XFWS-A 590 1-20 (1969) U.S. Fish Wildl, Serv. Spec, Sci. Rep. Fish.

...... = ...... ('I) ::::::: 0'> ...... ===C\J ...... -~---<0 co C\J Distributions of Fishes in Fresh Water of - - Katmai National Monument, Alaska, and ::...:-- Their Zoogeographical Implications

~----

-

::::::::::::::: ~ IS ::::::::::::::: ::::::::::::::: '-"'-"'- SPECIAL SCIENTIFIC REPORT-FISHERIES No. 590

::::::::::::::: 11 ~ ~ ~ .A335 ~ no.590

~ ~ ~ ...... ~ ...... ~ UNITED STATES DEPARTMENT OF THE INTERIOR ~ U.S. FISH AND WILDLIFE SERVICE ~ BUREAU OF COMMERCIAL F~

~ ' SPECIAL SCIENTIFIC REPORT --FISHERIES

Robert L. Hacker, Editor

Mary Fukuyama, Assoclate Editor

PUBLICATION BOARD

John A. Guinan Jobn M. Patton, Jr.

Robert L. Hacker Edward A. Schaefers

John 1. Hodges Parker s. Trefethen

Harvey Hutcblnis Robert c. Wtlaon

Lealie W. Scatter1ood, Chairman

Special Scientific Report--Fisheries are preliminary or progress reports and reports on aclendftc inveadlatlons of restricted scope. Established as Special Scientific llepona in 1940, nos. 1 to 67 were- issued from that date to 1949, when the newaertea,SpedalSclentlftcRepon.--Flsheries, with new serial numbering, was started. Special Sciendftc Report-Pisherlea ue distributed free to Ubrarles, research instltutlona, State apndea, and adentlats.

ARLIS Alaska Resources Llbii!Y tliDformation Selvkes Ancbo~,AK UNITED STATES DEPARTMENT OF THE INTERIOR s~ U.S. FISH AND WILDLIFE SERVICE ll BUREAU OF COMMERCIAL FISHERIES ,.A ~3~ N). c;qo

Distributions of Fishes in Fresh Water of Katmai National Monument, Alaska, and Their Zoogeographical Implications

By

WILLIAM R. HEARD, RICHARD L. WALLACE and WILBUR L. HARTMAN

ARLIS Alaska Resources Library & Information Services Library Building, Suite 111 3211 Providence Drive Anchorage, AK 99508-4614 C") 0> C\1 (0 co United States Fish and Wildlife Service ,....C\1 Special Scientific Report- -Fisheries No. 590 0 0 1.0 ,...._1.0 C") C") Washington, D.C.

October 1969 CONTENTS

Page

Introduction,. ••••••••••• ...... , ...... •...... ••••••••• ...... •...... •

Fresh-water areas of the monument ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 3 Bristol Bay Drainage •••••••••••••••••••••••••••••••.••••••••••••••••••••••••••••·••••••••••••••••••••••••••• 3 Shelikof Strait Drainage ...... ••••••••••••••••••••••••••• ••••••••••••• •••••••••••••••••• •••••••••••• 4

Methods and equipment .•••...... •...... •...... •••.•.•.•.....•...... •..... •••••• ..• .• .• . •.•.. ..•.... .• . . • . . 5

Annotated list of species ...... •..... 6 Fresh-water forms•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 6 Petromyzontidae ...... ••.•...... •••...... ••..••.•..•....•.....•... 6 Arctic lamprey, Lampetra japonica (Martens) ...... 6 Salmonidae ...... •...... •...••...... ••..•...... •.... 7 Humpback whitefish, Coregonus pidschian (Gmelin) ...... 7 Least cisco, Coregonus sardinella Valenciennes ...... 7 Pink salmon, Oncorhynchus gorbuscha (Walbaum) ...... 7 Chum salmon, Oncorhynchus keta (Walbaum) ...... 8 Coho salmon, Oncorhynchus kisutch (Walbaum) ...... 8 Sockeye salmon, Oncorhynchus nerka (Walbaum) ...... 8 Chinook salmon, Oncorhynchus tshawytscha (Walbaum) ...... 9 Pygmy whitefish, Pros opium coulterii (Eigenmann and Eigenmann) ...... 9 Round whitefish, Prosopium cylindraceum (Pallas) ...... 10 Rainbow trout, Salmo gairdnerii Richardson...... 10 Arctic char, Salvelinus alpinus (Linnaeus) •••••••••••••••••••••••••• ...... 10 Dolly Varden, Salvelinus malma (Walbaum) ...... 10 Lake trout, Salvelinus namaycush (Walbaum) ...... 11 Arctic grayling, Thymallus arcticus (Pallas) ...... 12 Osmeridae ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 12 Pond smelt, Hypomesus olidus (Pallas) ...... 12 Umbridae ...... •..•.•...... ~••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 12 Alaska blackfish, Dallia pectoralis Bean., ...... 12 Esocidae ...... ••.....•...•..•.....•....•...... •.....•...... •...... •...••••....•..•.•.•...... •.•..•• 12 Northern pike, Esox lucius Linnaeus••••••••••••••••••••••••••••••••••••••••••••••••••••••• 12 Catos tomidae .....•...-:-:-:=. =-:::-: ...... 12 Longnose sucker, Catostomus catostomus (Forster) ...... 12 Gadidae ...... •....•.....••...•.••...... •...... •.•.....•...... •.... · .•.•••...••..•. 13 Bur bot, Lot a lot a (Linnaeus) ...... ••••••••••••••••••••••••••••••••••••••••••••• •••••••••••••• 13 Gasterosteidae ...•=···•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 13 Threespine stickleback, Gasterosteus aculeatus Linnaeus ...... 13 Ninespine stickleback, Pungitius pungitius (Linnaeus) ...... 13 Cottidae . .•...... •...... ••...... •.•...... ••..• 13 Coastrange sculpin, Cottus aleuticus Gilbert ...... 13 Slimy sculpin, Cottus cognatus Richardson...... 14 Euryhaline forms ...... • .•..•.•...... •...... •...... •...... •...... •••...... •...... •. 14 Osmeridae ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 14 Arctic smelt, Osmerus dentex Steindachner ...... 14 Gadidae .•.•.•..•...... ••...... •...... •...... •..•.....•...... •.....•••...••. 14 Pacific cod, Gadus macrocephalus Tilesius ...... 14 Cottidae ...... •...... •..•.•.•...... •....•...... •.•.•....•. •••••••• ...... • 14 Pacific staghorn sculpin, Leptocottus armatus Girard ...... 14 Pleuronectidae ...... ••...... •••••••••••••••••••••••••••••••••••••••••••••••••• ••••••••••••••••••••••• 14 Starry flounder, Platichthys stellatus (Pallas) ...... 14 Zoogeographical implications •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 14 Summary •. ••••• ••••••••••••• •••....•.•••..•...... ••...••...•••.•..•....•••.••..•.•..•....•.•••....••.••..•...•...•.. 18 Acknowledgments •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 19 Literature cited •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 19

iii Distributions of Fishes in Fresh Water of Katmai National Monument, Alaska, and Their Zoogeographical Implications

By

WILLIAM R. HEARD,~ RICHARD L. WALLACE, 2 and WILBUR L. HARTMAN~

ABSTRACT

Katmai National Monument covers 10,916 km.2 on the base of the Alaska Penin­ sula and is divided by the Aleutian Mountain Range into two principal drainage areas. Streams north of the Aleutian Range flow into Bristol Bay of the Bering Sea, and those south of the mountains flow into Shelikof Strait of the North Pacific Ocean. The large multilake Naknek River system is the dominant drainage area on the Bristol Bay side of the monument, whereas small single lakes and short streams and rivers constitute many separate drainages on the Shelikof Strait side. Twenty­ four species of fish occur in the Bristol Bay drainages of the monument, but only eight species were collected in streams and lakes draining into Shelikof Strait. Evi­ dently the Aleutian Range has been a barrier to the southward movement of fresh­ water fishes in the monument. All eight species in Shelikof Strait drainages are capable of dispersal through salt water, whereas several forms in Bristol Bay drainages require fresh water for dispersal. Variable numbers of species occur in the interconnecting lakes of the Naknek River system. Naknek Lake, the downstream terminus of the lake system, contains 24 known species and each upstream lake contains fewer species than the one into which it drains. The present distribution of fishes in this system is discussed in terms of the sequential timing of species invasion and the postglacial development of barriers.

INTRODUCTION Katmai National Monument is at the base of aula. The area has a long history ofvolcanism the Alaska Peninsula in southwestern Alaska (Wahrhaftig, 1965), and glaciation was exten­ (fig. 1). It was established by Presidential sive during the Pleistocene (Muller, 1952; Proclamation in 1918, primarily to preserve Karlstrom, 1957). Several streams on each the spectacular features of the Valley of Ten slope of the Aleutian Range still originate Thousand Smokes--a volcanic phenomenon that from small, vestigial glaciers. One of the most resulted in 1912 from the eruption of the vol­ important features influencing the distribution cano Mt. Novarupta. In subsequent years the of fishes in the monument is the large complex boundaries were extended to include much of of interconnecting streams and lakes in the the area surrounding the Valley of Ten Thou­ Naknek River system. This system dominates sand Smokes, and the monument now encom­ the western half of the monument; it is here 2 passes 10,916 km. • Interesting accounts of the Bureau of Commercial Fisheries has a the discovery, description, and history of this cooperative arrangement with the National remote scenic area were given by Griggs Park Service to do biological research on (1922) and Cahalane (1959). sockeye salmon, Oncorhynchus nerka (Wal­ The monument has many prominent fea­ baum). This research began at Brooks Lake tures that influence the distribution of fishes. in 1940 and was expanded in 1 961 to include It is divided by the Aleutian Range along a all of the Naknek system. Some of the studies northeast-southwest axis so that the general have dealt with fishes associated with sockeye streamflow is to the sides of the Alaska Penin- salmon. We have been involved with various aspects ~ Fishery Biologist, Bureau of Commercial Fisheries of studies at Brooks Lake since 1957 and Biological Laboratory, Auke Bay, Alaska 99821. throughout the Naknek system since 1961. In 2 Graduate student, Department of Fisheries and Wild­ 1962, we began to collect fishes from other life, Oregon State University, Corvallis, Oreg. 97331. drainage systems in the monument.

1 CAPE DOUGLAS

0 25 50 s KILOMETERS

Figure I.--Alaska Peninsula and adjacent areas, showing location of Katmai National Monument.

No general treatment of the occurrence and mostly to data from reports 5 on commercial distribution of fishes inKatmaiNational Monu­ fishing in Shelikof Strait, These reports are ment has been attempted, although Cahalane3 concerned primarily with the species of salmon in an interim report of the National Park Serv­ that spawn in particular streams, ice Katmai Project mentioned a few species, Cahalane ( 1959) made a biological survey of and Greenbank4 reported on a sport fishery the plant and vertebrate animal life in the survey of the Naknek system, Several publi­ monument but did not attempt to include fishes, cations concerned principally with aspects of The present study attempts to fill this gap in the sockeye salmon biology and a few published general knowledg~ of fishes in the monument, studies concerned with other species within The purposes of this paper are (1) to report the Naknek system provide an important back­ the known occurrence and distribution of ground for the present study, Previous infor­ fresh-water fishes in Katmai National Monu­ mation concerning fishes in the monument ment and (2) to consider the zoogeographical other than in the Naknek system is confined implications of these findings,

3 Cahalane, Victor H. 1954. A biological survey of 5 Information on the Shelikof Strait ponion of the monu­ Katmai National Monument. In R. S. Luntey, Katmai Proj­ ment is included in the Management Reports, U.S. Bureau ect interim repon, Katmai National Monument, Alaska, of Fisheries and Bureau of Commercial Fisheries, Kodiak pp. 75-109. u.s. Dept. Interior, Nat. Park Serv. Island District, for 1924-59. These repons are on file at 4Greenbank, John. 1954. Spon fish survey, Katmai the Bureau of Commercial Fisheries Biological Labora­ National Monument. Manuscript on file, Bur. Commer. tory, Auke Bay. Reponsforthis area since 1959 are avail­ Fish. Biol. Lab., Auke Bay, Alaska, 30 pp. able from the Alaska Depanment of Fish and Game, Juneau.

2 I

I I I I I I I ~------'

<>~ ~ ~ I' t;:'l:-~ ~ "' ~ of~

s

5 0 5 10 15 20 25

KILOMETERS

Figure 2.--Katmai National Monument, showing principal fresh-water areas. Asterisks indicate generallocations where fish collections wete made in the King Salmon River system and along Shelikof Strait, 1962-64. Hundreds of collections have been made throughout the Naknek River system beginning originally at Brooks Lake in 1940. Kuliak Bay Lake, Kaflia Bay Lake, Devils Cove Lake, Hallo Glacier Lake,·Katmai Crater Lake, andKaguyakCraterLakeare not offi­ cial names. They are identified as such in this manuscript with nearby named geographic features for convenience only.

FRESH-WATER AREAS OF THE MONUMENT

The most important drainage feature of the are essentially isolated from other fresh monument is the arrangement of the Aleutian waters of the continent by the Aleutian Range. Range so that streams along the west side of the Alaska Peninsula flow into the Bering Sea BRISTOL BAY DRAINAGE through Bristol Bay and streams along the east side flow into the North Pacific Ocean The principal fresh-water areas in the through Shelikof Strait or Cook Inlet (fig. 1). monument that drain into Bristol Bay are in A second important feature is the number, the Naknek and King Salmon River systems. size, and relative complexity of drainage Of minor importance are the extreme head­ systems along the east and west sides of the waters of the Kejulik River in the Egegik monument (fig. Z). Although no major zoogeo­ River system, which lie just inside the south­ graphical barriers lie between stream sys­ ern boundary of the monument. These head­ terns draining into Bristol Bay and interior waters are small and of little significance to Alaska, streams flowing into Shelikof Strait the distribution of fish in the monument (fig. Z).

3 Those portions of the King Salmon system habitat is confined to the Bristol Bay area; that lie within the southwest corner of the it becomes a prominent feature of the tundra monument are Contact, Angle, and Takayoto west of the monument boundary. Creeks. Angle Creek is a tributary of Taka­ yoto Creek which joins Contact Creek to form the King Salmon River. Two small intercon­ SHELIKOF STRAIT DRAINAGE necting lakes, which Cahalane (1959) called Cozy Lakes, drain into Contact Creek. Fresh-water drainages in the Shelikof Strait The Naknek system has seven interconnect­ area are characterized by numerous short and ing lakes and is by far the largest and most isolated systems, many of which consist only complex drainage in the monument (fig. 2). of stream habitats or of streams and small, Murray and Hammersly, the uppermost lakes, insignificant lakes. Some drainage systems lie outs ide the north boundary of the monument; have lakes up to 4.8 km. long, but none have Coville, Grosvenor, Brooks, andidavainLakes large interconnecting lakes. Variety in stream all lie wholly in the monument; and the lower and lake habitat in Shelikof Strait drainages part of Naknek Lake is west of the monument. comes not from a single large complex sys­ All of the lakes are of glacial origin (Muller, tern but from numerous small and relatively 1952). Six of them essentially occupy single simple ones. The shoreline of the strait has basins, but Naknek Lake has six distinct gently curving bays and rugged fiords; about basins (fig. 2): Iliuk Arm, North Arm, North­ 20 distinct systems drain into the heads of west Arm, and South Bay are basins separated the major bays. These bays, from south to from each other by shallow, narrow connec­ north along the coast of the monument, are as tions; Bay of Islands is a distinct area at the follows: Kashvik, Katmai, Dakavak, Amalik, east end of North Arm; and the West End is Kinak, Miss ak, Kuliak, Kaflia, Kukak, Hallo, a shallow outwash plain of earlier glacial and Kaguyak (fig. 2). The streams that flow activity. into these bays are between 4.8 and 32.2 km. All of the lakes are dimictic oligotrophic long, and generally have steep gradients. Most lakes. 6 They are usually ice free from early of them are unnamed; notable exceptions are May until early December. Murray and Ham­ Katmai River and Alagogshak and Soluka mersly Lakes lie about 488 m. elevation and Creeks, which flow into Katmai Bay. Mageik are ice free 4 weeks less than the other lakes. and Martin Creeks flow into Katmai River. Surface water temperatures during the summer North of Hallo Bay the coastline becomes reach 12° C. in all lakes except the shallowest more regular and continues in a large sweep­ lake, Coville, where they reach 15° C. Thermal ing arc that includes the easternmost promon­ stratification is seldom strongly developed or tory of the Alaska Peninsula, Cape Douglas persistent during the summer. Oxygen is at or (fig. 1). Big and Swikshak Rivers (fig. 2) drain near saturation at all depths at all times of much of this section of the coastline. In the the year. Iliuk Arm is the only basin that northernmost part of the monument several receives drainage directly from glaciers and small streams flow north from glaciers on from the Valley of Ten Thousand Smokes; Mt. Douglas into Kamishak Bay of Cook Inlet it is extremely turbid; visibility is usually (fig. 1). less than 15 em. Coarse ash and pumice, principally from The largest streams in the Naknek system volcanic activity in 1912 (Griggs, 1922), is are American, Hardscrabble, Ikagluik, Margot, many meters deep over much of the terrain in Bay of Islands, and Headwater Creeks, and the southeast portion of the monument. Several Coville, Grosvenor, Savonoski, Rainbow, Ukak, streams in this area, including Katmai River, and Brooks Rivers. In general, these streams Soluka Creek, and unnamed streams that flow have midsummer flows in excess of 3 m. 3/sec. into the head of Kukak Bay, carry heavy silt Naknek River itself, which drains Naknek loads from the ash-laden watersheds. These Lake, lies outside the monument. streams have unstable beds and are exten­ A number of tundra ponds, beaver ponds, sively braided. Other large braided streams, and small tundra lakes occur along the west­ which have less ash and pumice on their ern margin of the monument. These bodies of watersheds but carry silt loads from glaciers, water are shallow, frequently contain sub­ are Swikshak River and two streams that flow merged and emergent aquatic vegetation, and into Hallo Bay. Not all streams along the occasionally have no surface connections with strait, however, are braided or carry heavy major stream systems. Often a series ofthese silt loads. Those flowing from Dakavak Lake, ponds. is connected by slow-flowing streams or from Kuliak Bay Lake, to and from Devils marshes that go nearly dry during periods of Cove Lake, Alagogshak Creek, and Big River low runoff. In the monument this aquatic all have fairly stable streambeds with moder­ ate gradients. 6 Hanman, Wilbur L., and Roben L. Burgner. 1968. Besides the principal terminal streams, Limnology of sockeye salmon-producing lakes in south­ short lateral streams drain the steep moun­ western Alaska. Manuscript on file, Bur. Commer. Fish. tain slopes along the sides of the fiordlike Biol. Lab., Auke Bay, Alaska, 51 pp. bays along Shelikof Strait. Many of these

4 streams are precipitous and have irregular largest--about 4.8 km. long and 1.0 km. wide. flows and unstable streambeds. In our limited Its watershed is heavily covered with pumice sampling in these habitats, principally in the and ash, which give a chalky blue-gray color Amalik and Kinak Bay areas, we found that to the water. The maximum depth is unknown, most of these streams had no established although gill nets were fished as deep as 21 m. ic hthyofauna. in 1962. Other smaller but generally clear The two largest streams along the Shelikof lakes where fish were collected include Kuliak Strait side of the monument are Katmai and Bay Lake, Kaflia Bay Lake, and Devils Cove Big Rivers. Griggs ( 1922) has described the Lake. Small hanging lakes with high waterfalls physical characteristics, some of them unique, occur at the heads of Amalik and Kinak Bays, of Katmai River. We collected fish on the east but we did not collect from them; we did col­ side of Katmai River about 3.2 km. above the lect, however, in the outlet stream from a mouth and from a clear-water tributary of the turbid glacier lake that forms along the face of river, Martin Creek. Big River, which is Hallo Glacier. about 32.2 km. long, arises to the west of Most lakes along Shelikof Strait had sum­ Kaguyak Crater and flows in a large sweeping mer surface temperatures similar to those arc into the northwest corner of Kaguyak Bay of lakes of the Naknek River system; Hallo (fig. 2). It has alternate stretches of pools and Glacier Lake was only 1° C. in mid-August riffles, numerous small oxbows and cutoff 1964. meanders, and a low-gradient, slow-flowing Crater lakes occur in the calderas ofKatmai lagoon section above the mouth. We collected and Kaguyak volcanoes. Although no attempt fish from the main stream and from small was made to collect fish in Katmai Crater flood plain cutoff ponds about 8.0 km. above Lake, one species was collected in Kaguyak the mouth. Crater Lake. Although they are generally small, the sev­ Notably lacking in Shelikof Strait drainages eral lakes on the Shelikof Strait side of the is any extensive interconnecting stream-lake monument provide a variety of lentic habitats complex similar to that in the Naknek system. for fishes in this area. Most of the lakes Except for common marine connections, Sheli­ appear to be of glacial origin and are rela­ kof Strait streams and lakes in the monument tively deep for their size. Dakavak Lake, are isolated from Bristol Bay drainages and which drains into the head of Dakavak Bay from each other by mountains of the Aleutian through a 3.2-km.-long outlet stream, is the Range.

METHODS AND EQUIPMENT

The intensity of fish collecting was much in Shelikof Strait drainages and 3 collections at greater in the Naknek River drainage than in 2 locations in the King Salmon River system either the King Salmon River or the Shelikof between 1962 and 1964. Locations of collections Strait drainage. The occurrence and distri­ from Shelikof Strait and King Salmon River bution of fishes in these two areas involve areas are shown in figure 2. fewer species, however, and appear less com­ The problems of moving personnel and plex than the Naknek system. The amount of equipment to certain areas in the monument fish collecting done in any one stream or lake were challenging. Small boats and amphibious varied greatly, depending on accessibility and aircraft were available for travel throughout the degree to which the location was involved the larger Naknek system lakes. Overland in research on sockeye salmon. No attempt treks were made on foot to small ponds and was made to document each collection from streams adjacent to larger waterways. We the Naknek system because hundreds of col­ used small single-engine float planes to reach lections in most of this system have been made the most remote areas, particularly along by the authors and many others engaged in Shelikof Strait. With aircraft, we used various salmon research. We collected from numerous sampling methods to collect in remote areas. smaller lakes, ponds, and streams in the Nak­ The procedure usually followed involved a nek system supplemental to salmon research 2- or 3-day air trip with a series of stops at to fill in gaps in the knowledge of fishes in collecting sites to set gill nets, small traps, the Naknek system. This report is the first or fyke nets and to make seine and hook and attempt to bring together knowledge gained line collections and general observations on from this sampling concerning the occur­ fishes at each site. A return flight tothe same rence of various species in specific lakes and area on the final day of the trip enabled us to streams. haul the nets and traps left overnight and to Because of the large size, the remoteness process the collections. The collector in re­ of the area, and the general similarity of the mote areas frequently wore a neoprene ex­ drainages along Shelikof Strait, no attempt was posure suit and searched for fishes by snor­ made to collect in all of the streams there. keling along the surface in shallow water. Ex­ We made about 25 collections at 13 locations posure suits were also used during the setting

5 and lifting of gill nets and other gear when contributed information on the occurrence and boats were not available. distribution of fishes considered in this report. We used several types of equipment to col­ Fish collections were processed by identi­ lect fishes. Beach seines and gill nets of fying and counting the number of each species various lengths and mesh sizes, tow nets collected. In some collections, length and 3.3 m. and 1 m. in diameter, small otter weights were recorded, and scales were col­ trawls, and small-meshed trap nets and fyke lected for age analysis. Lengths of fish, except nets were used routinely through most of the where otherwise noted, are reported as fork Naknek system. Much of this equipment was length. Field records from all Shelikof Strait also fished in other areas, although trawls and King Salmon River system collections, and tow nets required powerboats, available plus many of the Naknek system collections, only on the larger Naknek system lakes. Other are filed in the museum of the Bureau of Com­ sampling equipment included fish toxicants, mercial Fisheries Biological Laboratory at sport fishing gear, hand nets, fish spears, Auke Bay, Alaska. small bobbinet "habitat" seines, and small­ Identifications of fishes were made by the meshed collapsible nylon traps. Details on authors except where otherwise referenced. the use of much of this equipment have been The principal taxonomic keys used were by discussed elsewhere (Hartman, Heard, and Wilimovsky 9 and Clemens and Wilby (1961). Strickland/ Ellis,8 Heard, 1962; Merrell, Meristic counts follow Hubbs and Lagler 1964). ( 1958). Representative samples of all species Fish weirs or counting fences operated by collected in each of the principal drainages of personnel of the Bureau of Commercial Fish­ the monument have been deposited either in eries at various times in the Naknek system the museum of the Laboratory at Auke Bay or on Naknek River, Brooks River, American in the Oregon State University fish collection. Creek, Hardscrabble Creek, Hidden Creek, Nomenclature of fishes follows the American and West Creek, and along Shelikof Strait at Fisheries Society ( 1960) except where other­ the outlet of lower Kaflia Bay Lake have all wise noted.

ANNOTATED LIST OF SPECIES

The data on occurrence and distribution of FRESH-WATER FORMS fresh-water fishes in Katmai National Monu­ ment are presented in the following annotated Many of the fishes occurring in fresh-water list of 24 species from nine families known to areas of Katmai National Monument either are occur in the monument. Within the family or can be anadromous. A few forms that ap­ groups, six families- -Petromyzontidae, Os­ parently are confined to fresh water through­ meridae, Umbridae, Esocidae, Catostomidae, out their life in southwest Alaska occur in and Gadidae--have only one species; whereas estuarine environments in the Arctic (Walters, two families--Gasterosteidae and Cottidae-­ 1955). Only three families in the monument-­ have two species each; and one large family-­ Umbridae, Esocidae, and Catostomidae--are Salmonidae- -has 14 species. In addition to considered by Miller (1958) to be primary information on fresh-water forms, we provide fresh-water fishes. All 24 forms listed below, data on four euryhaline species, each from however, at least spawn and spend their juve­ a separate family, that occur in or near fresh nile stages in fresh waters of the monument. waters of the monument. Besides data on the occurrence and distribution of each species, Petromyzontidae the annotations include information on the life history and biology from various parts of the Arctic lamprey, Lampetra japonica (Mar­ monument. Such information generally is not tens).--Arctic lamprey were collected only in available from other sources. The quantity of the Naknek system, where fresh-water and this information varies greatly from species anadromous forms coexist. We collected ma­ to species--a result of circumstance, not ture, spawning, or spent individuals of the design. smaller fresh-water form in May and June from Coville, Grosvenor, and Brooks Rivers; from most tributaries of Brooks Lake; from 7 Hartman, Wilbur L., William R. Heard, and Charles two small tributaries of Naknek Lake, one in W. Strickland. 1962. Red salmon studies at Brooks Lake North Arm and one on the south shore of the Biological Field Station, 1961. Manuscript on file, Bur. lake near the monument boundary; and from Commer. Fish. Biol. Lab., Auke Bay, Alaska, 53 pp. a tributary of Coville Lake about 1.6 km. 8 Ellis, Robert J. 1963. The abundance and distribution north of the lake outlet. Only one specimen of of juvenile red salmon and associated species in lakes of the Naknek River system and Karluk Lake. Manuscript on 9 Wilimovsky, Norman J. 1958. Provisionalkeystothe file, Bur. Commer. Fish. Biol. Lab., Auke Bay, Alaska, fishes of Alaska. Manuscript on file, Bur. Commer. Fish. 80 pp. Biol. Lab., Auke Bay, Alaska, 113 pp. 6 the larger anadromous form was taken inside Least cisco, Coregonus sardinella Valen­ the monument--at the mouth of Brooks River, ciennes.--Least cisco were collected in the but several were collected in the Naknek River Naknek and King Salmon systems. They oc­ just outside the monument, curred in Coville, Grosvenor, and Naknek Immature adult Arctic lamprey were com­ Lakes but not in Brooks Lake, although they monly collected during the summer in tow nets were occasionally caught in lower Brooks in limnetic areas of Coville, Grosvenor, and River below the falls. Greenbank (see foot­ Brooks Lakes and of Iliuk Arm, South Bay, note 4) collected this species from a small and North Arm of Naknek Lake, They were lake east of the head of Bay of Islands which also caught in fyke nets (late summer and fall) he called "Jo-Jo" Lake, At one time Jo-Jo from Coville, Grosvenor, and Brooks Rivers. Lake undoubtedly was continuous with the Bay Actively feeding parasitic-phase lampreys of Islands, and high water levels may still con­ were observed attached to sockeye salmon, nect the two through a series of marshes; Oncorhynchus nerka (Walbaum); pygmy white­ this lake may also drain through a series fish, Prosopium coulterii~ 0 (Eigenmann and of marshes into the Savonoski River, In the Eigenrnann); rainbow trout, Salmo gairdnerii King Salmon River system least cisco were Richardson; and three spine stickleback, Gas­ collected in lower Cozy Lake, terosteus aculeatus Linnaeus, Parasitism by Most least cisco were caught in passive the lamprey, however, is not known to threaten sampling gear such as trap nets and gill nets, any fish population in the Naknek system, The species seemed to escape readily from Morphologically, the parasitic Arctic lam­ active sampling gear such as seines, tow nets, prey is almost identical with the nonparasitic and trawls, especially in clearer waters, American brook lamprey, Lampetra lamotteii of eastern North America, Lampetra lamotteii Pink salmon, Oncorhynchus gorbuscha (Wal­ has been reported in Alaska (Wilimovsky, baum).-- Pink salmon were collected in the 1954; Hubbs and Lagler, 1958), and it was ini­ Naknek and Shelikof Strait systems. They oc­ tially believed to occur in the Naknek system, curred in the monument portion of the Naknek Available evidence suggests, however, that all system in small numbers: 100 passed through lamprey in the Naknek system are parasitic Brooks River weir into Brooks Lake in 1962, (Heard, 1966), The report by Cahalane (see and about 200 spawned in Brooks River itself. footnote 1) that Lampetra tridentata (Gairdner) Spawning pink salmon were observed in Up-a­ occurred in Grosvenor Lake was apparently Tree, One Shot, Hidden, and Headwater Creeks based on hearsay and cannot be accepted, and in Coville River between 1960 and 1964. Although accurate counts are few and some Salmonidae locality records are not specific, pink salmon are the most abundant and widely distributed Humpback whitefish, Coregonus pidschian species of salmon on the Shelikof Strait side (Gmelin).--Humpback whitefish were collected of the monument, They have been reported only in the Naknek system, where they occur from every bay along the coastline and prob­ in most of the principal lakes, They were col­ ably spawn in streams at the head of most of lected in Iliuk Arm, South Bay, North Arm, these bays, Counts of adults through the Kaflia and West End of Naknek Lake and in Coville Bay weir in 1929 and 1934 were 291 and 375, and Grosvenor Lakes but not in Brooks Lake, respectively. In 1946, 100,000 pink salmon Apparently, the greatest density of humpback were caught in Swikshak Lagoon, whitefish in the monument is in Coville Lake, We collected and observed pink salmon in although numbers are substantial also in South the Shelikof Strait area of the monument in Bay basin of Naknek Lake. Ripe humpback 1962-64. We saw spawning runs of several whitefish were caught in the autumn in South thousand fish in Alagogshak Creek, Big River, Bay near the mouth of Brooks River but were and the stream that flows from Dakavak Lake never observed spawning in the river, Observ­ into Dakavak Bay, and smaller runs in Katmai ers in a low-flying aircraft saw thousands of River and the principal terminal streams that fish on riffles in the lowerportionofAmerican flow into Amalik and Kuliak Bays and Devils Creek on October 27, 1963. The fish were Cove of Kukak Bay. Our observations were believed to be humpback whitefish spawners made during limited surveys and should only from Coville Lake. The taxonomic and phyletic be considered as records of occurrence and relation of this fish and the lake whitefish, not as indicators of total production, We were Coregonus clupeaformis (Mitchill) in Alaska unable to observe streams in several of the is not clear (Walters, 1955; Lindsey, 1962), areas, including Kashvik Bay, the main arm of Kukak Bay, and all of the area east and ~o Although the American Fisheries Society checklist north of Big River. uses the specific taxon spelling coulter!, the latest judg­ Spawning by pink salmon in Katmai River ment of the International Code of Zoological Nomenclature was somewhat unexpected because the unstable holds that original spellings of "ii" on patronyms must be stream bottom of this heavily braided river is retained. This ruling also applies to Lampetra lamotteii composed primarily of water-soaked pumice, (LeSueur) and Salmo gairdnerii Richardson. Nevertheless, adult pink salmon spawned in

7 small discontinuous pockets of gravel along spawning areas of sockeye salmon in the Nak­ the edge of a small bluff on the east bank of nek system are the American, Hardscrabble, the river about 3.2 km. above the mouth. Margot, Headwater, and Bay of Islands Creeks Surface velocity was 0.3-0. 7 m./ sec. in this and Grosvenor, Brooks, and Naknek Rivers. area. Many smaller streams and some lake beaches Chum salmon, Oncorhynchus keta (Wal­ also support spawning populations, Juvenile sockeye salmon occurred in all areas of the baum).--Chum salmon were collected in the larger Naknek system lakes except Idavain and Shelikof Strait systems, They are Nak~~k and Murray Lakes, Sport fishing for adults is the least abundant salmon in the Naknek sys­ popular in Brooks River, tern, and the only information on their occur­ No fresh-water (kokanee) populations of rence and distribution is that some adults sockeye salmon are known to occur sympatri­ (usually fewer than 10) pass through the cally with anadromous populations in the Brooks River weir each year; a few have been monument. We discovered kokanee salmon in reported in American Creek. We did not see two lakes that drain into Shelikof Strait, and chum salmon in the Shelikof Strait area of the Greenbank (see footnote 4) reported them in monument, but they were reported in Big Jo-Jo Lake, River and in Katmai, Kuliak, Kaflia, Kukak, We saw several schools of young-of-the­ Hallo, and Swikshak Bays, either in the early year sockeye salmon (age 0) in lower Cozy U.S. Bureau of Fisheries management reports Lake of the King Salmon River system and or by biologists of Alaska Department of Fish and Game, collected about 100 fish on June 29, 1962. Apparently the largest run of anadromous Coho salmon, Oncorhynchus kisutch (Wal­ sockeye salmon in the Shelikof Strait area of baum).--Coho salmon, which are widely dis­ the monument is into two lakes at the head of tributed in the monument, were collected in Kaflia Bay, The lower lake has slightly brack­ the Naknek, King Salmon, and Shelikof Strait ish water either from seepage or possibly systems. Most of the information on this spe­ from flooding on extremely high tides, The cies is based on samples of juveniles taken in upper lake is 15 to 23 m, higher than the the summer because the adult salmon do not lower, and the two are connected by a steep appear until fall after much of the work on boulder-strewn stream about 305 m, long, sockeye salmon has been discontinued, Within The U,S, Bureau of Fisheries had a weir for the Naknek system, annual counts of adult coho counting adult salmon, presumably at the out­ salmon through Brooks River weir in 1957-62 let of the lower lake at Kaflia Bay in 1929-32, ranged from 194 to 1,844 (average 620). We and 1934; escapements of spawning sockeye collected juveniles or adult coho salmon in the salmon into the two lakes in these years aver­ Naknek system in Coville, Grosvenor, and aged about 12,500 (see footnote 5), The esti­ Brooks Rivers; Coville, Grosvenor, and Brooks mated escapement of sockeye salmon to the Lakes, and all basins of Naknek Lake; Hard­ two lakes in 1939 was 100,000 fish. OnJu1y 12, scrabble, Margot, Headwater, Hidden, One 1963, we saw about 2,000 adult sockeye salmon Shot, and West Creeks; and unnamed creeks off the mouth of the interconnecting stream from Coville, Grosvenor, and lower Naknek between the two lakes, We saw no adult salmon Lakes, The only coho salmon collected from in the stream proper, but collected one juve­ the King Salmon River system was a 62-mm, nile (about 70 mm. long) from a pool just juvenile from lower Cozy Lake on June 29, below the upper lake, so at least some adults 1962, The only area where we saw coho salmon ascend this precipitous stream, along Shelikof Strait was in Big River; 71 ju­ Sockeye salmon runs of unknown size were veniles (35-68 mm.) were seined there on reported in the early management reports and August 15, 1964. Accounts of good runs of by local residents of the Kodiak Island area to adult coho salmon in Kaflia Bay lakes and occur in a small lake at the head of the west streams are given in some early management arm of Kuliak Bay (known locally as Halferty reports. Although coho salmon have been Bay), We saw no adults on two collecting trips reported in Kashvik, Katmai, Kuliak, Kukak, to the Kuliak Bay Lake in August 1964, but Rallo, and Kaguyak Bays and in Swikshak collected several juveniles (age 0 and l; from Lagoon, they have not been reported from 52 to 92 mm.). The outlet stream had no bar­ streams that enter those waters. riers. or obstructions that would prevent anad­ romous adults from entering the lake. Sockeye salmon, Oncorhynchus nerka (Wal­ We collected juvenilesockeyesalmonin1962 baum).--Sockeye salmon, the most evidentand from Dakavak Lake and a small lake above widely distributed species in the monument, the Devils Cove portion of Kukak Bay. No rec­ were collected in the Naknek, King Salmon, ords of adult sockeye salmon have been found and Shelikof Strait systems, in the commercial statistics for the bays The Naknek system produces a large part of into which these lakes 4_rain, Moreover, the the valuable Bristol Bay catch of this species; streams flowing from these lakes have bar­ annual escapements to the system have varied riers, and we believe that the populations in between 0,3 and 2.2 million fish, The major both lakes are landlocked kokanee salmon.

8 The outlet stream at Dakavak Lake is now precipitous, but most of the lower portion is subterranean; it flows or seeps through a sluggish and forms a lagoon just before it 60- to 100-m. pumice and volcanic ash dike joins the Swikshak River. Both portions are across the outlet end of the lake. The origin of about 4 km. long. Sockeye salmon spawn in the dike is unknown, although it likely devel­ an area about 0.8 km. long between the two oped, at least in part, from the extremely sections where the gradient first increases heavy ash deposits that followed the 1912 sufficiently to provide a silt-free gravel sub­ volcanic activity in the Dakavak Lake area. strate.ll Upstream, this gradient becomes too Possibly ash and rock slides in the narrow steep for a stable stream bottom. An estimated outlet arm of the lake, combined with large 2, 700 sockeye salmon spawned in this stream volumes of floating pumice on the lake sur­ on August 20, 1959.l2 face, formed the dike. During its formation, Chinook salmon, Oncorhynchus tshawytscha and possibly later, surface water flowed over (Walbaum).--Chinook salmon were collected in the top of the dike, as evidenced by high-water the Naknek and King Salmon River systems. beach lines on the dike and margins of the They occurred only in small numbers within lake. The top of the dike, about 25 feet above the current lake level, coincides with the top the monument portion of the Naknek system. Adults were observed in Brooks River and in of a denuded high-water zone around the lake Headwater, Bay of Islands, and American margin (Cahalane 1 s [ 1959] frontispiece, photo­ Creeks between 1960 and 1964, Since 1940 graphed at Dakavak Lake on July 28, 1953, an average of 15 adults per year have passed clearly shows this denuded zone.) Undoubtedly through Brooks River weir into Brooks Lake. this dike has prevented anadromous fishes from going into the lake for several years. The only juveniles observed within the monu­ ment portion of the Naknek system were seined Nineteen sockeye salmon collected from the from Hidden Creek Bay in Brooks Lake in lake on June 30 and July 1, 1962, were 61 to 1961. Large numbers of chinook salmon occur 160 mm. long and averaged 125 mm. The larger of these (over 140 mm.) were in their outside the boundary of the monument in the fourth year of life and had maturing gonads. lower portion of the Naknek system. Records We believe these fish would have spawned in from a weir on the Naknek River at the head the fall. This population may have become of tidewater indicate that about 5,000 adult landlocked during the eruption of 1912. chinook salm_On...£aS§_~d through the weir-annu­ Anadromous sockeye salmon are probably ~from 1953 to 1957; most of these spawned excluded from the lake above Devils Cove in in Naknek River below the outlet of Naknek Lake. Substantial runs of chinook salmon Kukak Bay by a 5- to 6-m. waterfall on the spawn in Big Creek and King Salmon Creeks outlet stream. We caught 26 possible kokanee downstream from the Naknek River weir salmon, 28 to 139 mm. (average 105 mm.), by site.l3 We did not observe adult chinook setting a gill net overnight and by beach sein­ salmon in the King Salmon River system ing in this lake on July 1, 1962. One fish, within the monument, but we collected juve­ 136 mm. long, was in its fifth year of life; niles (34-51 mm.) from a small slough along seven others, over 118 mm. long, were in Takayoto Creek in August 1964. their fourth year. These older fish had matur­ ing gonads and probably would have spawned Pygmy whitefish, Pros opium coulterii in the fall. (Eigenmann and Eigenmann).--Pygmy white­ Perhaps the most unusual account of sockeye fish were collected only in the Naknek system. salmon in Katmai National Monument, one These small whitefish were collected from where this salmon apparently did not follow Iliuk Arm, South Bay, North Arm, and Bay of the usual pattern of using a lake for a juvenile Islands in Naknek Lake; throughout Grosvenor rearing area, was given by Griggs ( 1922, and Brooks Lakes; from the eastern end of p. 161). He quoted c. H. Gilbert who suggested Coville Lake; and from near the outlet in that 5 -year-old adult anadromous sockeye Hammersly Lake. They were particularly salmon collected in 1917 from a small un­ abundant in Brooks Lake and South Bay of named tributary of the Katmai River had spent Naknek Lake. The fish were collected in 2 years in fresh water, possibly in the tribu­ shallow areas and in the deepest known fresh­ tary or in Katmai River itself in the years water area in the monument (171 m., Iliuk immediately after the 1912 eruption. We had Arm). Large numbers of pygmy whitefish hoped to recheck this stream to learn if sock­ eye salmon were still there but were unable ll Personal communication, W. H. Noerenberg, Deputy to determine the stream's exact location. Commissioner for Commercial Fisheries, Alaska Depart­ Another group of anadromous sockeye salmon ment of Fish and Game, January 24, 1966. that does not have a lake nursery area occurs l2 Noerenberg, Wallace H. 1959, Stream surveys in the in the Swikshak River sys'tem. These fish Kodiak area, 1959. Univ. Wash., Seattle, Fish. Res. Inst., spawn in a small stream that runs somewhat Circ. 111, 40 pp. parallel with, and to the east of, Swikshak lJ King Salmon Creek, a tributary of Naknek River west River and enters the river about 1.6 km. above of King Salmon, should not be confused with the King its mouth. In its upper portion this stream is Salmon River drainage.

9 from South Bay move into lower Brooks River Lake are brilliantly colored and are referred in the summer and feed, primarily on imma­ to by residents of the area as "golden trout." ture insects, These fish spawn in Brooks River in November and December, The biology Dolly Varden, Sa 1 v e 1 in us malma (Wal­ of pygmy whitefish in the Naknek system is baum).-- Two types of Dolly Varden occur reported by Heard and Hartman ( 1966), within the monument, one in the Naknek system and one in Shelikof Strait drainages. Dolly Round whitefish, Prosopium cylindraceum Varden in the Naknek system generally live in (Pallas).--Round whitefish were collected in streams and are small, usually less than the Naknek and King Salmon systems, They 25 em, long. McPhail (1961), who recently occurred in Naknek, Coville, Grosvenor, and evaluated the Dolly Varden-Arctic char com­ Brooks Lakes; American and Headwater plex in North America, recognized this small Creeks; Hidden and West Creeks (tributaries stream-dwelling Dolly Varden as a northern of Brooks Lake); and Coville, Grosvenor, and form, distinct from the more widely distributed Brooks Rivers, In the King Salmon River southern form, The Aleutian Range on the drainage, they occurred in lower Cozy Lake, Alaska Peninsula apparently represents the Round whitefish spawn in the autumn, ap­ dividing line for these forms, McPhail's mate­ parently in streams similar to those inhabited rial from Karluk and Fraser Lakes on Kodiak by mountain whitefish, P, williamsoni (Girard), Island was characteristic of the southern form, in Montana (Brown, 1952), About 200 adults in and material from Brooks Lake was charac­ breeding condition were seined from South teristic of the northern form. Bay of Naknek Lake off the mouth of Brooks In the Naknek system we collected Dolly River on November 7, 1962. Both sexes were Varden from Grosvenor and Brooks Rivers, tuberculate over much of the body and had and from Brooks Lake, Up-a-Tree, One Shot, brightly colored orange ventral fins and a bold and Hidden Creeks, Greenbank (see footnote 4) longitudinal orange stripe along the body, reported that specimens he collected from Tony Malone, a trapper from Naknek, Alaska, Idavain Lake were Dolly Varden, but McPhail reported that whitefish 12 to 16 inches (30 to (1960), who made gill raker counts of Green­ 41 em,) long were abundant in late autumn in bank's specimens, suggested that the fish were Headwater Creek above the monument boun­ Arctic char, dary, We believe these fish were round white­ The southern form of Dolly Varden, which fish because only that species and pygmy is widely distributed in streams and lakes whitefish occur in Brooks Lake or its tribu­ along the Shelikof Strait area ofthe monument, taries, and pygmy whitefish never grow longer is distinguished from the northern form prin­ than about 20 em, cipally by fewer gill rakers and fewer verte­ brae; also, the southern form lives in lakes Rainbow trout, Salmo gairdnerii Richard­ and streams and may be anadromous or land­ s on.--Rainbow trout were collected only in the locked (McPhail, 1961), We collected speci­ Naknek system, We found them in Coville, mens of the southern form in lower Kaflia Grosvenor, and Brooks Lakes; South Bay and Bay Lake and in inlet and outlet streams of North Arm of Naknek Lake; Coville, Gros­ this lake; in Dakavak Lake, Kuliak Bay Lake, venor, and Brooks Rivers; American, Head­ Devils Cove Lake, and the outlet stream of water, West, and Margot Creeks; and one this lake below the high waterfalls; in a small unnamed creek in South Bay at the base of pond on Martin Creek (tributary of Katmai Dumpling Mountain, We have seen adult rain­ River); in Big River; and in Kaguyak Crater bow trout spawning in Brooks River in May Lake, An archaelogical survey team from the and have collected recently emerged fry in University of Oregon collected this fish in quiet water along the river in August, Green­ 1963 from a small stream near the old village bank (see footnote 4) presented age and growth of Kukak. Counts of gill rakers on the lower data for rainbow trout caught by sports fisher­ limb of the first gill arch of Dolly Varden that men in Brooks River, A moderately heavy we collected from various areas along Sheli­ sport fishery for this fish also occurs in kof Strait agree with McPhail's (1961) counts Naknek River, We do not know if steelhead for Dolly Varden from Karluk and Fraser occur in the Naknek system, Lakes on nearby Kodiak Island (table l), Arctic char from Idavain Lake were air­ Arctic char, Salvelinus alpinus (Linnaeus).-­ dropped into Kaguyak Crater Lake in 1956;~ 4 W e collected Arctic char from Iliuk Arm, but none were observed in our collections, South Bay, North Arm, and Northwest Arm of We collected nine specimens of char in Kagu­ Naknek Lake; Coville, Grosvenor, Brooks and yak Crater Lake in August 1964; gill raker Idavain Lakes; and American and Margot counts of the specimens, however, indicate Creeks, There are limited sport fisheries for they were Dolly Varden, distinct from Idavain Arctic char off the mouth of Margot Creek, in s orne unnamed creeks along the north shoreline ~ 4 E. Siler, a guide and bush pilot at King Salmon, de­ of the North and Northwest Arms, and in Ida­ scribed this 1956 capture, transpon, and air drop of vain Lake, Spawning Arctic char from Idavain about 20 "golden trout" to us at Brooks Lake in 1964. 10 Table 1.--Gill rakers in Arctic char and Dolly Varden from various parts of Katmai National Monument and in Dolly Varden from Karluk and Fraser Lakes on Kodiak Island

Gill rakers, lower half of first left gill arch I Species Fish in sample I I Range I Mean

Number Number Number Salvelinus alpinus Brooks Lake!! 1 14

Idavain LakJI 4 14-16 14.7 Idavain Lake 20 13-16 14.2 Salvelinus malma Southern fonn Dakavak Lake 15 9-11 10.2 Devils Cove Lake 12 10-12 11.1 Fraser Lake!! 38 8-12 10 Kaguyak Crater Lake 9 10-11 10.8 Karluk Lake!/ 33 8-12 11 Kuliak Bay Lake 18 9-12 10.4 Northern fonn Brooks Lake!!

(tributaries) 20 11-14 12.6 !I Data from McPhail (1961).

Lake Arctic char (table 1), Kaguyak Crater occur in Hammersly and Murray Lakes, Al­ Lake now has no outlet, although a low point though the species is found primarily in lakes, in the northeast rim of the crater suggests occasionally they are seen in the connecting that a stream may have flowed at an earlier rivers between lakes, such as Brooks and date when water levels in the crater were Coville Rivers, Grosvenor Lake has a limited higher (see Cahalane, 1959, plate 2, fig. 1), sport fishery for lake trout just below the out­ Sequential beach ridges attest to the former let of Coville River, where a concentration of occurrence of higher water levels inside the fish remains at least from ice breakup in the crater, Dolly Varden live in the surrounding spring until mid-September, This concentra­ Big River drainage, and apparently dispersed tion of lake trout is undoubtedly related to the into this unusual habitat, food that passes down the inlet stream, Lake trout have been observed at Colville River Lake trout, Salvelinus namaycush (Wal­ preying on migrating juvenile sockeye salmon, baum).--Lake trout were collected only in the They also prey heavily on sticklebacks and Naknek system. We collected them in almost pond smelt and occasionally on small rodents, all areas of Naknek, Coville, Grosvenor, and Greenbank (see footnote 4) presented data on Brooks Lakes but not in any of the smaller the average size of various age groups of lake lakes within the monument, although they trout from the Naknek system, and Lindsey

11 ( 1964) discussed the coexistence of lake trout fish in large lakes of the Naknek system seem and anadromous parasitic Arctic lamprey in to be associated with submerged aquatic vege­ the Naknek system, tation, Arctic grayling, Thymallus arcticus (Pal­ Esocidae las).--Arctic grayling were collected only in the Naknek River system.~' We collected them Northern pike, ~ ~ Linnaeus.-­ in Brooks River; Bay of Islands, Headwater, Northern pike were prominent in many small Hidden, and West Creeks; Brooks Lake; and lakes and ponds in the Naknek portion of the South Bay and North Arm of Naknek Lake, monument, They also live in the large lakes Fish in the lakes were usually taken near one and streams but usually in areas of relatively of the streams, Although their distribution shallow, quiet water with aquatic vegetation, was spotty, grayling were seasonally abundant Four such areas were Northwest Arm and Bay in certain streams, particularly Brooks River, of Islands in Naknek Lake, the Grosvenor One grayling was seined from the tnorainal River Lagoon just below the outlet of Gros­ cut between Iliuk Arm and South Bay in 1962, venor Lake, and backwater lagoons along lower about 4,8 km, from the mouth of Brooks River. American Creek, Northern pike were also Grayling spawn in Brooks River in May and collected in Iliuk Arm, South Bay, and North early June; we collected fry {19-22 mm, long) Arm of Naknek Lake; Coville, Grosvenor, and from the river on July 11, 1963, Sport fishing Brooks Lakes; Coville, Grosvenor, and Brooks for grayling is common in Brooks River (see Rivers; and West and Hidden Creeks, Smaller footnote 4), lakes known to have pike populations include beaver ponds at the head of Hidden Creek, a Osmeridae series of lakes and beaver ponds atthehead of West Creek (Greenbank [see footnote 4] re­ Pond smelt, Hypomesus olidus (Pallas).-­ ferred to the larger of these as "Pike" Lake), Pond smelt were collected in Coville and Gros­ Jo-Jo Lake, and a lake in the Savonoski Valley venor Lakes; North Arm, South Bay, and Iliuk about 3 km. east of the Ukak River that Green­ Arm of Naknek Lake; and Coville River, bank called "Murial" Lake, We did not sample Although we did not find them in Brooks Lake, Murial Lake, but Greenbank did, we collected them in West Creek and Brooks Although northern pike now occur only rarely River just below the outlet of Brooks Lake, so in the Brooks River area, they may have been they probably live in the lake, Pond smelt are abundant there in recent historical times, apparently more widely distributed and more University of Oregon archaeologists have col­ abundant in Coville Lake than in other parts of lected considerable quantities of fish remains the Naknek system, in aboriginal occupation sites near Brooks River Falls, Considerable numbers of pike Umbridae scales were found in the remains, most of which were from adult salmon, Radiocarbon Alaska blackfish, Dallia pectoralis Bean.-­ dates and stratigraphic series of volcanic Blackfish were collected in Coville, Grosvenor, ash deposits indicate that most of these fish and Brooks Lakes; Bay of Islands and North­ remains were deposited from A.D. 1500 to west Arm of Naknek Lake; and West Creek 1800, ~ 6 Several swamp and marsh areas along and the ponds at the head of West Creek, the lower 400 m, of Brooks River now have Blackfish usually inhabit shallow tundra lakes water in them only for a brief period in late and ponds (Walters, 1955) or sluggish marshes summer and early fall when the water level at along streams (Blackett, 1962), Normally they Naknek Lake is highest. A few centuries ago are not considered to inhabit large lakes or when the mean water levels of Naknek Lake deep water, In Brooks Lake, where they ap­ were high, the pike habitat in the present parently are more abundant than anywhere marshes and swamps was probably more sta­ else in the Naknek system, blackfish are com­ ble. Under such conditions pike were probably monly caught in water 15 to 20 m, deep; at abundant in the area, this depth they are still within the range of submerged vegetation, particularly Chara and Catostomidae Nitella. A trap net set 8 m, deep on the bottom of Brooks Lake near the outlet from July 5 to Longnose sucker, Catostomus catostomus September 9, 1963, took 583 blackfish. In gen­ (Forster).--Longnose suckers were collected eral, the occurrence and distribution ofblack- in Coville and Grosvenor Lakes; Iliuk Arm, South Bay, and North Arm of Naknek Lake; ~.5 We received an unconfirmed reponfrom sponfisher­ American Creek; and Coville, Grosvenor, men in 1966 that they had caught grayling in Contact Creek in the headwaters of the King Salmon River system. ~ 6 Cressman, L. S., and D. E. Dumond. 1962. Re­ Grayling are known to be common in the Ugashik River search on Northwest prehistory, prehistory in the Naknek system fanher out on the Alaska Peninsula (fig. 1), so it drainage, southwestern Alaska. Univ. Oreg., Eugene, Dep. is likely they also live in the King Salmon River system. Anthropol., 54 pp.

12 Brooks, and Ukak Rivers. We did not collect fish were collected in fresh water, except for them in Brooks Lake but found them in the lower Kaflia Bay Lake, which is slightly lower portion of Brooks River below the falls. brackish. The fresh-water areas were only In the summer, many suckers occupy lagoon a short and readily accessible distance from areas in lower Brooks River and in Coville salt water. and Grosvenor Rivers, apparently to feed. Lateral plates on adult threespine stickle­ We presume that they spawn in early spring, backs from Shelikof Strait drainages were because ripe females were collected just off strongly developed, whereas those on Naknek the mouth of Brooks River in May and early system specimens were weakly developed or June. Postlarval fry were taken in the Coville absent. River Lagoon in July. Ninespine stickleback, Pungitius pungitius Gadidae (Linnaeus).--Ninespine sticklebacks were collected in the Naknek, King Salmon, and Burbot, Lota Iota (Linnaeus).--Adult burbot Shelikof Strait systems. They live through­ were collected from Iliuk Arm, South Bay, out the Naknek system in the same lake and North Arm, and Bay of Islands of Naknek stream areas as threespine sticklebacks, ex­ Lake. One larval fish believed to be a burbot cept for Hardscrabble Creek and Savonoski was taken in a plankton tow in Coville Lake, River, where only ninespine sticklebacks were although no adults were taken in Coville or collected. In the large lakes, the ninespine Grosvenor Lakes. No burbot was collected in stickleback seems to be more of a benthic Brooks Lake. Although no detailed data or fish than its relative, although both species specimens are available, there is a report~7 occur in almost all lake habitats. One notable of dead burbot on the upstream face of Brooks difference in the distribution of the two spe­ weir in the late 1940's. By far the most ex­ cies of sticklebacks in the Naknek portion of tensive fish collecting in any area of Katmai the monument is that only ninespine stickle­ National Monument during the past 10 years backs live in many of the shallow bog lakes, has been in Brooks Lake, and we doubt that tundra ponds, and beaver ponds. Examples of burbot now live in this lake. this distribution include the lakes and beaver ponds above West Creek, beaver ponds at the Gasterosteidae head of Hidden Creek, and Nystrom Lake--a small shallow lake near the mouth of Brooks Threespine stickleback, Gasterosteus acu­ River. Bond and Becker~ 8 also noted this pe­ leatus Linnaeus.--Threespine sticklebacks culiarity in the Iliamna Lake area. Robert and were collected in the Naknek and Shelikof Alice Dewey, resident biologists at the BCF Strait systems. Within the Naknek system they Brooks Lake Laboratory, found that ninespine were collected from Coville, Grosvenor, and sticklebacks were more tolerant to warmer Brooks Lakes; all basins of Naknek Lake; aquarium water than threespine sticklebacks. Coville, Grosvenor, and Brooks Rivers; Head­ Perhaps high summer temperatures in shallow water, Hidden, West, Margot, and unnamed ponds prevent establishment of stable popu­ creeks draining into North Arm; and Gros­ lations of threespine sticklebacks. In the King venor and lower Naknek Lakes. Large num­ Salmon River system, ninespine sticklebacks bers of threespine sticklebacks move from were collected in a slough of Takayoto Creek Grosvenor Lake and South Bay into lagoon and in lower Cozy Lake, where they were very areas in Coville and Brooks Rivers in early abundant. Only two specimens were collected summer to spawn. We saw thousands of fry in from the Shelikof Strait area of the monument; these lagoons in late summer. The greatest these were both taken in Big River about concentrations ofthreespine sticklebacks in the 4.8 km. above salt water. Naknek system apparently are in the west end of Naknek Lake and in Northwest Arm (see Cottidae footnote 8). In Shelikof Strait drainages, we collected Coastrange sculpin, Cottus aleuticus Gil­ three spine sticklebacks inKatmai River, lower bert.--Coastrange sculpins were collected in Kaflia Bay Lake, and Big River. Adults and the Naknek, King Salmon, and Shelikof Strait fry were collected in July or August; ap­ systems. In the Naknek system they were col­ parently all of the populations were ana­ lected in Coville and Brooks Lakes; South Bay dromous because the extent of lateral plate of Naknek Lake; Grosvenor, Brooks, and Ukak development on the adults suggested that they Rivers; Hidden, One Shot, and West Creeks; were from estuarine or marine environments and unnamed creeks in Grosvenor and Naknek (see Clemens and Wilby, 1961, p. 353). The Lakes. In the King Salmon system they were

~ 7 George J. Eicher, Ponland General Electric Com­ ~ 8 Bond, C. E., and C. D. Becker. 1963. Key to the pany, in conversation with W. L. HanmanatBrooks Lake, fishes of the Kvichak River system. Univ. Wash., Seattle, June 1965. Fish. Res. Inst., Circ. 189, 9 pp.

13 collected in lower Cozy Lake, In the Shelikof Osmeridae Strait systems they were collected in Big Arctic smelt, Osmerus dentex Steindach- River, in a tributary of Dakavak Lake, and in .!l!tl'·--Arctic smelt move into Naknek River outlet streams of lower Kaflia Bay and Devils from Bristol Bay in late winter and early Cove Lakes, spring on their spawning migrations, They Slimy sculpin, Cottus cognatus Richardson.-­ were collected as far upstream as the Naknek Rapids, a fast-water section of Naknek River Slimy sculpins were collected in the Naknek about midway between the outlet of Naknek and King Salmon systems, In the Naknek sys­ Lake and the town of King Salmon, Some of tern, they were taken in Coville, Grosvenor, these smelt may move into Naknek Lake near Brooks, and Idavain Lakes; all basins of Nak­ the outlet of the river, nek Lake; Coville, Grosvenor, Brooks, Savo­ noski, and Ukak Rivers; and American, Mar­ got, Headwater, Hidden, One Shot, and West Gadidae Creeks, This sculpin also occurs in Ham­ Pacific cod, Gadus macrocephalus Tile­ mersly Lake, In the King Salmon system, sius.--Greenbank (see footnote 4) reported slimy sculpins were collected in Takayoto that a Pacific cod was taken by an angler in Creek, Naknek River just below the outlet of Naknek Both species of cottids in the monument Lake, He speculated that cod might occasionally spawn in the spring, although their spawning stray into the lake, According to Gunter ( 1956) sites and behavior have not been observed, Greenbank's account was the first euryhaline Although the two species occur sympatrically record of this species. in streams and lakes of the Naknek system, collections from Brooks Lake and its tribu­ taries suggest the coastrange sculpin is usu­ Cottidae ally the dominant stream form and the slimy Pacific staghorn sculpin, Leptocottus arma­ sculpin is usually prevalent in lakes. Slimy tus Girard.-- We collected 10 staghorn scul­ sculpin was the only cottid collected from pins (14-65 mm, total length) from Katmai depths greater than 30 m. in any lake. River 3,2 km, above the stream mouth on EURYHALINE FORMS August 21, 1964,

At least four marine fishes occur in fresh­ Pleuronectidae water portions of streams that lie within the Starry flounder, Platichthys stellatus (Pal­ monurr.ent or flow from the monument, We las).--Several starry flounders have been collected specims of three of the species be­ taken in Naknek River as far upstream as tween 1962 and 1964, and Greenbank (see foot­ Naknek Rapids, and C. J. DiCostanzo~ 9 re­ note 4) reported the fourth, These fishes were ported that he has observed this fish in Nak­ collected upstream from the upper zones of nek River at the outlet of Naknek Lake. We tidal influence, and although no corroborative collected eight small starry flounders (34- salinity data are available for any of the col­ 44 mm, total length) from a shallow riffle lections, the species are all listed as euryha­ area of Katmai River about 3,2 km, above the line forms by Gunter (1956), stream mouth on August 21, 1964.

ZOOGEOGRAPHICAL IMPLICATIONS Additional collecting will undoubtedly reveal The term "primary freshwater fishes" was new locality records, but sufficient data are used by Miller ( 1958) for families that have now available to define the basic distribution been restricted to fresh water throughout of fish in the principal drainages of Katmai their known history, His classification applied National Monument, Striking differences occur to fishes in the monument would include only in these drainages, and the mountains of the blackfish, pike, and suckers, although Wal­ Aleutian Range (fig, 2) appear to have been an ters (1955, p, 334) reported that the longnose effective barrier to the southward dispersal of sucker "readily enters coastal waters" in the many fishes now in the Bristol Bay drainages, Arctic, Other cold-water fishes such as white­ All fishes along the Shelikof Strait side of fishes, lake trout, grayling, and burbot that the monument--salmon, char, sticklebacks, are normally confined to fresh-water streams and sculpins--are from groups known to have and lakes in the southern parts of their ranges high tolerance to salinity and to be capable do invade coastal waters in the Arctic (Wal­ of marine dispersal, None of the principally ters, 1955). We have no evidence that any of fresh-water fishes north of the Aleutian Range--whitefishes, lake trout, grayling, ~ 9 Chief, Salmon Investigations, Bureau of Commercial blackfish, pike, suckers, and burbot--occur Fisheries Biological Laboratory, Box 155, Auke Bay, in the Shelikof Strait drainages (table 2). Alaska 99821. Personal communication, 1964.

14 Table 2.--Fishes collected in principal drainages of Katmai National Monument

Bristol Bay drainages

Common name Scientific name Naknek River King Salmon Shelikof system River Strait system drainages

Arctic lamprey Lampetra ~nica X

Humpback whitefish Coregonus £idschian X

Least cisco C. sardinella X X

Pink salmon Oncorhynchus gorbuscha X X Chum salmon ---0. keta X Yx Coho salmon 0. kisutch X X X

Sockeye salmon 0. nerka X X X

Chinook salmon Q. tshawytscha X X

Pygmy whitefish Prosopium coulteri X

Rounl whitefish f· ~lindraceum X X

Rainbow trout Salmo z.airdnerii X

Arctic char Salvelinus al£inuS X

Dolly Varden S. malma X X

Lake trout ~. namaycush X

Arctic grayling ~llus arcticus X

Pond smelt liypanesus ~ X

Alaska blackfish Dallia ~toralis X Northern pike -----Esox lucius X Longnose sucker Catostamus catostamus X Burbot ---Lota lota X Threespine stickleback Gasterosteus aculeatus X X

Ninespine stickleback Pungitius ~itius X X X

Coastrange sculpin Cottus aleuticus X X X

Slimy sculpin ~· ~natus X X

1 Not personally verified by us (see footnote 5).

15 these fishes now live in brackish waters in Naknek system because it is a stream environ­ the vicinity of Katmai National Monument, and ment without the ecological diversity of a assume their dispersal in southwest Alaska multilake-stream complex. was essentially by fresh-water pathways or The Naknek River system has by far the via bridges of low salinity during early post­ most diverse ecological habitats of the drain­ glacial stages. ages in Katmai National Monument. This The presence of pygmy whitefish (Kendall, interconnecting stream-lake complex has also 1917), pond smelt (Narver, 1966), and black­ been open to invasion by fresh-water fishes fish (Roos, 1959) in the Chignik River system from the Bering Refugeum (see McPhail, 1963, supports the concept of the Aleutian Range and Flint, 1945) without significant barriers barrier. This system is about 274 km. south­ since early postglacial times. It has the west of the monument at a point on the Alaska largest number of fishes (24 species) of the Peninsula where the mountains become dis­ major drainages in the monument, and in­ continuous. It drains into the Pacific Ocean cludes all species found in the other drainages southwest of Shelikof Strait, but the headwaters (table 2). are flat, low in elevation, and only slightly We believe that open avenues offresh-water separated from other streams that flow north­ dispersal account for the principal differences west into Bristol Bay. Headwater stream trans­ in fishes in the Naknek system and Shelikof fer, which in the Chignik system would amount Strait drainages. Although no single Shelikof to a lowlands transfer, or flooding as discussed Strait drainage approaches the ecological di­ by Walters (1955, pp. 329-330), could easily versity of habitats in the Naknek system, they account for the occurrence of these fresh­ afford a variety of stream and lake environ­ water fishes in the system, illustrating that ments, These environments along the Shelikof where the Aleutian Range breaks down the Strait side of Katmai National Monument sug­ fishes from Bristol Bay drainages are dis­ gest that many (though probably not all) fishes persed southward. in the Naknek system could survive in various That pygmy whitefish, pond smelt, and parts of the Shelikof Strait systems if given blackfish occur in the Chignik system and access to them. Differences between fishes in apparently in no other drainages flowing into the Naknek and King Salmon River systems, the Pacific Ocean side of the Alaska Peninsula on the other hand, appear mostly related to illustrates that most Shelikof Strait drainages the lack of a diverse habitat in the King Salmon are isolated not only from fresh-water areas River system, This lack is particularly note­ north of the Aleutian Range but also from worthy in that portion of the King Salmon each other. Adjacent drainages along Shelikof River system that lies within the monument, Strait, particularly in much of the monument, Additional collections in the King Salmon flow parallel to each other but are separated River downstream from the monument should and isolated bybuttress ranges and deep fiords. help clarify this point, It seems unlikely that species incapable of The number of species in various lakes of tolerating high salinity would be dispersed the Naknek system differed considerably, All along the Shelikof Strait coastline. 24 known species in the Naknek system occur Similarities in the ichthyofauna of Shelikof in Naknek Lake (table 3), Naknek Lake as the Strait drainages and nearby Kodiak Island downstream terminus of the multilake system emphasize the isolation of these areas from is the first area to be invaded by fish through continental sources of fresh-water fishes. existing drainage patterns, All upstream lakes The only fishes reported from Karluk or Bare that drain directly or indirectly into Naknek Lakes on Kodiak Island (Greenbank and Nel­ Lake have fewer species than Naknek Lake, son, 1959) that were not found in fresh waters and in each upstream lake a barrier or poten­ along Shelikof Strait of the monument were tially adverse environment now exists between chinook salmon, Arctic char, and rainbow it and Naknek Lake, trout. Each of these species could occur in The heterogeneity of species in different Shelikof Strait drainages because they com­ lakes of the Naknek system appears to be re­ monly live in the sea during part of their lated to the development of barriers and the life cycle. sequential timing of the invasion of species The scarcity of species in the King Salmon into the system, Physical barriers such as River system portion of the monument may be falls or rapids or ecological barriers such as related to the limited diversity of habitat in unfavorable environments may have prevented the headwaters. Although we collected only dispersal of certain species within the system, eight species in this system, zoogeographically Development of barriers before the invasion it has the same potential for fresh-water of species would prevent dispersal beyond fishes as the Naknek system because it lies the barriers, whereas earlier invasion before north of the Aleutian Range barrier and ad­ barriers formed could allow more extensive jacent to the Naknek system. The collections dispersal, in the King Salmon River system were ad­ The Naknek system has been extensively mittedly small, but this system probably does glaciated, Muller (1952) discussed at least not have all of the fishes that occur in the four major glacial episodes in the system,

16 Table 3.--·0ccurrence of fishes in the major lakes and basin complexes in the Naknek River system of southwest Alaska [Surface elevations in meters above sea level for each basin are in parentheses] Coville- .Hammersly- Brooks Lake Idavain Lake Naknek Lake Grosvenor Lake MJ.rray Lake Fishes present (10 m.) (19m.) (32 m.) (232 m.) (488 m.)

Arctic lamprey X X X Humpback whitefish X X Least cisco X X Pink salmon X X X Chum salmon X X X Coho salmon X X X Sockeye salmon X X X X Chinook salmon X X X Pygmy whitefish X X X X Round whitefish X X X Rainbow trout X X X X .Arctic char X X X X X Dolly Varden X X X Lake trout X X X X Arctic grayling X X Pond smelt X X X Alaska blackfish X X X Northern pike X X X Longnose sucker X X Burbot X x(?) Threespine stickleback X X X Ninespine stickleback X X X Coastrange sculpin X X X Slill\Y sculpin X X X X X

Total known species 24 20 23 2 6 the earliest of which was probably pre-Wis­ Although Brooks Lake drains directly into consin. The present lakes and basins in the Naknek Lake, four species in Naknek Lake-­ Naknek system were formed by glacial activity humpback whitefish, least cisco, longnose during Wisconsin times, which Muller (1952) sucker, and burbot--apparently do not occur called the Brooks Lake glaciation. This glacial in Brooks Lake. These fishes probably were episode consisted of one main glacier--Naknek barred from Brooks Lake by a 1.5- to 2-m. Glacier--which occupied all the present basins falls that developed on Brooks River. These of Naknek Lake and had prominent tongues species may have been late invaders of the projecting through the present Coville -Gros­ Naknek system. During certain phases of the venor and Brooks Lake basins. The last major postglacial development of the system, Brooks glaciation in the Naknek system probably and Naknek Lakes were one lake, but con­ ended about 8,000 B. C.; a minor read vance tinued downcutting of end moraines on Naknek into Iliuk Arm of Naknek Lake ended about Lake subsequently lowered the lake level, sepa­ 4,000 B.C. {Karlstrom, 1957). rated Brooks and Naknek Lakes, and developed Invasion of fish in the Naknek system prob­ Brooks River. The time of sufficient separa­ ably began with the initial glacial recession tion of the two lakes to allow the formation of and the first development of postglacial drain­ Brooks Falls has been suggested at 2,000 age systems. At various times the entire sys­ B. C. 20 Many species now in Brooks Lake, tern, including Naknek River, was covered such as pygmy whitefish, pond smelt, black­ with ice, so the fresh-water fishes invading fish, sticklebacks, and sculpins, probably could the system would have come directly or in­ not surmount Brooks Falls as it now exists; directly from the Bering Refugeum. Anad­ romous and salinity-tolerant forms could 2 0o. E. Dumond, Depanment of Anthropology, Univer­ have come from the Bering Sea or Pacific sity of Oregon,' Eugene, Oregon. Personal communication, Ocean. 1965.

17 these species probably were present when complexes (table 3) and were probably among Brooks and Naknek Lakes were continuous, the earliest fishes to invade the Naknek system. Only one species in Naknek Lake, Arctic Hammersly and Murray Lakes, the highest grayling, does not occur in the Coville-Gros­ lakes in the Naknek River system {about venor Lakes complex (table 3 ). Three anad­ 488 m,) have only six species (table 3), These r omous f i s he s, pink, chum, and chinook lakes drain into Coville Lake through 87-km.­ salmon, occur only rarely in this area. The long American Creek. This creek has no com­ absence of grayling in these lakes is puzzling plete barriers, and anadromous sockeye salmon because no major physical barriers exist annually spawn at the outlet of Hammersly between Grosvenor and Naknek Lakes, and the Lake. A series of precipitous rapids in Ameri­ Grosvenor River and lower American Creek can Creek, however, probably forms an effec­ appear to be excellent grayling habitat, If gray­ tive barrier to many species in Coville Lake ling are a relatively late arrival and still in and lower American Creek that do not occur in the process of dispersal in the system, cau­ Hammersly and Murray Lakes. tious speculation suggests that their dispersal Hammersly and Murray Lakes and upper may be limited by the extreme turbidity of American Creek at one time may have drained Iliuk Arm and Savonoski River, The general north into the present Alagnak River system distribution and abundance of grayling in the (fig. 1). The Coville-Grosvenor tongue of the system are limited, and no large numbers of main Naknek Glacier pushed north almost to this species are available to probe routinely the Alagnak River (Keller and Reiser, 1959, through adverse environments. Other more plate 29) and coalesced with ice from the basin complex environmental factors maybe limiting of Nonvianuk Lake (Muller, 1952, p. 62), Drain­ grayling distribution in the system, age diversions associated with the retreat of We know of only two fishes, Arctic char and glaciers in this area may have caused the slimy sculpin, that inhabit Idavain Lake, which headwaters of what is now American Creek to drains into Naknek Lake through the 26-km.­ flow north into the Alagnak River, Certain long Bay of Islands Creek. Two series of falls, physiographic features of the area suggest about 4.8 and 16,1 km. below the lake, have this possibility (see footnote 19). If this diver­ probably prevented access to the lake by other sion did occur some of the fishes now in fishes since an early stage of glacial recession. Hammersly and Murray Lakes could have Arctic char and slimy sculpin are the only invaded the lake via the Alagnak River system species common to all of the major basin rather than the Naknek River system.

SUMMARY 1, Katmai National Monument is divided by 6. The last major glaciation of the Naknek the Aleutian Mountain Range into two princi­ system ended about 10,000 years ago. Sequen­ pal drainage areas: (1) streams flowing into tial timing of species invasion and the post­ Bristol Bay of the Bering Sea {dominated by glacial development of barriers have strongly the multilake Naknek River system) and influenced the present distribution of fishes in (2) streams flowing into Shelikof Strait of the this system, All upstream lakes in t,he Naknek North Pacific Ocean (characterized by many system have fewer species than the lakes short streams and rivers and a few small below them, and barriers exist between each lakes). upstream and downstream lake. 2. Twenty-four species of fresh-water 7. Arctic char and slimy sculpin, the only fishes, many of which are or can be anad­ species common to all basin complexes of the romous, occur in the monument. All 24 spe­ Naknek system, were apparently early in­ cies occur in the Naknek system, at least vaders, Lake trout, pygmy whitefish, rainbow eight of them occur in the King Salmon River trout, and sockeye salmon are present in all system, and at least eight of them occur col­ basins except Idavain Lake. lectively in the 18 to 20 systems draining into 8. Least cisco, humpback whitefish, and Shelikof Strait. longnose suckers ·could have been late invaders 3. Four marine fishes, all known euryha­ of the Naknek system because these fish ap­ line forms, occur in fresh water in or near parently are restricted from Brooks Lake by the monument. a falls on Brooks River. Sufficient separation 4. The Aleutian Range has been a barrier of Brooks and Naknek Lakes to allow the de­ to the southeastward movement offresh-water velopment of Brooks Falls is estimated to fishes within the monument area, All species have occurred around 4,000 years ago. in Shelikof Strait drainages are capable of 9. Hammersly and Murray Lakes, although dispersal through salt water, whereas many not in the monument, are important in the forms in Bristol Bay drainages require fresh distribution of fishes in the Naknek system. water for dispersal, These lakes and the upper portion of Ameri­ 5, Dolly Varden occur, apparently through can Creek may have drained north into the natural dispersion, in the crater lake in the Alagnak River system during early post­ caldera of Kaguyak Volcano, glacial drainage development. 18 ACKNOWLEDGMENTS Much of our collecting could not have been vided logistics support along Shelikof Strait. done without the help and skill of local bush The Alaska Department of Fish and Game gave pilots and Fish and Wildlife Service amphibious information on the occurrence of salmon along aircraft pilots. Pete Rice, Mike Novak, and Shelikof Strait, Carl Bond of Oregon State Uni­ Don Dumond, archaeologists from the Uni­ versity reviewed the manuscript. The National versity of Oregon, collected fishes and pro- Park Service encouraged the study.

LITERATURE CITED AMERICAN FISHERIES SOCIETY, COMMIT­ Alaska. U.S. Fish Wildl. Serv., Fish. TEE ON NAMES OF FISHES. Bull. 65: 555-579. 1960, A list of common and scientific names HUBBS, CARL L., and KARL F. LAGLER. of fishes from the United States and 1958. Fishes of the Great Lakes Region. Canada, Amer, Fish. Soc., Spec. Publ. Cranbrook Inst. Sci., Bull. 26 (Rev. ed.), 2, 102 pp. 213 pp. BLACKETT, ROGER F. KARLSTROM, T. N. V. 1962. Some phases in the life history of the 1957. Tentative correlation of Alaskan Alaskan blackfish, Dallia pectoralis. g~acial sequences, 1956. Science 125: Copeia 1962: 124-130. 73-74. BROWN, C. J. D, KELLER, SAMUEL A., and HILLARD N. 1952, Spawning habits and early development REISER. of the mountain whitefish, Prosopium 1959. Geology of the Mount Katmai area, williamsoni, in Montana, Copeia 1952: Alaska. U.S. Geol. Surv., Bull. 1058G: 109-113. 261-298. CAHALANE, VICTOR H. KENDALL, W.C. 1959. A biological survey of Katmai Na­ 1917. A second record for the Coulter 1 s tional Monument. Smithsonianinst., whitefish (Coregonus coulteri Eigen­ Misc. Coll. 138(5): 1-246. mann). Copeia 45: 54-56. CLEMENS, W. A., and G. V. WILBY. LINDSEY, C. C. 1961. Fishes of the Pacific coast of Canada. 1962. Distinctions between the broad white­ Fish. Res. Bd, Can., Bull. 68 (2d ed.), fish, Coregonus ~· and other North 443 pp. American whitefishes. J. Fish. Res. FLINT, R. F. Bd. Can. 19: 687-714. 1945. Glacial map of NorthAmerica. Geol. 1964. Problems in zoogeography of the lake Soc. Amer., Spec. Pap. 60 (pt. 1 ), 1 p. trout, Salvelinus namaycush. J. Fish. GREENBANK, JOHN, and PHILIP R. NELSON. Res. Bd. Can. 21: 977-994. 1959, Life history of the threespine stickle­ McPHAIL, J. D. back, Gasterosteus aculeatus Linnaeus, 1960. Annotated bibliography on Arctic in Karluk Lake and Bare Lake, Kodiak North American freshwater fishes. Island, Alaska. u.s. Fish Wildl. Serv,, Univ. British Co 1 u m b i a, Vancouver, Fish. Bull. 153: 537-559. Inst. Fish., Mus. Contrib. 2, 20 pp. GRIGGS, ROBERT F. 1961. A systematic study of the Salvelinus 1922, The valley of Ten Thousand Smokes. alpinus complex in North America. Nat. Geogr. Soc., Washington, D.C., J. Fish. Res. Bd. Can. 18: 793-814. 340 pp. 1963. The post-glacial dispersal of fresh­ GUNTER, GORDON. water fishes in northern North Amer­ 1956. A revised list of euryhalin fishes of ica. Ph. D. Thesis, Dep. Zool,, McGill North and Middle America, Amer. Mid­ Univ., Montreal, 167 pp. land Natur, 56: 345-354. MERRELL, THEODORE R., JR. HEARD, WILLIAM R. 1964. Ecological studies of sockeye salmon 1962. The use and selectivity of small­ and related limnological and climato­ meshed gill net at Brooks Lake, Alaska. logical investigations, Brooks Lake, Trans. Amer. Fish. Soc. 91: 263-268. Alaska, 1957. U.S. Fish Wildl. Serv., 1966. Observations on lampreys in the Nak­ Spec. Sci. Rep. Fish. 456, iv + 66 pp. nek River system of southwest Alaska. MILLER, ROBERT RUSH. Copeia 1966: 332-339. 1958. Origin and affinities of freshwater HEARD, WILLIAM R., and WILBUR L.HART­ fish fauna of Western North America. MAN. In Hubbs, C. L. (editor), Zoogeography, 1966. Pygmy whitefish (Pros opium coulteri) pp. 187-222. Amer. Ass. Advance. Sci. in the Naknek River system of southwest Publ. 51.

19 MULLER, ERNEST HATHAWAY, Alaska. Trans, Amer, Fish, Soc. 88: 1952, The glacial geology of the Naknek 253-260. District, the Bristol Bay region, Alaska. WAHRHAFTIG, CLYDE. Ph. D. Thesis, Univ. ill., Urbana, 1965. Physiographic divisions of Alaska. 98 pp. U.S. Geo1, Surv,, PJ;"of. Pap. 482, 52 pp. NARVER, DAVID WELLS, WALTERS, V, 1966. Pelagia! ecology and carrying ca­ 1955, Fishes of western Arctic America pacity of sockeye salmon in Chignik and eastern Arctic Siberia, Amer, Mus, Lakes, Alaska, Ph, D, Thesis, Univ, Natur. Hist,, Bull, 106: 255-368, Wash,, Seattle, 348 pp, WILIMOVSKY, NORMAN J. ROOS, JOHN F. 1954, List of the fishes of Alaska. Stanford 1959, Feeding habits of the Dolly Varden, Ichth. Bull. 4: 279-294 Salvelinus malma (Walbaum), at Chignik, Ms #1890

ARLIS Alaska Resources r Jbrary & Information Services Anchorage. AK

20 GP 0 877-890 As the Nation's principal conservation agency, the Depart­ ment of the Interior has basic responsibilities for water, fish, wildlife, mineral, land, park, and recreational re­ sources. Indian and Territorial affairs are other major concerns of America's "Department of Natural Resources." The Department works to assure the wisest choice in managing all our resources so each will make its full contribution to a better United States-- now and in the future.

UNITED STATES DEPARTMENT OF THE INTERIOR U.S. FISH AND WILDLIFE SERVICE BUREAU OF COMMEROAL FISHERIES WASHINGTON, D.C. 20240 POSTAGE AND FEES PAID U.S. DEPARTMENTOF THE INTERIOR Al OFFICIAL BUSINESS THIRD CLASS

Return this sheet to above address, if you do NOT wish to receive this material O, or if change of address is needed 0 (indicate change including ZlP Code).