth was not expected in Iliuk Arm because
glacial flour makes the water quite opaque which
would result in little light penetration and thus low
photosynthetic activity. Both of these apparent
anomalies in growth (no slowing by September and
good apparent growth) are probably caused by the re-
cruitment of the larger fish from the upsystem areas
during the summer.

The calculated growth curves for age I sockeye
salmon in Iliuk Arm for 1961-63 in the three units
combined (Fig. 17) resemble those of age 0 fish in that
they do not show a decrease during the summer. There
was no trend in length of age I fish from one end of the
basin to the other (N-15, N-14, N-13).

The length frequency diagrams for samples of
juvenile sockeye salmon from Iliuk Arm for 1961-63
also indicate a general uniformity among the three
sampling units (Fig. 18). The length frequencies are
generally bimodal, reflecting the presence of the two
age classes — age 0 and age I.

South Bay. — The seasonal changes in apparent
growth and length frequencies of young sockeye salm-
on from South Bay are similar to those from Iliuk
Arm. The data were too few to permit analysis of
growth by sampling unit, but do permit considerations
of apparent growth for the entire basin (all units com-
bined). As in Iliuk Arm the apparent growth of age 0
fish had slowed little if at all by 1 September (Fig. 19).

The apparent growth curves for age I sockeye
salmon in South Bay for all units combined for 1961-63
(Fig. 20) do not have the same shape as those for Iliuk

• G-192. JULY 27, N = 43
<• G3B>4, NO SAMPLE

I ' I ' I ' T r i > |-i

• G. 182, AUG 6-IO.N^TO
[ «G-3ft4. AUG 10.N=I87
* GROSVENOR RIVER
AUG 9-U.N=338

G-lfl2,NO SAMPLE
G-364, AUG 27. N = 9

• G-2 .JULY 24, N = 29

• G-4.JULY Z9,N=7I8

* GROSVENOR RIVER
AUG I0,N=239

G-2 SEPT 2,N= 75
1 C-384, NO SAMPLE
► GROSVENOR RIVER

AUG 19 - SEPT 9, N; 485

30 42 54 66 78 90 MOO

Figure 15. — Length frequency distributions of juvenile sockeye salmon (age 0 and age I combined) captured
in tow nets in sampling units of Grosvenor Lake and in fyke nets in Grosvenor River for several time periods
between July and September 1961-63.

Figure 16. — Curves of apparent growth of age 0 sockeye salmon captured in tow nets in
sampling units of Iliuk Arm, 1961-63.

1961

1962

o N-15 /
° N-14 /
• N-13 /•

1963

0 30 40 50 60 70 80 90 0 30 40 50 60 70 60 90 0 30 40 50 60 70 80 90

DAYS SINCE MAY 31

Figure 17. — Curves of apparent growth of age I sockeye salmon captured in tow nets in
sampling units of Iliuk Arm, 1961-63.

OJ 10

ro 01

OJ u> —

o> ii "

CJ z z

OJ OJ

(\J oj oj

>-•>->■
— ' i >

=> => z>

■5-3-3

in <r io

hi i

• o

■

to io £>

2tS

•

z z z

1 <J- -

OJ OJ OJ

55; 5i

-D -1 "5

in <r ro

z z z

•

• 0

•

• 0 X

1 1 1

1 1

•

o.l

•X

CM & fO

^■ro-

il M f

Z Z Z

C\J — —

555

c- — ^

Z) => =>

S. I

"> "3 "3

/ 1

- -

.c i •

in ^ to

— — —

~^v

z z z

• OX

o— =r

=0*

00 * ^_

o^V

•o 7T^»»

J> N

1 1 1

I I In. I

a
<

1 1 1 I

J L

1 1 1 1

<
w

z
01

z z

o X

J L

o
■ o

l£H£)

tDCJ.

—

<3"

»f>

-j:

«•!

15, AUG. 24
■98

14, AUG. 24

137
-13, AUG. 28

200

*<^

z z z z z z

"""""■* 1

• ox

B^^

■— ^*

! 1 !

I...I

- o
-o

lOCTi

O »n
oj —

=
-f

■g

u
a

S 3

■o —

10 U)

w c

UJ01

c =»

111

J3 ~

a s

IT
O

_ 3

„ <

ADN3n03<d3 lN3D«3d

30 10 50 60 70 80 90
DAYS SINCE MAY 31

30 40 50 60 70 60 90

Figure 19. — Curves of apparent growth of age 0 sockeye salmon captured in tow nets in
sampling units of South Bay 1961-63.

Arm — a decrease in growth rate and average length
appears in the South Bay data. The apparent negative
growth of age I fish in South Bay may be due to the
combining of data to produce a single curve. On the
other hand, it may be due to the loss of larger fish to
the summer migration of smolts in the Naknek River.
If this summer migration of smolts did not involve fish
from Iliuk Arm, the observed difference in growth
curves of South Bay and Iliuk Arm would result.

The length frequency diagrams of juvenile sockeye
salmon from South Bay (Fig. 21) resemble those from
Iliuk Arm. In general, there is a bimodality indicating
two age groups (age 0 and age I). In those periods
when all three sampling units of South Bay were sam-
pled, the curve for age I fish from unit 6 (the most
downsystem area) was to the right of those of the other
two units — larger age I fish were relatively more
abundant in the downsystem portion of South Bay.
Age I and older sockeye salmon in South Bay were
similar in length to those in Iliuk Arm (Table 21).

West End. — Apparent growth curves for sockeye
salmon in the West End are available only for age 0
fish and only in 1962 and 1963 (Fig. 22). The growth of
age 0 fish from West End differed from that in Iliuk
Arm and South Bay in that the growth in West End

fish tended to decrease during the summer. The aver-
age lengths of age 0 fish were, however, usually grea-
ter in samples from West End than in those from other
lakes of the system on the same date (Table 21). Al-
though the samples were small, within West End the
average lengths were generally greatest in N-4, the
area adjacent to South Bay. The average lengths (in
millimeters) of age 0 sockeye salmon in samples from
N-4, N-2, and N-l were as follows:

Date

N-4

N-2

N-l

66.0

64.2

62.1

62.3

58.6

57.8

('")

61.7

56.6

20 August 1962
31 August 1963
3 September 1964

Age I and older sockeye salmon were too scarce in
tow net catches from West End to permit construction
of growth curves. The average length of this group
was greater here than in Iliuk Arm or South Bay.

The length frequency graphs for samples of juvenile
sockeye salmon from West End (Fig. 23) are unimodal
because of relatively few age I fish in the tow net
catches.

'"No sample taken.

0 30 40 50 60 70

90 0 30 40 50 60 70 60 90 0 30 40 50 60 70 80 90
DAYS SINCE MAY 31

Figure 20. — Curves of apparent growth of age I sockeye salmon captured in tow nets in
sampling units of South Bay 1961-63.

oj -

" o

Z K>

r- oo
eg oj

3 Z>
-5 "5

01 UP

z z

•

O X

"~ 1

•

1 >

•

^^^

•

00 •

tfl /

" UJ UI t

Z _|_l i

- 0. 0- /

"25 /

O WW \

3 oo »

< Z Z

•

— - - •

— en to

•

z z z

• ox
1

10 10

*» o» —

X" —

1 J^--"

^>x

•

CO z z

K-"*-s^ J7*"^

2 r^r"-"

OJ OJ

^ en en

K^»

JULY
JULY
JULY

• X

"^ 1 1

_ en to

• o

z z z

• ox

• 0 X

X
X

— —«x

1 1 1

i i

O
1 1 i I 1

• ox

T-

o- ^

UJ "" ^^>>

_, .. uj 7

o_ z -i -»

SAM
Y 15,
SAMP

5 ^° 1

z ^ z *

- - _

— <J) ID

Z Z Z

1 1 1 1 1 I...I

•

z'<

en s
~ ro
OJ UJ —

•

1 1 , AUG. II, N =
9, NO SAMPL
6, AUG. 10, N =

• x

X
X

z z z

• OX

1 1

i i

1 ,, ,1

•
•

hul

O <f> O tf> O u~>
ro oj oj — —

ADN3n03cH lN33d3d

uj 5 5

N < <t

D O O

< z z

— en io

z z z

Z _l

UJ uj

-1 _l

■ a. a

O Z 2

2 o o

< z z

— en to

Z. 2. Z

I IM

- en

U5y-j

to en

2
5

10 to

to en

—

. o
o

o
en

1962 y\

• N-4,2 ,a 1

1 1 1 1

1963 yf

Figure 22. — Curves of apparent growth of age 0 sockeye salmon captured
in tow nets in sampling units (combined) of West End 1962-63.

North Arm. — Although North Arm is the largest
basin in the Naknek system (about one-quarter of the
system's surface area), apparent growth curves and
length frequency curves could not be constructed be-
cause so few fish were captured in tow nets. The only
sample captured near 1 September that had more than
20 age 0 fish was obtained in 1963; the average length
of these age 0 fish (Table 21) was close to the average
for the system. Only one large sample of age 1 sockeye
salmon was taken in tow nets in North Arm and this
happened during a daytime test tow on 13 August
1963. The average length of 225 fish in this tow was 93
mm. The average length of 18 fish taken during regular
tow netting in North Arm on 1 September 1963 was 94
mm. The average lengths of the age I fish in these two
samples were several millimeters longer than the aver-
age lengths in similar samples elsewhere in the system.

Northwest Basin. — Northwest Basin is small and
relatively insignificant in the production of sockeye
salmon in the system and consequently was not sam-
pled as intensively as the other basins. Too few data
were obtained to permit construction of growth or
length frequency curves. Average size data are avail-
able for two dates — 20 August 1961 and 1 September
1963 (Table 21). The size of age 0 fish was about aver-
age for the system. Age I and older sockeye salmon
occurred only occasionally in tow net catches in
Northwest Basin and the general size of these fish was
similar to those from Iliuk Arm and South Bay.

Brooks Lake. — Juvenile sockeye salmon were gen-
erally as scarce in tow net catches in Brooks Lake as
in North Arm, but because more tow netting was done
in Brooks Lake data were sufficient to permit con-
struction of growth curves. The calculated curves de-
scribing the apparent growth of age 0 sockeye salmon
in Brooks Lake (Fig. 24) were based on samples of one
or more fish. Although the minimum sample size was
small, all the points fall close to the calculated curves.
These curves show the typical (for the Naknek sys-
tem) declining rate of growth in late August. The aver-
age lengths of age 0 sockeye salmon on 20 August and

1 September were generally about average for the sys-
tem (Table 21). Age I and older fish were seldom cap-
tured in Brooks Lake, but in general they were about
the same size as comparable fish in Iliuk Arm and
South Bay.

Causes of Differences in Size of Juvenile
Sockeye Salmon on 1 September

Differences in the size of juvenile sockeye salmon
within a year between areas and within areas between
years have commonly been reported for other sys-
tems. These differences may be due to one or more
factors, of which I will consider the following for the
Naknek system: (1) real differences in rates of growth,
(2) differences in time of recruitment of fry and result-
ing differences in number of growing days by a given
date, (3) differences in rates of dispersion of large and
small or fast- and slow-growing fish, and (4) differ-
ences in size of fry at time of emergence.

Real differences in rates of growth. — Differences in
the rates of growth of juvenile sockeye salmon within a
system are most likely due to differences in the avail-
ability of food and in water temperatures. A reduction
in the average size of juvenile sockeye salmon has
often been directly or indirectly attributed to large
numbers of feeders, both sockeye salmon and other
species such as sticklebacks. Some examples in sys-
tems of western Alaska are the Wood system (Burg-
ner. 1964); the Kvichak system;11 and the Chignik sys-
tem (Narver and Dahlberg, 1964). Examples in other
areas are: British Columbia, Babine Lake of the
Skeena system (Johnson, 1958) and Cultus Lake of the
Fraser system (Foerster, 1944), and the east coast of
Kamchatka Peninsula, USSR, Lake Dalnee (Krogius,
1961).

For the seven largest lakes of the Naknek system
(Northwest Basin is excluded because of too few sam-
ples), the mean surface water temperatures in the
month of August, the mean number of age 0 and age I
sockeye salmon per tow, and the mean fork lengths
of the age 0 fish for the years 1961-63 are shown in
Table 22; the mean number per tow of the three
species offish most commonly taken in tow nets with
the juvenile sockeye salmon (pond smelt, threespine
sticklebacks, and ninespine sticklebacks) are also
shown in the table. Some of the differences in the size
of the age 0 sockeye salmon are probably due to real
differences in growth rate. The largest age 0 fish gen-
erally occurred in Coville Lake and West End (Table
22). These two basins also had the greatest average
combined catches of sockeye salmon juveniles and as-
sociated species in tow nets and the highest surface

"Orra E. kerns. 1966. Abundance and size of juvenile sockeye
salmon and major competitor species in Iliamna Lake and Lake
Clark. 1964 and 1965. Univ. Wash., Fish. Res. Inst. Circ. 66-15.
35 p.

J L

I I I I

ZL
0>

■3

J L

U3
Z

"cn

in 10
"(0 01

- m

•

Z 2

C\J cm

3 3

~3 ~3

•

CM —

z z

• O

•

__ o

1 1 1 1

1 hill

J L

J I LlllI

■ en

5
2

_ (X) u>

'in

J I I L

Cm
©

a
E
n

Is
I*

•s <

1 =

« «

o -o

a, .2

m •-

3 g.

B
O

a
2
<

III,

J L

I i i J

"U)01

J L

'ml

o tf>

OJ —

X3N3n03dd !N30a3d

3
U

Figure 24. — Curves of apparent growth of age 0 sockeye salmon captured in tow nets in Brooks Lake

1961-63.

water temperatures during August. The lakes with the
smallest combined catches of sockeye salmon and as-
sociated species. North Arm and Brooks Lake, were
intermediate in size of age 0 sockeye salmon and in
water temperatures. I conclude that the abundance of
associated species such as pond smelt and stickle-
backs is not restricting the growth of juvenile sockeye
salmon. The growth of juvenile sockeye salmon seems
to be more directly related to temperature, but the
mechanism is not known.

Differences in time of recruitment of fry. — A greatly
prolonged period of recruitment of fry from the spawn-
ing grounds has been reported for Karluk Lake on
Kodiak Island (Burgner et al., 1969) and in 1 yr at
Kitoi Lake on Afognak Island (Smoker, 1957). In the
Naknek system a late recruitment of recently emerged
sockeye salmon fry was apparent only in Grosvenor
Lake and the length frequency graphs for this lake
(Fig. 15) show two groups of age 0 fish in August 1962

and 1963. The location of the peak of the larger groups
indicates a smaller average length than in the rest of
the system, which is probably due in part to late
emergence and lake entry. Grosvenor Lake is inter-
mediate in the system in regard to summer water
temperatures (Table 22) and in productivity (Burgner
et al.. 1969).

Differences in rates of dispersion of large and small or
fast- and slow-growing fish. — Differences in the aver-
age size of age 0 sockeye salmon at various distances
from the major spawning grounds have been reported
in Lake Aleknagik of the Wood River system (Pella,
1968) and Iliamna Lake of the Kvichak system (see
footnote 12). In these two lakes differences in lengths
of juvenile sockeye salmon could be explained by the
more rapid migration of larger and faster growing fish.
Within the Naknek system the earlier migration of the
larger fish is apparent in the differences between the
size of the summer migrants in Coville River and the
fish collected at the same time with tow nets in Coville

Table 22. --Mean surface water temperatures in August, mean number of age 0 and age I sockeye
salmon and of pond smelt and threespine and ninespine sticklebacks, and mean fork lengths of
age 0 sockeye salmon in seven lakes of the Naknek River system, 1961-63.

Mean3 fork

Mean1

length of

surface

age 0

water
temperature

Mean

number

of fish per tow2

Four
species

sockeye

Sockeye

salmon

Pond

Threespine

Ninespine

salmon

Lake or basin

(°C)

Age 0

Age I

smelt

sticklebacks

combined

(mm)

Coville Lake

14.4

140

Grosvenor Lake

10.8

<0.S

<0.5

Iliuk Arm

10.2

<0.S

South Bay

11.7

<0.S

West End

12.5

<0.S

<0.5

108

192

North Arm

11.8

<0.5

<0.S

Brooks Lake

11.9

<0.S

<0.5

<0.S

JMean of all observations made in each lake during tow netting in August 1961-63.
2Mean for 1961-63, August 16 to September 1, for species other than sockeye salmon, and
post-August 11 for sockeye salmon.

3Mean for 1962 and 1963 on August 20.

Lake. The migration of larger fish may also be the
cause of the reversal between 20 August and 1 Sep-
tember of the relation between the average size of age
0 sockeye salmon in Iliuk Arm and the average size in
Coville Lake (Table 21). The average size of age 0 fish
was smaller in Iliuk Arm than in Coville Lake on 20
August (1961, 1962, 1963), but by 1 September the fish
were larger in Iliuk Arm than in Coville Lake.

Differences in size of fry at time of emergence.

— Differences in the average sizes of fry produced by
different spawning groups within a system have been
documented (Raleigh, 1967; Brannon, 1967). McCart
(1967) considered the question of differences in size of
fry and suggested that they could result from differ-
ences in the size of adults and of eggs. Although de-
tailed study may reveal differences in the size of fry
within the Naknek system, the similarities of shape
and in location of peaks of length frequency graphs for
late July during this study do not indicate differences
in the size of fry at the time they leave the gravel.

Species Commonly Associated with
Juvenile Sockeye Salmon

Data on size, length frequency, and growth of
species commonly captured with juvenile sockeye
salmon in tow nets are too few to permit description of
growth. Therefore I discuss data ori length frequency
in some of the lakes for only three species — threespine
sticklebacks, ninespine sticklebacks, and pond smelt.

Two to several age groups were present in the length
frequency samples of each associated species, usually
including age 0 fish. The variation in the rate of cap-
ture of age 0 fish with season and species and in year-
class strength from year to year makes it difficult to
compare the abundance either between species or
within a species at different times.

Threespine sticklebacks. — The length frequencies of
threespine sticklebacks from nine samples collected
with tow nets from 1961 to 1964 are presented in Fig-
ure 25. Although these samples represent diverse
areas and times, two important facts were evident. ( 1)
Age 0 threespine sticklebacks did not appear in tow
net catches in appreciable numbers until late August
when they ranged to about 30 mm fork length. When
sticklebacks hatch in early July they are about 5 mm
long and they grow to about 7 mm in their first week.12
I substantiated these laboratory observations by visual
observations of small threespine sticklebacks close to
shore near the outlet of Coville Lake during seining
and diving in July 1963. (2) Although it is probable that
only two age classes other than age 0 made up most of
the population, the older classes usually could not be
separated on the basis of length because of a broad

12Based on observations of progeny of a pair of threespine
sticklebacks from Brooks Lake that spawned in an aquarium. (W.
R. Heard, National Marine Fisheries Service, Auke Bay Fisheries
Laboratory. Auke Bay. AK 99821, pers. comm.)

overlap in length offish assumed to be age I and older.

These two general observations also appear to be
true for threespine sticklebacks in Karluk Lake and in
Bare Lake on Kodiak Island (Greenbank and Nelson,
1959) and in lakes of the Wood system (Rogers, 1968).
When comparing my data with those of Greenbank
and Nelson it appears that they overlooked the real
age 0 fish when they did appear in the length frequency
graphs (only on 27 August 1954 for Bare Lake and
probably from 17 August to 13 September according to
length frequency histograms for fish from Karluk
Lake). As a result, Greenbank and Nelson may be 1 yr
off in assigning ages to fish represented by portions of
these histograms. Kerns (1961), however, was able to
separate age I threespine sticklebacks from age 0 and
age II and older fish by length.

European workers also have difficulty in separating
age groups of threespine and ninespine sticklebacks on
the basis of size distribution because of the slow
growth of the age I and older fish and a resulting over-
lap in size of the various year groups (Jones and
Hynes, 1950).

The largest threespine stickleback I measured was
66 mm in fork length and came from West End. It
appears that few threespine sticklebacks survive after
spawning in their third or fourth summer.

Ninespine sticklebacks. — Length frequency data are
available for only four samples of ninespine stickle-
backs (Fig. 26). It appears probable that three age
classes, 0, I, and II, are present in the length fre-
quency tabulations, but their definition by length is not
possible because of the broad overlap in length. Wal-
lace (1969) could not separate the age classes of nine-
spine sticklebacks from the Naknek system, although
he examined otoliths as well as length frequencies. A
higher proportion of ninespine sticklebacks than
threespine sticklebacks was in the 60 mm and greater
size groups. The relatively fewer ninespine than
threespine sticklebacks less than 36 mm may be due to
differences in habitat preference or size of age 0 fish of
the two species. The ninespine sticklebacks were
more abundant than the threespine sticklebacks only
in the shallower water of Coville Lake. The largest
ninespine stickleback collected was 72 mm in fork
length.

Pond smelt. — Length frequency data are presented
for four samples of pond smelt from Coville Lake and
one from West End in Figure 27. As with sticklebacks,
age 0 pond smelt did not appear in the tow net catches
until late August. The fork length of 73 pond smelt
collected with a small-mesh dip net near the outlet of
Coville Lake on 18 July 1962 ranged from 26 to 48 mm.
These fish were probably all in their second summer.
It appears that at least three age classes, 0, II, and II,
are in the samples represented in Figure 27 and that
there is broad overlap in length of the age I and older
fish.

>-
<

ro ^

o *S.

=> ^^

i l 1 l i

«•

< *

I ~

^ •

H 1-

• ^""^^

3 Q. « ■

O UJ ^

(/) t/> ^^-^„^^

-^^^^

1 1 1

iii"

<\l

—

■o •■

H >■

a. _^«^ ■

III

UJ _----^

" ^ ;

i i i i i

z 00

OD .

« X
Uj Co

* ~

cc o

O 3

z <

~ ;

o <t

J I I I l_

-J

J I l_

o
en

<i> Z

I-

» o

in z

. 0J or

* o

o
.S

If)

>- 10

< 0i

CD -

I *

•

r- ~

•

3 O

O =>

CO <

1 1 1 1

1 1

a
z

•

1-"

•

1 1 1

1 1

X0N3n03ad INSOdSd

SOUTH BAY
AUG.3I,I964,N=I9

I i | 7| i | ' | I I i | l I I I I | I I l | l | I *~* I i I I I ' I ' I I I ' I ' I t | i | l | l | i |

FORK LENGTH (MM)

Figure 26. — Length frequency distributions of four samples of ninespine sticklebacks captured in tow nets of
lakes of the Naknek River system, 1961, 1963, and 1964.

Wallace (1969) presented more data on this species
in the Naknek system and concluded from analysis of
otoliths that as many as six age classes are present in
the lakes, but that broad overlap of lengths for each
age prevents separation of age I and older fish by
length.

Pygmy whitefish and least cisco.— Pygmy whitefish
and least cisco were seldom captured in tow nets, but
were frequently taken in gear that sampled near shore
or close to the bottom. The life history of the pygmy
whitefish in the Naknek system was described by
Heard and Hartman (1966) who found populations of
large and small types of pygmy whitefish in the sys-
tem. The maximum ages and sizes reported by Heard
and Hartman were age III and 84 mm for the small
type and age V and 163 mm for the large type.

Information on the life history of the least cisco in
the Naknek system was compiled by Wallace (1969).
The range in ages and sizes he encountered were age 0
to VI and length 61 mm to 336 mm. The largest and
oldest individuals came from Coville Lake.

PREDATION ON JUVENILE
SOCKEYE SALMON

Predators have been considered to be both signifi-
cant in determining the abundance of juvenile sockeye
salmon (Foerster and Ricker, 1942; Rounsefell, 1958)
and not significant (DeLacy and Morton, 1943; Roos,
1959). During the present study it became obvious that
the Naknek system contained many potential sockeye
salmon predators and. although no specific studies of

predation were made, data and observations were col-
lected incidentally. This information is included here
to add to the overall knowledge of the biology of the
system and to aid in the planning for future studies. In
the apparent general order of importance, the fish that
prey on sockeye salmon in the Naknek system are:
lake trout, Arctic char, Dolly Varden, rainbow trout,
northern pike, and juvenile coho salmon. The burbot
and humpback whitefish are probable predators. Arc-
tic terns. Sterna paradisaea, and Bonaparte's gulls.
Lams philadephia, appear in large numbers and feed
actively at the mouth of Coville River during the
summer migration of age 0 sockeye salmon. Mergan-
sers and other fish-eating ducks occur throughout the
system.

Lake Trout

Many studies have shown that fish frequently con-
stitute the major portion of the diet of lake trout. Van
Oosten and Deason (1938) present a summary of ear-
lier literature on food habits and present additional
data. More recent studies were done by Miller and
Kennedy (1948) and Rawson (1951). Lake trout are
often found in deep cool water, but have been ob-
served in shallow water when water temperatures
permit (Rawson, 1951; Connecticut State Board of
Fisheries and Game, 1942). Lake trout occur in most
of the Naknek system (Table 3), but information on
their food habits is available only for a group occurring
in shallow water at or near the outlet of Coville River
in Grosvenor Lake.

Lake trout feed voraciously near the mouth of

z —
w 10'

_ CD

_<o

Ijj

•tf

—

s
o

UJ 10

* £7)

< —

-> r-

Ul~

dee

> o

O 3

o <

1 1

IX)

> o-

O UJ
O to

omomo o<r>om
n — — ro cm eg —

ADN3n03dJ lN3Da3d

Table 25. --Stomach contents of IS lake trout1 captured by angling in 1963 in Grosvenor Lake near
the mouth of Coville River on July 3-5 (before presmolt sockeye salmon migration) and of 21 lake
trout2 captured on August 4-5 (during migration).

Dates of
capture

Stomachs

containing

food

Stomachs with
sockeye salmon

Sockeye salmon
per stomach

Mean

Range

Stomachs with
other fish

Other fish
er stomach
Range

Mean

July 3-5
August 4-5

0
21

4-167

6
21

30.1
u6.6

0-2
1-20

!Mideye fork length range 47.0-60.9.
2Mideye fork length range 46.2-58.9.

3Some of these were salmonlike but were too digested for positive identification.
4Most of these were too digested to be identified, but many were the size of sockeye salmon
found in some stomachs.

Coville River during the summer migration of pre-
smolt sockeye salmon from Coville Lake and are read-
ily taken by angling in the vicinity both before and
during this migration. In 1963, 39 lake trout were cap-
tured by angling — 18 on 3-5 July before the presmolt
migration and 21 on 4-5 August when many age 0 sock-
eye salmon were migrating from Coville Lake. The
stomach contents were examined. None of the trout
captured on 3-5 July contained sockeye salmon and half
of the stomachs were empty (Table 23). In the trout
collected on 4 and 5 August every stomach contained
sockeye salmon.

Lake trout have been observed in this area all sum-
mer (June to September) and it is possible that they
constitute a local resident population rather than being
fish attracted from wide areas of the lake. On 12 Au-
gust 1964 I fished with sport gear in Grosvenor Lake
near each of the four smaller stream tributaries (omit-
ting Hardscrabble Creek) and hooked lake trout read-
ily in each area. Visual observation from a slowly
moving boat along the shores of the south side of
Grosvenor Lake (the tributaries are all on the north
side) revealed many solitary lake trout.

Two lake trout (both about 47 cm long) were taken
by angling in Grosvenor River near Grosvenor Lake
on 20 May 1962. These fish contained age 0 sockeye
salmon, ninespine sticklebacks, and sculpins.

I determined the length frequencies for 70 lake trout
collected by angling in Grosvenor Lake at the
mouth of Coville River and for 26 taken in gill nets
(about 10-cm stretch mesh) in areas C-4 and C-5 of
Coville Lake in 1963 (Table 24) and for 94 collected by
angling in Grosvenor Lake near the mouth of Coville
River in 1964 (Table 25). The gill nets fished in Coville
Lake included small mesh sizes down to those which
held ninespine sticklebacks and so could have cap-
tured smaller lake trout, which are notably absent in
the length frequencies. Most of the fish fell in the 45-
to 58-cm size range, but fish as short as 40 cm and as
long as 69 cm were collected. The length frequency
data indicate the presence of several year classes (be-
cause of the wide range in length and an expected slow

growth) and the absence of the larger lake trout found
in some northern lakes (for example. Great Slave
Lake [Rawson, 1951]) and of the smaller sizes.

Humpback Whitefish

Humpback whitefish are widely distributed in the

Table 24. --Length frequencies of lake trout
captured in Grosvenor Lake at the mouth of
Coville River by angling, June 22 to August 5,
and in units C-4 and C-5 of Coville Lake in
gill nets,1 June 22 to July 20, 1963. The mid-
eye- fork lengths of fresh dead fish were mea-
sured with calipers to the nearest millimeter.

Length

Fish captured

group

June

22 -August 5

June 22-July 20

(cm)

Grosvenor Lake

Coville Lake

40.1-42.0

42.1-44.0

44.1-46.0

46.1-48.0

48.1-50.0

50.1-52.0

52. 1-54.0

54.1-56.0

56.1-58.0

58.1-60.0

60.1-62.0

62.1-64.0

64.1-66.0

66.1-68.0

68.1-70.0

Total

^e net was 10-cm stretch mesh and the
other consisted of equal length sections of
different sizes: 9.5 mm, 12.7 mm, 19.0 mm,
22.2 mm, and 25.4 mm. About half the trout
were captured in the 10-cm net and the rest
in the smaller sizes, but the exact sizes were
not recorded.

Table 25. --Length frequencies of lake trout
captured in Grosvenor Lake at the mouth of
Coville River by angling, July 29 to September
6, 1964. The total lengths of the live fish
were measured with a tape measure to the
closest higher inch.

Length

group

Inches

Centimeters

Fish captured

14.1-15.0

35.8-38.1

15.1-16.0

59.2-40.6

16.1-17.0

40.9-43.2

17.1-18.0

43.4-45.7

18.1-19.0

46.0-48.3

19.1-20.0

48.5-50.8

20.1-21.0

51.1-53.3

21.1-22.0

53.6-55.9

22.1-23.0

56.1-58.4

23.1-24.0

58.7-61.0

24.1-25.0

61.2-63.5

25.1-26.0

63.8-66.0

Total

Naknek River system (Heard, Wallace, and Hartman,
1969), but data on length frequency and food are avail-
able only for fish collected with gill nets in Coville
Lake. Studies of the food offish in Great Slave Lake
(Larkin, 1948) and Great Bear Lake (Kennedy, 1949)
both indicated that the food of the closely related lake
whitefish, Coregonus clupeaformis, in these northern
lakes was mainly invertebrate animals and that fish
occurred only rarely. Larkin (1948) found fish remains
in only one sample (number of fish stomachs in sample
not given) for Great Slave Lake and Kennedy (1949)
reported that none of the fish in 86 samples from Great
Bear Lake contained food (202 fish examined).

Briefly, the results of the examination of the
stomachs of 38 humpback whitefish (mideye-fork
length 12.1 to 55.5 cm) taken from Coville Lake 14
July to 12 September 1963 are as follows: 23 of the 38
stomachs held no food; 4 contained unidentifiable
mass or "white paste"; 7 contained snails or mussels;
and 4 contained evidence of fish ( 1 pond smelt, 1
whitefish, 1 sculpin, and 3 unidentifiable fish remains).
It appears that fish were more important in the diet of
humpback whitefish in Coville Lake than in the lake
whitefish of Great Slave and Great Bear Lakes. The
fork lengths of 165 humpback whitefish (Table 26) col-
lected with gill nets in Coville Lake from 22 June to 12
September 1963 ranged from 121 to 560 mm and sev-
eral modes were apparent. The presence of several
modes in the length frequency indicates several year
classes; a preliminary study of scale samples from
these fish indicated that the ages ranged from 4 to 12
yr.

Arctic Char and Dolly Varden

Arctic char and Dolly Varden cannot be differen-

tiated without detailed examination and may have
been confused in many instances in the present study.
Therefore, I will refer to both species as char unless
the identification is certain.

Char occur throughout the system in lakes and fre-
quently in the streams and probably eat juvenile sock-
eye salmon when they are available. Arctic char were
taken with lake trout, but in fewer numbers, in the gill
nets in Coville Lake and by angling in Coville and
Grosvenor Rivers. The stomachs of a few Arctic char
captured by angling in Grosvenor Lake near Coville

Table 26. --Length frequencies of humpback
whitefish captured in gill nets in Coville
Lake, June 22 to September 12, 1963. The fork
lengths of fresh dead fish were measured to the
nearest millimeter.

Length group (mm)

Fish captured

121-130
161-170
171-180
181-190
191-200
201-210
211-220
221-230
231-240
241-250
251-260
261-270
271-280
281-290
291-300
301-310
311-320
321-330
331-540
341-350
361-370
571-380
381-390
391-400
401-410
411-420
421-430
431-440
441-450
451-460
461-470
471-480
481-490
491-500
501-510
511-512
513-520
521-530
531-540
541-550
551-560
Total

1
1
1
1
1
4
5
5
4
0
3
2
0
3
6
11
0
1
0
1
0
2
2
2
17
12
16
17
11
6
7
1
4
5
3
2
2
2
2
0
2

165

River contained food similar to that of lake trout from
the same area. A char (about 40 cm long) was seen
feeding in upper Grosvenor River on 20 May 1962 and
was captured by angling. This fish contained several
sockeye salmon fry, two of which were still alive.

Other Species

Only general observations are available on the other
pisciverous fish in the Naknek River system — rainbow
trout, juvenile coho salmon, northern pike, and bur-
bot.

Rainbow trout inhabit most of the larger streams in
the system and were often taken by angling in the lakes
near the mouths of these streams. Sportsmen fish for
this species in American Creek and Coville, Brooks,
and Naknek Rivers; fish above 60 cm are commonly
caught in these locations. No food studies have been
made here, but rainbow trout have been observed
feeding on young sockeye salmon that were migrating
from stream spawning grounds to the lakes and from
lake to lake via connecting rivers either as presmolts
or smolts.

Juvenile coho salmon were taken in appropriate gear
in many streams and beach areas in the system, but
were virtually absent from tow net samples. Because
of their relatively small size (no juvenile coho salmon
over 130 mm were taken), I would expect coho salmon
to be most effective as predators on sockeye salmon
during the first few weeks after the sockeye salmon
leave the gravel — in streams and lake margins before
the sockeye salmon become pelagic.

Northern pike are widely distributed in the Naknek
system wherever suitable habitat is found. The lake
areas where northern pike seem to be abundant are the
shallow north end of Coville Lake and the shallow
waters of Northwest Basin. Generally the habitat in
which northern pike are abundant does not contain
many juvenile sockeye salmon. Possible exceptions to
this occur in Grosvenor River near Grosvenor Lake
and in the upper Naknek River where lagoons contain-
ing northern pike are closely connected to river areas
containing migrating sockeye salmon. Sockeye salmon
have not been reported in stomachs of northern pike
from this system.

Burbot have been captured in Iliuk Arm, South
Bay, and North Arm (Heard, Wallace, and Hartman.
1969), but were never abundant. They were caught in
gill nets and trap nets in South Bay, in trap nets in
North Arm, and in seines in Iliuk Arm. In Lake
Michigan, the stomachs of lake trout and burbot that
were captured in gill nets contained the same kinds of
fish, but the burbot contained only 74% fish by volume
and the lake trout contained 98% (Van Oosten and
Deason, 1938). Both species were predators on
coregonids. No data are available on the diet of burbot
in the Naknek system, but apparently so few are pres-
ent that they would not be a significant predator even
if sockeye salmon were important in their diet.

General Significance of Predation

Although many species of fish and birds are poten-
tial or known predators on juvenile sockeye salmon in
the Naknek system, the role of predators in determin-
ing freshwater survival is unknown. The abundance of
smolts from the escapement of 1961 shows that smolt
production per adult may be high in spite of predation
in the Naknek system. In 1961 a relatively small es-
capement of about 350,000 adult sockeye salmon en-
tered the Naknek system, of which about 220,000 went
to the most distant spawning grounds, American
Creek. The production of smolts from the total es-
capement to the system in 1961 was about 32 smolts
per adult (see footnote 3) — the highest rate recorded
for the Naknek system between 1956 and 1963. The
survival of these smolts to returning adults in 1966 and
1967 was about 15.5%, I3 very close to the long-term
average of about 16.5% (Burgner et al., 1969). A dif-
ferent distribution or abundance of adult sockeye
salmon or predators might result in a much different
effect on survival in another year.

SUMMARY AND SIGNIFICANCE FOR
RESOURCE DEVELOPMENT

Although most stocks of sockeye salmon have the
same general life history, each stock is unique because
it has its own combination of biological and physical
environments. The principal objective of this study
was to determine the distribution, abundance, and
growth of juvenile sockeye salmon in the Naknek
River system, Bristol Bay, Alaska. The work was
done from 1961 through 1964.

The Naknek system contains eight interconnected
and generally biologically discrete lakes or basins with
different ratios of potential spawning grounds to rear-
ing area for sockeye salmon and different densities of
juvenile sockeye salmon and associated species of
fish. The sockeye salmon was the most common and
abundant fish in all basins, followed by threespine
sticklebacks, ninespine sticklebacks, and pond smelt.
Eighteen other species of potential competitor or
predator fish were present.

Juvenile sockeye salmon in the pelagic areas had a
characteristic pattern of abundance in tow net catches
during the summer of 1961-64. For the entire system
the abundance of age 0 fish increased from early sum-
mer to midsummer and then declined to late August.
The abundance in late August varied by a factor of
about 2.5 and, although data are available for only 4
years, the abundance appears to be independent of
variations in the number of parents from 1960 to 1963.

In July the catches of age 0 sockeye salmon in each
basin were about proportional to the abundance of
contiguous spawning grounds, but by late August this
relation no longer existed. This change was at least

l:,C.J. DiCostanzo, National Marine Fisheries Service. Auke Bay
Fisheries Laboratory, Auke Bay, AK 99821, pers. comm.

partly due to migration of age 0 fish — generally from
basins of greater abundance offish to others of lesser
abundance. The larger and faster growing fish proba-
bly were the first to migrate. Not all basins were in-
volved in these migrations.

In the Naknek system smolt production has varied
only about twofold with parent escapements of
350,000 to 2,000,000 (escapements of less than 300,000
have produced markedly fewer smolts). Several fac-
tors are suggested as contributing to this relatively uni-
form production of smolts. The maintenance of a
minimum level of fry production is enhanced by the
presence of several major spawning units or races in
widely separated spawning grounds of different types.
This combination helps ensure against a total loss of a
year's production of eggs and alevins due to adverse
physical conditions on the spawning grounds. Exam-
ples of the value of having different types of habitat
are: scouring action of floods would not affect beach
spawning areas; extreme freezing would not greatly
reduce the flow in major rivers connecting lakes; and
warm dry weather causing low lake levels and low
flows in small streams would increase the flow of
streams fed by snow and icefields. The possibility of
full utilization of fry is greatly enhanced by the pres-
ence of several connected lakes and the migratory be-
havior of the juvenile sockeye salmon during their first
summer.

No indications that the population of juvenile sock-
eye salmon was near its upper limit were apparent
during this study. In other systems the first obvious
effect of too high populations is a reduction in growth.
Such a reduction was not evident in juveniles in the
lakes of the Naknek system in 1961-64 and apparently
did not occur in the period 1957-65, as evidenced by
the size of age I smolts — age I smolts from the Naknek
system are as large as, or larger than, those of other
Bristol Bay systems (Burgner et al., 1969). Much of
the variation in the average length of age I smolts
(-8.5% to + 6.6% of the mean of 99.4 mm) in the Nak-
nek system is thought to be due to variations in grow-
ing conditions in the spring just before the smolts leave
(Burgner et al., 1969).

The data on abundance and growth of juvenile sock-
eye salmon and the distribution of the escapement and
spawning grounds indicate the possibility that produc-
tion of sockeye salmon in the Naknek system could be
greatly increased. Two of the major basins. North
Arm and Brooks Lake, which constitute about 35% of
the system, are now producing relatively few
juveniles. The low production of juvenile salmon in
both basins appears to be the result of too few fry
being produced by the spawning grounds, but the
reason for the low production of fry differs in the two
basins: North Arm has limited but heavily used
spawning grounds, whereas Brooks Lake has appar-
ently adequate but lightly used spawning grounds.

North Arm contains about 24% of the rearing area
of the system but only about 2% of the spawning

grounds (and usually receives about 2% of the es-
capement) and the basin does not receive juveniles
from other areas. Even full use of all the present
spawning grounds in North Arm would probably re-
sult in too few juveniles to use the rearing area fully.

Ninety-five percent of the area of potential spawn-
ing grounds and 90% of the escapement in North Arm
are in Bay of Islands Creek and most of the rest is
distributed among seven small streams. Bay of Islands
Creek runs about 27 km from a high tundra lake down
to North Arm. A falls impassable to salmon is located
about 14 km upstream from North Arm. The probable
difficulty in making the falls passable and the potential
of the stream above the falls for production of sockeye
salmon are unknown. It is possible that the present
production of sockeye salmon in North Arm could be
increased significantly by simply making all of Bay of
Islands Creek accessible to spawners. The lake prob-
ably could support 10 to 20 times the present density of
juvenile salmon. The increase in fry production re-
quired to produce the numbers of lake residents North
Arm could support could be obtained from a combina-
tion of enhancement techniques used elsewhere.

The reason for the low production of juvenile sock-
eye salmon by Brooks Lake is not clear. The major
spawning area. Headwater Creek, has an estimated
spawning ground capacity of about 40,000 adult sock-
eye salmon, but the largest number recorded in the last
20 yr was about 1 1 ,000. An intensive study of the biol-
ogy of the sockeye salmon of Headwater Creek could
be expected to reveal the time, place, and cause of
mortality in fresh water. With this information action
could be taken to bring Brooks Lake into full produc-
tion.

Three factors in the biology of juvenile sockeye
salmon of the Naknek system are of special signifi-
cance to the managers of the resource: (1) the abun-
dance of smolts each spring has been fairly constant
for the system as a whole and not closely related to the
abundance of the parents, or from 1961-64, apparently
even to the abundance of age 0 fish during early sum-
mer, (2) the apparent growth of juvenile sockeye
salmon and potential competitor species was not re-
lated to the abundance of these fish in any lake of the
Naknek system, and (3) two major lakes, constituting
about 35% of the rearing waters, do not receive age 0
sockeye salmon from other basins and are supporting
relatively few sockeye salmon.

These three factors and their causes and effects
could form the basis for a program to increase the
production of sockeye salmon by the Naknek River
system.

The question of what escapement of adult sockeye
salmon is needed to ensure full production of juveniles
is of primary importance to fishery managers. From
1961 to 1964. as few as 350,000 adult spawners were
apparently adequate in the Naknek system. However,
the special circumstance involved, i.e., the majority of
fish in this low escapement used one spawning area

with probable special benefits, must be considered. It
seems that to ensure full production with adequate in-
surance against catastrophes, every major spawning
ground should be utilized every year. On the basis of
the smolt-escapement data, Burgner et al. (1969)
placed the desired escapement for the Naknek system
at 600,000 to 1.000,000 fish. The present study indi-
cates that escapements in this range probably fully
utilize the present combination of spawning and rear-
ing areas without danger of overburdening the food
supply.

ACKNOWLEDGMENTS

The original planning and development of proce-
dures in 1960 and 1961 were done by Charles J. Di-
Costanzo, Wilbur L. Hartman, and Richard R. Straty.
The organization and direction of field crews and de-
velopment of techniques through 1961 were accom-
plished largely by R. L. Wallace and W. R. Heard.
Wallace continued as field leader through 1963. The
extensive sampling was done by the cooperative ef-
forts of about 30 different seasonal aids from 1961
through 1964. The analysis of variance tests of results
of tow net sampling were done under the guidance of
James C. Olsen. The estimates of the numbers of fry
migrating during Latin-square sampling were done
under the guidance of Jerome J. Pella with a computer
program he designed.

LITERATURE CITED

ANDREW. F. J., and G. H. GEEN.

1960. Sockeye and pink salmon production in relation to pro-
posed dams in the Fraser River system. Int. Pac. Salmon
Fish. Comm.. Bull. 11. 259 p.
BRANNON, E. L.

1967. Genetic control of migrating behavior of newly emerged
sockeye salmon fry. Int. Pac. Salmon Fish. Comm., Prog.
Rep. 16, 31 p.
BURGNER, R. L.

1958. A study of fluctuations in abundance, growth, and surviv-
al in the early life stages of the red salmon (Oncorhynchus
nerka Walbaum) of the Wood River Lakes, Bristol Bay,
Alaska. Ph.D. Thesis, Univ. Wash., Seattle, 200 p.
1962. Sampling red salmon fry by lake trap in the Wood River
Lakes, Alaska. In Ted S. Y. Koo (editor). Studies of Alaska
red salmon. Univ. Wash. Publ. Fish., New Ser. 1:315-348.
1964. Factors influencing production of sockeye salmon (On-
corhynchus nerka) in lakes of southwestern Alaska. Int. Ver.
Theor. Angew. Limnol. Verh. Proc. 15:504-513.
BURGNER, R. L., C. J. DICOSTANZO. R. J. ELLIS, G. Y.
HARRY, JR.. W. L. HARTMAN. O. E. KERNS. JR., O. A.
MATHISEN, and W. F. ROYCE.
1969. Biological studies and estimates of optimum escapements
of sockeye salmon in the major river systems in southwestern
Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 67:405-459.
COCHRAN, W. G., and G. M. COX.

1957. Experimental designs. 2d ed. Wiley & Sons, N.Y., 611 p.
CONNECTICUT STATE BOARD OF FISHERIES AND
GAME.
1942. A fishery survey of important Connecticut lakes. Conn.
Geol. Nat. Hist. Surv., Bull. 63, 339 p.

CRADDOCK, D. R.

1961. An improved trap for the capture and safe retention of
salmon smolts. Prog. Fish-Cult. 23:190-192.
DELACY, A. C, and W. M. MORTON.

1943. Taxonomy and habits of the charrs. Salvelinus malma
and Salvelinus aplinus, of the Karluk drainage system.
Trans. Am. Fish. Soc. 72:79-91.

FOERSTER. R. E.

1944. The relation of lake population density to size of young
sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board
Can. 6:267-280.

FOERSTER, R. E., and W. E. RICKER.

1942. The effect of reduction of predaceous fish on survival of

young sockeye salmon at Cultus Lake. J. Fish. Res. Board

Can. 5:315-336.
GREENBANK. J., and P. R. NELSON.

1959. Life history of the threespine stickleback Gasterosteus

aculeatus Linnaeus in Karluk Lake and Bare Lake. Kodiak

Island, Alaska. U.S. Fish Wildl. Serv., Fish. Bull.

59:537-559.
GROOT. C.

1965. On the orientation of young sockeye salmon (Oncorhyn-
chus nerka) during their seaward migration out of lakes. Be-
haviour, Suppl. 14. 198 p.

HARTMAN. W. L., W. R. HEARD, and B. DRUCKER.

1967. Migratory behavior of sockeye salmon fry and smolts. J.
Fish. Res. Board Can. 24:2069-2099.
HEARD. W. R.. and W. L. HARTMAN.

1966. Pygmy whitefish Prosopium coulteri in the Naknek River
system of southwest Alaska. U.S. Fish Wildl. Serv.. Fish.
Bull. 65:555-579.

HEARD, W. R., R. L. WALLACE, and W. L. HARTMAN.

1969. Distributions of fishes in fresh water of Katmai National
Monument, Alaska, and their zoogeographical implications.
U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 590. 20 p.
JOHNSON. W. E.

1956. On the distribution of young sockeye salmon (Oncorhyn-
chus nerka) in Babine and Nilkitkwa Lakes, B.C. J. Fish.
Res. Board Can. 13:695-708.

1958. Density and distribution of young sockeye salmon (On-
corhynchus nerka) throughout a multibasin lake system. J.
Fish. Res. Board Can. 15:961-982.
JONES, J. W., and H. B. N. HYNES.

1950. The age and growth of Gasterosteus aculeatus, Pygos-
teus pungitius, and Spinachia vulgaris, as shown by their
otoliths. J. Anim. Ecol. 19:59-73.

KENNEDY, W. A.

1949. Some observations on the coregonine fish of Great Bear
Lake. N.W.T. Fish. Res. Board Can., Bull. 82. 10 p.

KERNS. O. E, JR.

1961. Abundance and age of Kvichak River red salmon smolts.
U.S. Fish Wildl. Serv., Fish. Bull. 61:301-320.
KROGIUS, F. V.

1961. O sviaziakh tempa rosta i chislennosti krasnoi (On the
relation between rate of growth and population density in
salmon). Tr. Soveshch. Ikhtiol. Kom. Akad. Nauk SSSR
13:132-146. (Translated by R. E. Foerster. 1962, 17 p.. 6 fig .:
available Fish. Res. Board Can.. Transl. Ser. 411)
LARK.1N. P.A.

1948. Pontoporeia and Mysis in Athabaska. Great Bear and
Great Slave Lakes. Fish. Res. Board Can.. Bull. 78.
33 p.
MCCART. P.

1967. Behaviour and ecology of sockeye salmon fry in the
Babine River. J. Fish. Res. Board Can. 24:375-428.

MERRELL, T. R, JR.

1964. Ecological studies of sockeye salmon and related lim-
nological and climatological investigations. Brooks Lake,
Alaska. 1957. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish.
456, 66 p.
MILLER, R. B., and W. A. KENNEDY.

1948. Observations on the lake trout of Great Bear Lake. J.
Fish. Res. Board Can. 7:176-189.
NARVER, D. W.

1968. Identification of adult sockeye salmon groups in the
Chignik River system by lacustrine scale measurement, time
of entry, and time and location of spawning. In R. L. Burgner
(editor). Further studies of Alaska sockeye salmon. Univ.
Wash. Publ. Fish., New Ser. 3:113-148.
NARVER, D. W., and M. L. DAHLBERG.

1964. Chignik sockeye salmon studies. In Ted S. Y. Koo
(editor). Research in fisheries. . . .1963. p. 18-21. Univ.
Wash.. Coll. Fish., Contrib. No. 166.
PELLA, J. J.

1968. Distribution and growth of sockeye salmon fry in Lake
Aleknagik, Alaska, during the summer of 1962. In R. L.
Burgner (editor). Further studies of Alaska sockeye salmon.
Univ. Wash. Publ. Fish.. New Ser. 3:45-1 II.
RALEIGH, R. F.

1967. Genetic control in the lakeward migrations of sockeye
salmon (Oncorhynchus nerka) fry. J. Fish. Res. Board Can.
24:2613-2622.

RAWSON, D. S.

1951. Studies of the fish of Great Slave Lake. J. Fish. Res.
Board Can. 8:207-240.
ROGERS, D. E.

1968. A comparison of the food of sockeye salmon fry and
threespine sticklebacks in the Wood River Lakes. In R. L.

Burgner (editor). Further studies of Alaska sockeye salmon.

Univ. Wash. Publ. Fish., New Ser. 3:1-43.
ROOS. J. F.

1959. Feeding habits of the Dolly Varden. Salvelinus malma

(Walbaum). at Chignik, Alaska. Trans. Am. Fish. Soc.

88:253-260.
ROUNSEFELL, G. A.

1958. Factors causing decline in sockeye salmon of Karluk
River. Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 58:83-169.

RUGGLES, C. P.

1966. Juvenile sockeye studies in Owikeno Lake, British
Columbia. Can. Fish Cult. 36:3-21.
SCHEFFE, H.

1959. The analysis of variance. Wiley & Sons, N.Y., 477 p.
SIEGEL, S.

1956. Nonparametric statistics for the behavioral sciences.
McGraw-Hill, N.Y., 312 p.

SMOKER, W. A.

1957. Kitoi Bay research station. Alaska Fish. Board and
Alaska Dep. Fish. Annu. Rep. 1956. Rep. No. 8. p. 35-39.

SNEDECOR. G. W.

1956. Statistical methods, applied to experiments in agriculture
and biology. 5th ed. Iowa State Coll. Press, Ames, 534 p.
VAN OOSTEN, J., and H. J. DEASON.

1938. The food of the lake trout (Cristivomer namaycush
namaycush) and of the lawyer [Lota maculosa) of Lake
Michigan. Trans. Am. Fish. Soc. 67:155-177.
WALLACE. R. L.

1969. Some aspects of the comparative ecology of fishes as-
sociated with juvenile sockeye salmon, Oncorhynchus nerka
(Walbaum). in the lakes of the Naknek River system, Alaska.
PhD. Thesis, Oreg. State Univ., Corvallis, 160 p.

GPO 991 -397

648. Weight loss of pond-raised channel catfish [Ictalurus punctatus) during holding in
processing pianl vats By Donald C. Greenland and Robert L. (iill I lecember 1971, iii + 7
- , 2 tables Fur sale bv the Superintendent of Documents. I S Government
Printing Office, Washington D C 20402.

649 Distribution of forage of skipjack tuna (Euthynnus pelamis) in the eastern tropical
Pacifii B> Maurice Blackburn and Michael Laurs. January 1972, iii + 16 pp .

tahles For sale bj the Superintendent of Documents, U S Government tainting Office,
Washington D.C 20402

650 Effects ol s antioxidants and EDTA on the development of rancidity in Spanish

mackerel [Scom beromorus maculatus) during frozen storage. By Robert N. Farragut.
Februarj L972, i\ + 12 pp., 6 figs., 12 tables. For sale by the Superintendent of
Documents, U S Government Printing Office, Washington, D.C. 20402.

651 The effect of premortem stress, holding temperatures, and freezing on the
biochemistry and quality of skipjack tuna. By Ladell Crawford. April 1972, iii + 23 pp
figs.. 4 tables. For sale by the Superintendent of Documents, U.S. Government Printing
Office Washington, D.C. 20402.

653 llu use of electricity in conjunction with a 12.5-meter (Headrope) Gulf-of-Mexico
shrimp trawl in Lake Michigan. By James E. Ellis. March 1972. iv + 1(1 pp., 11 figs., 4
tables For sale by the Superintendent of Documents, U.S. Government Printing < Iffn e,
Washington, D.C -'0402.

654. An electric detector system for recovering internally tagged menhaden, genus
Breuoortia Bj K 0. Parker. Jr February 1972. iii + 7 pp , 3 figs , I appendix lable. For

sale bv the Superintendent of Documents. U.S. Government Printing Office, Washington,
D.C. 20402

655. Immobilization of fingerling salmon and trout bv decompression Bv Doyle F.
Sutherland. March 1972. iii + 7 pp . 3 figs . 2 tables. For sale by the Superintendent of
Documents, U.S. Government Printing Office. Washington. D.C 1040

662 Seasonal distribution of tunas and billfishes in the Atlantic. By John P Wisi and
Charles W Davis. January 1973, i\ + 24 pp., 13 tigs.. 4 tables. For sale bj the Superinten-
dent of Documents, U.S. Government Printing Office, Washington. D C 2H402

663. Fish larvae collected from the northeastern Pacific Ocean and Puget Sound during
April and May 1967. By Kenneth D Waldron. December 1972. iii + 16 pp., 2 figs., I table.
4 appendix tables. For sale by the Superintendent of Documents, U.S. Government Prim
ing Office. Washington, D C 20402.

6fi4. Tagging and tag-recovery experiments with Atlantic menhaden. Brevoortia tyran-
nic By Richard L. Kroger and Robert L, Dryfoos December 1972, iv -f 11 pp., 4 figs.. 12
tables For sale by the Superintendent of Documents, U.S. Government Printing Office.
Washington, D C 204Q !

665. Larval fish survey ol Humboll B.n . < lalifornia. By Maxwell B. Eldndge and Charle..
F, Bryan. December 1972, iii + 8 pp., 8 figs., 1 table. For sale by the Superintendent ol
Documents, U.S. Government Printing Office, Washington. D.C. 20402.

666, Distribution and relative abundance of fishes in Newport River, North Carolina. B>
vVilliam R Turner and George V Johnson September 1973. iv + 23 pp.. 1 fig.. 13 tables,
For sale bv the Superintendent of Documents. U.S. Government Printing Office,
Washington. DC 20402

667, An analysis of the com nun ial lobster I Homarus ar.iericanus) fishery along the coast
of Maine. August 1986 through December 1970 Bv James C, Thomas. June 1973 * * 57
pp 18 figs., 11 tables. For sale by the Superintendent of Documents. U.S. Government
Printing Office. Washington. D.C '"I".

668. An annotated bibliography ol the i miner, Tautogolabrus adspersus (Walbaum). By
Fredrii M Serchuk and David W Frame May 1973, ii + 43 pp. For sale by the
Superintendent of Documents, U.S. Government Printing Office, Washington. D.C.
20402

656. The calico scallop. Argopecten gibbus. By Donald M Allen and T J Costello May
1972. in + 19 pp., 9 figs., 1 table. For sale by the Superintendent of Documents, U.S
Government Printing Office. Washington. D.C 20402

669 Subpoinl prediction tor direct readout meteorological satellites. By L. E. Eber
August 1973, iii + 7 pp., 2 figs 1 table Ft silt- b> the Superintendent of Documents,
I S Government Printing Office Washington D.C 20402

657 Making fish protein concentrates by enzymatic hydrolysis. A status report on
research and some processes and products studied by NMFS Bv Malcolm R. Hale.
November 1972, v + 32 pp., 15 figs., 17 tables. 1 appendix table For sale by the
Superintendent of Documents, U S Government Printing Office, Washington. D.C.

20402

658. List of fishes of Alaska and adjacent waters with a guide to some ol their literature
By Jay C. Quast and Elizabeth L. Hall. July 1972. iv + 47 pp. For sale by the Superinten
dent "l Documents, U S Government Printing Office, Washington, D.C. 20402.

670 I'nharvested fishes m the U.S. commercial fishery of western Lake Erie in 1969. By
Harr\ 1 > Van Metei lul) 1973, Iii + 11 pp., 6 figs., 6 tables. For sale by the Superinten-
dent ol Document ' S Government Printing Office, Washington, D.C 20402.

671 Coastal upwelling indices west coast ol North America, 1946-71. By Andrew

• 103 pp 6 figs . 3 tables. 45 appendix figs. For sale by the
Superintendent ol Documents U.S Government Printing Office, Washington. D.C.
!0402

659 The Southeast Fisheries Center bionumeric code Part I: Fishes By Harve) R
iullis, Jr., Richard B Roe, and Judith C. Gatlin. July 1972, xl * 95pp 2figs Forsaleby
the Superintendent ol Documents, U.S. Government Printing Office, Washington. D.C.

20402

672 Seasonal occurrence ol young Gull menhaden and other fishes in a northwestern
Florida estuarj ByMarlinE TagatzandE Peter H Wilkins August 1973. iii 4 14 pp 1

l ig . 4 tallies Forsaleby the Superintendent ol Documents, U S. Government Printing Of-
fice Washington, D I !040 !

660 A freshwater fish electro-motivator {FFEMl-its characteristics and operation Bj
James E. Ellis and Charles C. Hoopes. November 1972. in + 11 pp.. 9 tigs

661. A review of the literature on the development of skipjack tuna fisheries in the cen-
tral and western Pacific Ocean. By Frank J. Hester and Tamio Otsu, January 19
1 3 pp., [fig. For sale by the Superintendent of Documents. V S Government Printing Ol
fice. Washington, D.C. 20402

67 Abundance and distribution ol inshore bent hie fauna off southwestern Long Island,
NY By Frank W Steimle, Jr and Ri< hard H Stone. December 1973, iii + 50 pp.. 2 tigs.,
5 appendix tables.

674 Lake Erie bottom trawl explorations, 1962-66. By Edgar W. Bowman. Januan 1974,
iv + 21 pp., 9 tigs , 1 table, 7 appendix tables.

MBL WHOI Library - Serials

■ "IIM i

5 WHSE 01921

UNITED STATES
DEPARTMENT OF COMMERCE

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION POSTAGE AND FEES PAID

FOURTH CLASS

NATIONAL MARINE FISHERIES SERVICE M II K H I IX ^ ^ U S DEPARTMENT °F COMMERCE

SCIENTIFIC PUBLICATIONS STAFF | V/ V^ I \ III Vrft^^Ww COM 2 1 0

ROOM 4 50
1107 N E 45TH ST
SEATTLE WA 98105

OFFICIAL BUSINESS

Marine Biological Laboratory
Library - Periodicals
Woods Hole, Ma 02fA3

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "NOAA technical report NMFS SSRF"

See other formats