Conservation of Marine Birds of Northern North America

---

---

The international symposium "Conservation of Marine Birds of Northern North America" was convened because of a growing awareness that not all was well with our marine birds. The symposium provided a forum for scientists, governmental administrators, conservationists, and laypeople to discuss the diverse topics and issues that we must all understand if we are to act both responsively and responsibly to assure that marine birds will not be lost through our neglect.

The symposium was cosponsored by the Natural Resources Council of America, National Audubon Society, National Wildlife Federation, and the U.S. Department of the Interior, Fish and Wildlife Service; additional support was provided by the Canadian Wildlife Service, the International Association of Game, Fish, and Conservation Commissioners, the Pacific Seabird Group, the Sierra Club, the Smithsonian Institution, the Wildlife Management Institute, and the Wildlife Society.

Persons interested and knowledgeable in the many and varied aspects of marine bird conservation were invited to participate in this symposium. There were 139 registered and several score of unregistered participants in attendance. Major topics treated were: (1) socioeconomic considerations and conservation of marine birds; (2) the marine environment of birds; (3) status of marine bird populations on land and sea; (4) the biology and ecology of marine birds in the North; (5) conflicts between the conservation of marine birds and uses of other resources; (6) programs and authorities related to the conservation of marine birds; and (7) conservation of marine birds in other lands.

The objective of the symposium was to identify problems and the needed information and programs necessary for the conservation of marine birds of northern North America. For the purpose of this symposium the term "northern North America" referred to the coasts of Washington, British Columbia, Alaska, Yukon Territory, and Northwest Territories and the adjacent North Pacific and Arctic Oceans. "Marine bird" was defined as being any bird using marine or estuarine waters. Speakers were asked to describe the status of information or the state of the art as it pertained to their topic within the limitations set by the objective of the symposium. Examples from other regions and of bird species not found in the regions of concern were to be used for comparative purposes when little pertinent information was known for regions or species of concern. Speakers were asked to identify the gaps in the knowledge and methodology that are most critical to their topic.

I believe that this symposium was particularly successful in that it provided a timely forum for many scientists who were about to embark on studies of marine birds in those areas of Alaska and California being considered for outer continental shelf oil and gas exploration and development. These published proceedings may be of lesser importance from that standpoint because some data, particularly those on populations, are out of date. However, I believe that the proceedings will long be of importance to biologists and administrators alike in charting their respective courses to ultimately assure conservation of this valuable avian resource.

Many people from many organizations and agencies worked hard to put together the symposium in the relatively short time of about 8 months. Nathaniel P. Reed was the person primarily responsible for bringing this symposium to fruition. The Steering Committee was composed of Daniel A. Poole, John S. Gottschalk, David N. Nettleship, Amos S. Eno, C. Eugene Knoder, Warren G. King, Louis Clapper, Robert Hughes, Fred G. Evenden, James C. Bartonek, and me. James C. Bartonek, Warren G. King, David N. Nettleship (Co-chairmen), C. Eugene Knoder, David A. Manuwal, William H. Drury, and Spencer G. Sealy served on the Program Committee. David A. Manuwal and Terence R. Wahl arranged trips for persons to observe pelagic birds off the Washington coast and other birds on Skagit Flats. C. Eugene Knoder handled financial matters. John A. Sayre and Richard Bauer made arrangements for facilities and entertainment. Elaine Rhode prepared the program and abstracts for printing. John Pitcher kindly contributed the artwork used in this publication as well as that used in the program and abstracts.

George Reiger made general introductions to the symposium; Spencer G. Sealy, Daniel W. Anderson, and I served as Session Chairmen; and James C. Bartonek served as General Chairman. Elvis J. Stahr was guest speaker at the symposium banquet.

Most credit for the success of this symposium goes to the 52 persons who as authors, coauthors, or summarizers of sessions presented much meaningful information in their presentations, during recorded discussions, and during many informal occasions. I wish to make special recognition of Ian C. T. Nisbet for his skillful summary of the symposium.

Editorial assistance in preparing the proceedings was provided by Judith Brogan.

Harvey K. Nelson

Chairman of Symposium and
Director of Wildlife Resources

---

Migratory birds make up a resource that is shared by many people of many nations. Public awareness of marine birds—their manifold values, ecological requirements, and problems—is prerequisite to their protection. I am proud that the Fish and Wildlife Service can further this needed awareness by publishing these proceedings of the international symposium "Conservation of Marine Birds of Northern North America."

Lynn A. Greenwalt, Director
Fish and Wildlife Service

---

by

M. T. Myres

Department of Biology, University of Calgary
Calgary, Alberta, Canada T2N 1N4

During the last 30 years extensive literature in the fields of physical and biological oceanography has accumulated that is not readily accessible to the nonprofessional student of seabirds and not as widely understood by career seabird ornithologists as it should be. This literature in oceanography and marine fisheries is as extensive in Russian and Japanese together as in the main languages of Western Europe combined; this abundance compounds the problem of becoming familiar with it if, as a student of seabirds, one's interest in the literature is initially somewhat marginal. Nevertheless, to achieve the best possible appreciation of the oceanographic influences affecting seabirds, particularly in the north Pacific Ocean and its adjacent embayment seas, it is necessary to make the effort.

Because of the rigor of carrying out their primary duties while at sea, only a very few North American and European oceanographers or fishery biologists have found time to interest themselves in seabirds and then, with a few notable individual exceptions, only as an off-duty pastime. The reason is not far to seek. It is far less important to examine the ecology of organisms at the next highest level of the food chain to the ones that are the primary concern than it is to examine the next lowest level (the food of the fishes or, in the case of phytoplankton, the physical and chemical environment in which the organisms grow best).

Seabirds are at the very top of the marine food chain, and they are not wholly aquatic in any case since they mainly travel through the air rather than the water and reproduce on land rather than in the sea. Only with the relatively recent recognition that seabirds contribute to the recycling of nutrients back into the ocean to an important degree, have seabirds gained a new scientific constituency.

At about the same time, governments have begun to recognize that seabirds are relatively easily examined indicators of the presence of unseen chemical pollutants in coastal seas, perhaps primarily for the very same reasons that they were previously so largely ignored; namely, that they are at the top of the food chains (and so accumulate the most-persistent and least-degradable pollutants) and that the on-land failures in their reproductive biology are readily visible.

During the last 10 years, it has become evident that yet another fundamental science is even more basic to the achievement of a balanced and in-depth understanding of the influence of the environment upon seabirds—the combined field of astrophysics, geophysics, and climatology. New developments in this field (when they are not published in Nature or Science) appear in journals that are less familiar to seabird ornithologists than those in which the fishery biologists and biological oceanographers publish their findings.

Unfortunately, important advances in understanding the dynamics and energy transport mechanisms of both the atmosphere and the water masses of the oceans are not being picked up by students of seabirds because of the natural lag in communication that occurs between disparate disciplines. Only in the last few years have oceanographers and climatologists been invited to address gatherings of ornithologists, and the modesty with which they have sometimes done so has limited the impact of their offerings.

At this symposium, it was left to a biologist with no pretentions in either physics or mathematics to demonstrate the need for seabird ornithologists to understand basic environmental processes well beyond their usual range of interests. I did so with a series of slides taken from this "other" literature, and I had intended to include in the published version of this paper an extensive bibliography, subdivided into category groupings, so that seabird ornithologists could make their own selection of the points in the spectrum at which they most needed information.

Unfortunately, limitations upon space in this volume, daily additions to the exploding literature, and my own inability to keep up with understanding this have forced me to omit any references and not to attempt to expound detailed specific physical mechanisms.

Thus unencumbered here, I shall briefly outline instead what I perceive to be some of the significance for seabird ornithology and conservation of the rapidly expanding understanding of the oceans, the air-sea interface, atmospheric dynamics, and influences upon the world's climate of extraterrestrial events.

There is no need to dwell on the well-known events that could be mentioned under this heading. Seabird ornithologists are familiar with the fact that the atmosphere is the medium of seabirds both when searching the ocean for feeding areas and when on migration, and also a violent enemy, as when particular storms cause occasional "wrecks" of seabirds inland from coastlines. As a refinement of the former, Manikowski of Poland suggests that seabirds respond to the passage of weather systems, so that their distribution over the open ocean may be constantly changing. Whereas some species may attempt to avoid the stormy conditions of low-pressure areas (cyclonic conditions), others more highly specialized for exploiting the aerodynamic properties of wind over a moving water surface may possibly, instead, try to avoid large high-pressure regions (anticyclonic conditions with little or no wind). My student, Juan Guzman, is attempting to determine whether this may be so; if it is, it might be possible, for example, to predict some things about the distribution patterns and population structure of southern hemisphere shearwaters while they are visiting the oceans of the northern hemisphere during the nonbreeding season.

In comparison with the "wrecks" brought about by storms, which are of short duration and not usually very serious, seabird ornithologists are also familiar with relatively brief and localized disasters caused by changes in the ocean itself. The best-known example is a slight change in the boundary of an ocean current (or other shift in the position of a distinctive water mass) that results in the failure of food fishes to appear as they normally would, close to breeding sites of conspicuous colonial seabirds, such as the periodic shift in the El Niño off the west coast of South America. A scarcely studied refinement of this type of event would be the effects of less-pronounced oceanic changes that might reduce the planktonic food supply of nocturnally active, burrow-nesting seabirds. In such instances, the effects might also be a breeding failure for only one or two seasons; in all probability such events occur, but whether they are as likely to be detected by us is problematical. However, the populations of most seabirds are probably already adapted to survive short-term crises of this type because, having long adult life spans, reproductive adults that fail to raise young one year may mostly live to succeed in doing so in the next or succeeding year, when the oceanic "anomaly" has disappeared. What constitutes an "anomaly" will be considered again shortly.

A third critical condition for seabirds may be local or widespread, temporary or final, or some combination of these. A single local spill, or outfall, of a chemical pollutant will be short term if we can take steps to alleviate the consequences or stem the flow. Alternately, we may consider it to be long term if we take the view that it is one additional act of violence resulting from the "progress" of Industrial Man, and that it is never going to shift into reverse gear. We may say that the effect on seabird populations of spills of oil products or chemical pollutants in coastal waters of a region will be a "final solution" for any that become wholly extinct before the oil wells go dry or the industries fail. On the other hand, the effect will have been merely a perturbation of the population if the species survives and outlives these activities. Recent upturns in populations of peregrine falcons (Falco peregrinus) and pelicans (Pelecanus sp.) in certain places where environmental controls have been enacted give us hope that crises of several years' duration can be withstood by at least those species that once were common in relation to their respective food sources or available safe breeding habitats. The really critical features to document are the means whereby abandoned breeding sites are reoccupied and the time it takes.

It must never be forgotten that we know almost nothing about the ecology of subadult or nonreproductive adult seabirds during the years they are at sea unconfined by membership in a breeding unit and that we know almost nothing about the activities of pelagic seabirds in the nonbreeding season. These birds may be far from land and hard to study, but what happens during those phases of their lives is basic to the composition of the colony and condition of the birds when breeding. A start would be to learn everything that is known and is being discovered about the oceans by oceanographers and, thus forearmed, go looking for the seabirds with certain questions clearly in mind.

A scientist's working life lasts only a few decades, and few studies of seabirds by a single author or agency have been continued for longer than 5-10 years on any one problem. Further, while we as individuals may live to be equally active in a certain field of research 20 years hence, our collective conscience and collective muscle consist of several levels of government that tend to exhibit 4- or 5-year changes of direction and priorities. Certainly, the civil service may live on as an inertial recorder of collective experience. Certainly, too, those who live under one form or another of dictatorship or, as in some Canadian provinces, where conservative patterns of voting occur, may experience a continuity of research and development and conservation policies that exceed the 4- to 5-year turnaround pattern that is most common. Yet, even these more continuous systems may come to an end quite suddenly because of economic or political happenstance.

The point of this digression is to show that seabird ornithologists must not rely on government programs to provide continuous data over a long period of years—not, at least, in most countries. Monitoring the biological circumstances of seabirds is not the same as recording the temperature regularly by machine at a weather station, since this activity is unlikely to be terminated unless the society collapses altogether. We may know that in some countries the amateur naturalist exists in such numbers that records of seabirds will continue to be made whatever the circumstances. Nevertheless, planning of censuses that will be repeated every 10 years is best assured if government and career biologists combine with the amateur element, so that any one of them can continue the work if any other element should be incapacitated. At any one time, either the amateur or the government or the university personnel may be the prime mover, and each of these forms now exists in various countries.

What the scientific literature in the fields of the geophysical, atmospheric, and oceanographic disciplines demonstrates is that natural climatic oscillations probably range in length from the 11-year sunspot cycle through several decades (or a human lifetime) to several hundred years. So, when our children are the new trustees of seabird colonies 20 or 40 years hence, they must interpret their data using the full range of physical as well as biological data that we can leave for them. Indeed, the information is, I believe, already available over a long enough period (since 1940 at least) to allow some speculative interpretations of what may have been happening to our seabird populations, whether or not we knew or had any evidence of it.

I have already suggested that extraterrestrial events, particularly the 11-year sunspot cycle, are increasingly believed to influence the atmosphere of this planet. The Chinese and Japanese have remarkably precise records of the northern limits of certain agricultural crops at particular times, the phenology of flowering, and the freezing of lakes. These demonstrate long-term trends in overall climate in eastern Asia that extend over hundreds of years. The climate of Japan is influenced by the high-pressure area in winter over mainland East Asia. There is evidence that severe ice conditions in the Bering Sea during the early 1970's may have been due to an eastward shifting of this high-pressure area. Again, the water mass of the Kuroshio Extension and the West Wind Drift takes several years to travel across the Pacific Ocean, and there is an established temperature variation that travels like a slow wave with it. Off Japan, the Kuroshio Current periodically develops meanders which slow the speed of the eastward flow. Cold and warm "pools" of water approach the west coast of Canada and the western United States from time to time.

Ocean currents are driven by the atmospheric motion above them, which consists of several convective cells between the equator and each pole. The outcome is zonal winds, such as the trade winds and the westerlies. However, as the influence of the sun on the atmosphere is variable, the input of heat and the extent of the major high-pressure areas vary, as does the path of the jet stream. The recent droughts in northern Africa and unusually heavy rains in Australia are both linked to a southward shift of the Intertropical Convergence Zone in the atmosphere and a "corrugation" of the wind circulation from a more normal zonal (latitudinal) path. These shifts in the atmospheric circulation are almost certainly transmitted also to the ocean currents and the marine ecosystem, with the influence being felt for a long period of years.

One of the oceanic domains of the North Pacific is the transitional domain, which lies east-west where the West Wind Drift impinges upon the coasts of British Columbia and Washington State. It is precisely in this sector that there was a well-documented "temperature anomaly" in 1957-58. Since an anomaly implies something completely out of the ordinary, I seriously question the appropriateness of the term for an event that may or may not be recurrent (at the time it was a pronounced variation from the oceanographic records accumulated up to that time, but the period had not been a very long one). It is no coincidence that the numbers of albatrosses recorded at Ocean Weather Station "Papa" was higher during this warm-water "anomaly" than subsequently (indeed, an 18-year record of the seabirds recorded at "Papa" also exhibits other interesting fluctuations from the base-line data in certain years).

Recent analyses of sediments from off the coast of California have demonstrated long-term fluctuations in sardine populations extending back at least 1,800 years, with increases lasting 20-150 years and spaced 20-200 years apart. The number of anchovies declined steadily. Yet until now, El Niño events have been treated as anomalies in that region as well as off the coast of Peru. Just as we recognize that different species of fish follow the warm water north on such occasions, we must also recognize the rather distinct seabird species assemblage that is trapped, as it were, in the Gulf of California. Clearly, like the termination point of the West Wind Drift at about the 45-55° parallel, the coast of Baja California and southern California State, from the 25-35° parallel where the California Current begins to swing away from the coast to the west as the North Equatorial Current, is another zone of instability.

I think that it is no accident that the southern limit of several northern species of North Pacific seabirds ends in southeastern Alaska or northern British Columbia, and that the northern limit of the ranges of several other species occurs in Washington State or southern British Columbia. Indeed, the west coast of Vancouver Island is not rich in species, and several of those that exist are not present in great numbers. This is a region of rather more variable conditions than elsewhere, and species evidently find that it is difficult to colonize and it quickly becomes unsuitable again. Since 1940, indeed, there has been a parallel decline in the annual mean sea-surface temperature at a number of coastal recording stations in British Columbia, and this seems to have been a rebound from a less well-documented rise in sea-surface temperatures during the 20 years before that, which culminated in a peak around 1940. Salinity has likewise trended downwards during the last 30 years. The seabird colony size data before 1960 are so nonquantitative that it is impossible to be sure what changes in seabird populations and breeding sites may have taken place in response to these physical changes.

The lesson is that we must now examine all future census and distribution data with trends in sea-surface temperature and salinity in mind as two of several likely factors influencing them. We must no more ignore data outside our own field than a salmon ecologist might.

We know little of the accuracy of censuses of seabird numbers made between 1850 and 1950. There has been a tendency to assume that numbers of seabirds at long-established colonies have been relatively unchanging, even though the expansion of some species into previously unrecorded breeding sites in low numbers is well documented. Contraction of breeding ranges, likewise, has most commonly been attributed to the influence of man. Recent literature from the physical sciences, on the contrary, suggests that seabird numbers at particular colonies are most unlikely to have been stable for any great length of time, at least at high or middle latitudes and particularly at points where boundaries between currents impinge on continental coasts. Indeed, some early estimates of colony sizes may not have been as much in error as we may have assumed, neither when apparently too large nor when apparently unlocated by previous visitors.

The halving of a large colony over a period of 20 to 50 years in the middle of the range of a species and the establishment and disappearance of smaller breeding groups at opposite extremes of the range (both latitudinally and longitudinally), may equally reflect natural long-term climatic or oceanographic changes and may naturally be reversed at some time in the future, perhaps within half a century. The implication for conservation of seabird colonies that are at the contracting end of a species' range is that cultural rather than biological criteria may be the best determinants.

---

by

George J. Divoky[1]

U.S. Fish and Wildlife Service
Fairbanks, Alaska

Sea ice is a major factor in the distribution and ecology of many of the birds treated in this symposium. Sea ice is defined here as ice formed by the freezing of seawater and includes both free floating pack ice and the more stable shorefast ice. Since icebergs are composed of ice of land origin, they are not discussed.

Before discussing the specific relationship of birds and sea ice in the Beaufort, Chukchi, and Bering seas, I list the general effects that arctic ice can have on seabirds. For purposes of discussion these effects can be divided into negative effects, or disadvantages, and positive effects, or advantages.

Negative Effects

Sea Ice Decreases the Surface Area of Water

The decrease in the surface area of water is the simplest and most immediate effect that sea ice has on birds. Ice acts as a barrier that restricts the availability of food in the water. Surface feeders are the most severely affected since, in general, ice cover of 50% reduces the possible feeding area by half. The effect on diving species is not as severe since, if open water is scattered throughout the ice, diving species still have access to much of the prey in the water column and benthos. When open water is scarce, however, diving species can become concentrated in the available water, resulting in intense competition for available prey. In certain situations the open water is used only as a migratory pathway, but open water is necessary for birds that must roost or feed.

Sea Ice Reduces Primary Productivity in the Water Column

Ice inhibits phytoplankton blooms in the water column, thus decreasing the biological productivity of ice-covered waters. This inhibition occurs in two ways:

• By decreasing light penetration of the water column.—Much of the sunlight reaching the ice is reflected by the ice and by snow on the ice. The amount of light reaching the water depends on the angle of the light, thickness of ice, and amount of snow cover. When the layer of under-ice algae forms, it absorbs light and further reduces the amount of light reaching the water (Bunt 1963). This reduction in light reduces the depth of the euphotic zone.

• By increasing the stability of the water column.—Increased stability of the water column reduces the upwelling of nutrient-rich waters into the euphotic zone. Ice stabilizes the water column primarily by preventing wind-driven movement of surface waters and by forming a layer of meltwater at the surface in the spring and summer (Dunbar 1968).

Sea Ice Reduces Benthic and Intertidal Biota

Benthic flora and fauna can be reduced by the presence of ice in two ways: In shallow water ice can freeze to the bottom for much of the year and prevent the establishment of plant and animal populations; and when ice floes are pushed together, they form underwater ice keels that can scour the bottom when the ice moves. Both of these events not only act directly to decrease benthic populations but also disturb the sediment, making it less suitable for colonization. In areas with heavy ice scour, sessile benthic populations can be greatly reduced, although motile species may move into scoured areas during the ice-free period in summer. In addition to preventing the establishment of sessile benthic animal populations, ice scour also prevents the establishment of beds of kelp and eelgrass (Zostera marina), thus decreasing the diversity and productivity of arctic inshore waters. Both kelp and eelgrass beds are important feeding sites for birds in areas south of the region affected by ice scour.

Sea Ice Allows Terrestrial Predators Access to Breeding Sites

The formation of ice between the mainland and offshore islands allows the arctic fox (Alopex lagopus) and other predators access to the islands used by breeding birds. Foxes can become permanently established on islands that have food sources during the period when birds are absent from the island. Often, however, there is little to attract foxes to the islands other than breeding birds. Because moats form around many islands before the breeding birds arrive, foxes are primarily a problem when moat formation is incomplete or when the breakup of ice is late. Arctic foxes are found on the pack ice throughout the summer and thus can visit islands that are separated from the mainland by open water but are adjacent to the pack ice.

Advantages

Sea Ice Provides a Matrix and Substrate for an Ice-associated Plankton Bloom and an Associated Under-ice Fauna

The first detailed studies on the blooms of diatoms that occur in the lower levels of ice were done by Appollonio (1961). The importance of this bloom in the energy budgets of arctic and subarctic seas has only recently been realized (Alexander 1974; McRoy and Goering 1974). In areas where ice is present throughout the year, the plankton bloom supports a population of under-ice invertebrates. These populations have been little studied but apparently consist primarily of copepods and amphipods (Mohr and Geiger 1968). Feeding on the invertebrates associated with the ice are two species of fish, polar cod (Arctogadus glacialis) and arctic cod (Boreogadus saida). Andriashev (1968) used the term cryopelagic to describe such fish, which are found in the midwater zone but also are associated with ice during some part of their life cycle.

The underside of multi-year ice has numerous ridges and pockets that provide a heterogeneous environment for the under-ice fauna. This environment is protected from disturbance from currents and wave action by ice keels acting as barriers, which also provide shelter from predators in the same manner as a coral reef. The overall effect of the under-ice flora and fauna is to increase the diversity of surface waters in arctic seas by creating an inverted benthic biota.

Sea Ice Provides Hauling Out Space for Marine Mammals

The mammals that inhabit the ice in the Chukchi and Bering seas and their adaptations to the pack ice environment were discussed by Fay (1974). Many of these species frequently haul out on the ice, where they provide food in the form of feces, placentas, and carcasses.

Sea Ice Provides Roosting Sites

Ice provides a hard substrate that allows seabirds to leave the water to roost. This allows such species as the Larus gulls, which typically roost on hard substrates, to occur in large numbers well offshore.

Sea Ice Reduces Wind Chill

The unevenness of the upper surface of the ice reduces the speed of winds directly over the ice, thus providing a microhabitat and reducing the amount of wind chill for birds sitting on and next to the ice.

Sea Ice Decreases Wave Action

Ice floating on the water reduces the surface disturbance of the water. Although swells pass through areas with much ice cover, waves do not. In addition, surface waters on the lee side of ice floes and cakes usually have little surface disturbance. Surface feeders may be able to locate prey more easily because of these reductions in surface disturbance.

The retreat of the pack ice each spring and the formation of new ice each fall greatly affect a large area of the Arctic Ocean off the coast of Alaska and much of the Bering Sea. Specific ways in which birds are affected by ice in the western Arctic are discussed on a seasonal basis. All observations are my own, unless otherwise stated.

Winter

Chukchi and Beaufort Seas

From late November to mid-April, ice cover of the Chukchi and Beaufort seas is almost complete. The only areas where birds can be expected to winter in these seas are the chronic lead systems. Such lead systems are found off Wainwright and Point Barrow and south of the Point Hope-Cape Thompson area (Shapiro and Burns 1975). Only the black guillemot (Cepphus grylle) is known to regularly winter offshore from Wainwright and Point Barrow (Gabrielson and Lincoln 1959; Nelson 1969). In the Point Hope-Cape Thompson area, glaucous gulls (Larus hyperboreus), the common murre (Uria aalge), and the thick-billed murre (U. lomvia) occur throughout the winter (Swartz 1967). It is likely that black guillemots are also found in this area.

The lack of chronic lead systems in the Beaufort Sea precludes the presence of wintering seabirds. The one species that may be found wintering in the Beaufort is the Ross' gull (Rhodostethia rosea). Ross' gull is believed to winter primarily in the Arctic Ocean (Bailey 1948). The number of sightings that have been obtained in both the eastern and western Arctic indicate that the species may winter over much of the Arctic Ocean. It may thus be expected to occur in both the Chukchi and Beaufort seas during winter.

Ice cover—not prey abundance—plays the major role in severely limiting bird numbers in the Arctic Ocean in winter. Prey is known to be abundant in parts of the Arctic Ocean during the period of ice cover. In the Chukchi Sea, Eskimos fishing through the ice can catch 23 kg of arctic cod per person per day (D.C. Foote, unpublished data). Eskimos jig for the fish at considerable depths, and the cod do not appear to be as common directly below the ice as they are in summer. The effects of new ice (which forms on the underside of the ice during the winter) on the under-ice fauna are not known. The abundance of amphipods in ice-covered waters in winter is demonstrated by the experience of the Greeley Expedition in the eastern arctic. They discovered that any scrap of food thrown into a lead was quickly consumed by amphipods. Nets were made to catch the amphipods and the availability of this food source played a major part in the survival of the expedition (Schmitt 1965).

Aside from the food found in leads in the ice, the only food available to birds in the Beaufort and Chukchi seas in winter is carrion and the feces of mammals found on the pack ice. The presence of the arctic foxes on the pack ice during the winter demonstrates the availability of scavenging opportunities on the ice. Arctic foxes on the pack ice live on feces and the remains of seals killed by polar bears (Ursus maritimus). Polar bear and seals are both common in the Beaufort and Chukchi seas in winter, but no scavenging seabirds are found there in the winter. It was thought that the ivory gull (Pagaphila eburnea) was associated with marine mammals during the winter, but they are now known to winter at the Bering Sea ice edge, where they feed on fish and crustaceans (Divoky 1976). The only birds associated with polar bear kills in the Chukchi Sea in March are ravens, Corvus corax (T. J. Ely, Jr., personal communication).

Bering Sea

Ice begins to cover the northern Bering Sea in November and reaches its maximum by February, when it usually extends as far south as the edge of the continental shelf, and covers nearly 75% of the surface of the Bering Sea (Lisityn 1969). Coverage can vary greatly from year to year. In certain years Bristol Bay may be completely covered and in others ice is found only in the northern part of the Bay. Almost all ice in the Bering Sea is first-year ice. This ice tends to be flat on the top and underside and in general lacks the extensive keels and pressure ridges found on multi-year ice.

The Bering Sea ice has a number of large-scale features of importance to birds. The "front" is a zone of ice south of the consolidated pack that is composed of small floes, ice pans, and brash ice. This zone is prevented from forming large floes by the action of swells from the open water to the south. The front continually changes in width. When winds are from the south, it is compressed into a narrow band; when winds are from the north, it is a broad zone composed of bands of ice interspersed with open water.

Polynias (areas of open water) are found immediately south of the large islands in the northern Bering Sea. They are formed by the southward movement of ice caused by the prevailing winds. This movement causes ice to be pushed away from the south side of islands, leaving areas of open water. Large polynias are associated with St. Lawrence, St. Matthew, and Nunivak islands and with the south side of the Seward Peninsula (Shapiro and Burns 1975).

The most biologically active area of the Bering Sea in winter is the ice front. Studies of primary productivity in April show that production at the surface in the ice front is high (1.98 mg C/m³ per h). Surface waters directly under the pack ice have much lower production (0.29 mg C/m³ per h), and that in the water south of the ice is lower yet. At this time production within the ice is very high (more than 5 mg C/m³ per h) (McRoy and Goering 1974). Because this phytoplankton bloom is trapped in the ice, it is not available to grazers. Thus, before the spring melt the ice front is the only area where a large quantity of phytoplankton is available to higher levels of the marine food chain.

The winter distribution of birds in the Bering Sea correlates well with the findings on primary productivity. Densities south of the ice and the continental shelf average less than 10 birds/km². At the ice front during one cruise in March, densities exceeded 500 birds/km². Densities at the ice front increase from south to north; they drop in the region where the ice front grades into more consolidated pack ice, and are less than 0.1 bird/km² in the consolidated pack.

The most numerous species at the ice front are common and thick-billed murres, which constitute more than 90% of all birds seen. Irving et al. (1970) were the first to report on the large number of murres at the ice front. Feeding flocks of 25,000 individuals have been observed at the front, in which densities were as high as 10,000 birds/km². No other diving species is common at the ice front. The parakeet auklet (Cyclorhynchus psittaculus) is seen on most cruises, but only during a small percentage of observation periods and always in low numbers. Black guillemots are common north of the ice front and stragglers are occasionally seen at the front. Pigeon guillemots (Cepphus columbus), least auklets (Aethia pusilla), and crested auklets (A. cristatella) are irregular visitors to the front.

Surface feeding species commonly found at the ice front include the northern fulmar (Fulmarus glacialis) and five species of gulls. The fulmar is common south of the ice and is found only in the southern portion of the front. Three species of Larus are found at the ice front. The most common is the glaucous-winged gull (Larus glaucescens); the glaucous gull is less frequently seen. The slaty-backed gull (L. schistisagus), a species that breeds in Asia, is most common west of St. Matthew Island (McRoy et al. 1971). The black-legged kittiwake (Rissa tridactyla) is common in open water south of the ice but is also found throughout the entire width of the front. The ivory gull is unique in that it is found only at the ice front in winter. In addition to these species, the fork-tailed storm-petrel (Oceanodroma furcata) is a regular but uncommon visitor to the ice front in winter. Densities of surface feeding species at the ice front are low when compared to the high densities of murres, and do not regularly exceed 10 birds/km².

The primary food consumed by birds at the ice front is pollock (Theragra chalcogramma). An amphipod (Parathemisto libellula) and the euphausiids are less important. Examination of the stomach contents of birds and fish show that large feeding flocks are usually associated with schools of pollock feeding on P. libellula and euphausiids.

The habitat of the consolidated pack in the Bering Sea is markedly different from that at the ice front. Whereas the front is characterized by bands of ice interspersed with open water and ice coverage rarely exceeding 4 oktas (4/8), the consolidated pack consists primarily of large expanses of unbroken ice. Small leads are formed by the shifting of the ice caused by currents and wind. Ice coverage is usually 7 to 8 oktas. The southern part of the consolidated pack, which grades into the ice front, has frequent leads. Most of the species found at the ice front can be found in the southern part of the consolidated pack, but murres are most common. Their numbers decrease, however, in the more northerly pack, where leads are less frequent. Black guillemots, in contrast, increase with increasing ice cover, and reach their greatest abundance in the small leads constantly forming and refreezing deep within the ice. Because they exploit this habitat, they are dependent on the formation of lead systems. I have often seen leads a quarter mile wide refrozen to the point where new ice covered all but a small patch of open water; black guillemots were frequently crowded into this open water. Before the lead closes completely the guillemots must fly to an open lead. When winds are light and temperatures low, lead systems fail to form as rapidly as usual, and when they do they refreeze quickly, causing a loss of the preferred habitat of wintering black guillemots. A severe winter in the White Sea in 1965-66 decreased the amount of open water and caused an increased black guillemot mortality (Bianchi and Karpovitsch 1969). On a windless day in March I conducted bird observations in the Bering Sea ice where no leads or open water were encountered. The only bird seen was a black guillemot flying over the ice. In situations such as this, where black guillemots are prospecting for open water, they may use the "water sky" and steam fog associated with leads as visual aids. "Water sky" is the reflection of the dark water in the clouds over the lead, and contrasts sharply with the "ice sky." The presence of "water sky" allows birds to detect open water from a distance of many miles.

Aside from birds found in and near island-associated polynias, only murres and black guillemots are regularly found on the consolidated pack ice in winter.

The polynia associated with islands in the consolidated pack provide refuge(s) for seabirds. Fay and Cade (1959) found the polynias south of St. Lawrence to be most important to oldsquaws (Clangula hyemalis). King eiders (Somateria spectabilis), common eiders (S. mollisima), and oldsquaws are common in the St. Matthew Island polynias (McRoy et al. 1971). Because these polynias are in shallow-water areas, they provide feeding opportunities for benthic feeding species.

Spring

Chukchi and Beaufort Seas

In April and May a lead system develops from the Bering Strait north to Cape Lisburne and then northeast to Point Barrow. The lead is a flaw lead that occurs between the shorefast ice and the free-floating pack. It is a major migration route for a number of species of birds, primarily eiders. East of Point Barrow in the Beaufort Sea, no similar well-defined large lead exists. Consequently, there is a greater chance of bird mortality occurring in the Beaufort Sea than in the Chukchi Sea because the early migrants are unable to find open water. In 1960, 10% of all the king eiders that migrate through the Beaufort Sea died during a late freeze (Barry 1968). Additional records of eider mortality due to late breakup or sudden freezes were presented by Palmer (1976).

In late May, rivers that empty into the northern Chukchi and Beaufort seas begin to flow. The shorefast ice is still present at this time and the rivers flow over the ice. For large rivers, such as the Colville and the Sagavanirktok, the area of ice covered by water is considerable. Openings in the ice develop sometime after the river runoff starts and the river water drains through the ice.

This river overflow plays an important role in the breeding biology of certain island nesting species, since the overflow surrounds islands and prevents arctic foxes from reaching the islands. The overflow also allows birds to sit in the water near breeding sites. It is not known whether river overflow contains prey items available to birds. After the overflow drains through the ice, the shorefast ice that has been covered with river overflow decomposes quickly, and patches of open water occur early in areas just seaward of major river deltas. For this reason the largest breeding colonies on barrier islands in the northern Chukchi and Beaufort seas are all found near the mouths of large rivers. Islands away from rivers become isolated from the pack ice by moats, which are caused by the absorption of solar radiation by the islands and the melting of the ice immediately adjacent to them. Moat formation is not as predictable and uniform as river overflow.

Bering Sea

When the ice in the Bering Sea begins to melt in April, the edge of the pack does not recede northward as is frequently thought. Rather, there is a general decomposition of ice throughout the pack. The leads that are constantly forming in the ice no longer freeze. As melt continues and ice becomes rotten, leads form with increasing frequency. This manner of ice decomposition is important to birds. The leads that form deep in the pack ice provide feeding and roosting areas near the large seabird colonies found north of the ice edge, and are used by certain tundra-nesting ocean migrants such as eiders, red phalaropes (Phalaropus fulicarius), and jaegers (Stercorarius spp.). If ice decomposition is retarded by persistent low temperatures, the initiation of breeding may be delayed at northern Bering Sea colonies and for some tundra species.

At the time of decomposition the large standing stock of phytoplankton present in the pack ice is released into the water. No information is available on fish and invertebrate populations that are associated with the decomposing ice. The quantity of organic carbon released is considerable, although it is not known what fish or invertebrate populations are supported by this plankton as soon as it is released. For birds breeding in areas where ice is present in the initial stages of breeding, the phytoplankton released by the disintegrating ice could play an important part in the birds' energy budgets.

Summer

Chukchi and Beaufort Seas

In the northern Chukchi and Beaufort seas the nearshore marine environment is dominated by sea ice in June and July. In June the coastal areas are characterized by a snow-free tundra teeming with nesting waterfowl and shorebirds next to an expanse of sea ice almost completely devoid of bird life. In areas where river outflow does not occur, the use of nearshore waters usually begins when a moat forms along the shoreline. Amphipods and other invertebrates are found in this moat, especially at stream mouths. Limited but regular use of the moat occurs, primarily by loons (Gavia spp.), oldsquaws, and arctic terns (Sterna paradisaea).

As the snow on top of the shorefast ice begins to melt, ponds form on top of the ice. As melt proceeds, these melt ponds merge into long, parallel channels and may cover well over 50% of the ice surface. Only when thaw holes form and the melt ponds are connected to the water under the ice is food present in the channels. Amphipods are then seen swimming in these channels. Bird use of these channels is not extensive.

It is usually late July before the nearshore ice begins its rapid decomposition. Ice in the lagoons is the first to melt. Ice seaward of the barrier islands decomposes more slowly because of the presence of keels and pressure ridges. As the ice melts, the in-ice algal bloom is released into the water. These algae are important because they provide at least 25 to 30% of the productivity in coastal waters and allow the biological growing season to begin before the open-water plankton bloom occurs (Alexander 1974). In nearshore areas close to Barrow, large populations of mysids and amphipods are associated with the decomposing ice. At least in certain areas, these ice-associated zooplankton populations are a major food source for nearshore migrants, especially red phalaropes, arctic terns, and Sabine's gulls (Xema sabini).

The effects of ice scour on the shoreline and the nearshore bottom of the Chukchi and Beaufort seas is demonstrated by the absence of sessile benthic fauna and flora. The effect this absence has on birds is seen in the feeding habits of nearshore birds. Oldsquaws and eiders, which frequently feed on molluscs, feed instead on motile benthos species such as mysids, amphipods, and isopods. The emperor goose (Philacte canagica) is absent from the northern Chukchi and Beaufort seas, apparently due to the absence of eelgrass beds. Ice scour is the major cause of the absence of eelgrass in northern Alaska (C. P. McRoy, personal communication).

The offshore ice in the Chukchi decomposes more rapidly than that in the Beaufort, largely because Bering Sea water enters the Chukchi through the Bering Strait (Coachman and Barnes 1961). By late July the Chukchi is usually ice free as far north as Icy Cape. In the Beaufort, however, ice decomposition occurs slowly through June and July, and only in August does a definite strip of open water develop between the shore and the edge of the pack ice. The amount of open water varies greatly from year to year. In certain years the Beaufort is not navigable due to the lack of open water.

Aerial censusing in June and July shows that bird densities on the offshore ice are extremely low. In August and September, when shipboard censusing can be conducted, densities on the pack ice in both seas are about 10 birds/km². Unlike the Bering Sea, where densities south of the ice are much less than on the ice, bird densities south of the ice in the Beaufort and Chukchi seas are slightly higher in the open water south of the ice, averaging about 20 birds/km². In the Chukchi the principal species encountered on the ice are the black-legged kittiwake and the thick-billed murre. In the Beaufort, red phalaropes, oldsquaws, and glaucous gulls are the most common species.

Numerous arctic cod are associated with the underside of the summer pack ice. Shipboard censusing in the ice is complicated when cod are stranded on ice floes, as the ice shifts under the weight of the ship. Gulls, arctic terns, and jaegers gather behind the ship to feed on these fish; mixed flocks of more than 100 birds are common. In the absence of a ship to provide the disturbance needed to make large numbers of cod available, these birds are dependent on locating the fish in the surface waters next to ice floes. Because cod frequently swim over underwater ice shelves they are highly visible from above and should be easily accessible to aerial feeders.

Fall

Chukchi and Beaufort Seas

By the time ice formation begins in late September or early October, most seabirds have left the Arctic on their southward migration. The principal exception is the oldsquaw, which does not begin its migration until September. Some oldsquaws remain in nearshore waters until they are driven out by the formation of new ice. In contrast to the spring mortality, there are few records of extensive bird mortality in the fall due to lack of open water. One instance was reported for 1975, when nearshore waters froze early and flightless eiders were seen sitting on the ice near Pt. Lay in the Chukchi Sea. The birds were in a weakened condition, apparently due to their inability to obtain food (W. J. Wiseman, personal communication).

In the offshore waters the species associated with the pack ice in September are the same as those in August. In late September, however, ivory and Ross' gulls become the most common species at the ice edge in the Chukchi. Glaucous gulls and black guillemots are also associated with the advancing ice edge (Watson and Divoky 1972). Except for the Ross' gull, which apparently winters in the arctic basin, these species remain with the ice as it advances into the Bering Sea.

Bering Sea

Little is known about bird distribution in the Bering Sea during ice formation because cruises in rapidly forming ice are potentially hazardous. It is not known if the large numbers of birds found at the ice edge in March are present in December and January.

The principal effect of the arctic pack ice is to lower biological productivity and bird densities in the areas it covers. Unlike the antarctic pack ice, which supports a large biomass of pagophilic species, the number of pagophilic species supported by the arctic pack ice is small. Only the ivory gull, Ross' gull, and black guillemot have specific adaptations to the ice environment. The Ross' gull and guillemot winter in the pack ice, and the ivory gull is associated with ice throughout the year. The total biomass of these species is low. Other species which are regularly associated with the arctic pack, such as murres and black-legged kittiwakes, are also found in large numbers away from the ice. In addition, these species are usually associated with ice for limited periods during the year—murres primarily in winter and spring and kittiwakes primarily in summer.

The difference in the antarctic and arctic pack ice systems is largely due to the antarctic pack ice being surrounded by ocean, whereas the arctic pack ice is, in general, surrounded by land. The high productivity associated with the antarctic pack ice is due primarily to the mixing that occurs at the edge of the pack ice. There is little opportunity for mixing to occur next to the arctic pack ice, except where it is next to large expanses of boreal waters. This occurs in the Bering Sea in winter and spring, in the North Atlantic, and to a minor extent in the Chukchi Sea in summer and fall (Dunbar 1968). The limited geographic range and seasonal nature of high productivity at the arctic pack ice edge has been a major factor in preventing a well-developed pagophilic avifauna.

The importance of the in-ice algal bloom and its associated under-ice fauna is not yet clear. It is probably most important in areas such as the Beaufort Sea, where productivity in the water column is low. Although considerable numbers of seabirds are regularly found in the summer pack ice feeding on arctic cod and zooplankton associated with the ice, bird densities south of the ice are usually greater than those in the ice. The only species that appear to depend on the ice-associated fauna for much of their food are the three pagophilic species mentioned above.

---

by

James C. Bartonek[2]

U.S. Fish and Wildlife Service
Fairbanks, Alaska

and

Spencer G. Sealy

University of Manitoba
Winnipeg, Manitoba, Canada

Habitats for breeding marine birds are found along much of the 4,100-km coastline of Alaska that fronts on the Chukchi and Bering seas. Seasonal sea ice and an extensive outer continental shelf are dominant features that contribute to the productivity of these marine waters, which sustain populations of fishes, birds, and mammals that are of considerable and diverse values to man (Kelley and Hood 1974).

Our purpose in this paper is to describe the distribution, abundance, and relative status of some of the nearly 100 species of marine birds breeding within this region and the information base from which the descriptions are derived. Although the selection is admittedly arbitrary, we discuss mainly the colonial nesting species because they are generally in greater jeopardy from lost breeding habitat and from catastrophes than are the species that are widely dispersed or solitary in nesting. Because we believe matters affecting the conservation of marine birds will be geographically oriented, we discuss the status and distribution of breeding birds on that basis, rather than by the more traditional taxonomic approach. We use the terms "colony" and "colonies" somewhat loosely and interchangeably to include any aggregation of birds of the same or different species nesting in proximity to each other, even those on the same island or headland, although populations may be miles apart and occupy different kinds of habitats. The nature of this paper and the scale of our maps do not allow for detailed resolution of each colony's location (for the most part this information is not available), but rather facilitates a general impression of status.

Most place-names used by us are shown in Fig. 1; the others may be located by referring to Orth's (1967) gazetteer on Alaska.

There is no adequate catalog of marine bird colonies and other avian habitats for the Bering-Chukchi region or for Alaska as a whole. King and Lensink (1971) described the waterfowl populations and major lowland habitats of the State and listed only a few of the many colonies of cliff-nesting birds. LeResche and Hinman (1973) identified a few additional colonies, provided fragmentary information on composition and abundance at some of these sites, and delineated areas of wetland habitats on maps in their statewide atlas on wildlife. General and occasionally site-specific information on the location, but rarely on population size and composition, of colonies can be gleaned from the 321 species accounts presented by Gabrielson and Lincoln (1959) and from the general works by Bent (1919, 1921, 1922, 1923, 1925, 1927, 1929), Dement'ev and Gladkov (1951), Dement'ev et al. (1951, 1952), Palmer (1962), Fisher (1952), Tuck (1960), and others. The birds on the Asiatic side of these waters, which are not treated in this paper, were described by Portenko (1973).

Information on the status of waterfowl in the region is generally more detailed than that for most other groups of birds because waterfowl have been the object of systematic surveys since the late 1940's as part of the continent-wide effort to manage populations for sport hunting. Because the emphasis of these surveys has been directed toward the species of ducks important to hunters in the "lower 48" States, data are not adequate to measure changes in populations for most sea ducks and marine geese nesting in this region. These surveys have, however, enabled biologists to delineate waterfowl habitats and make reasonable estimates of populations for some of the more abundant and conspicuous species (King and Lensink 1971; U.S. Fish and Wildlife Service [FWS] 1973c; U.S. National Park Service [NPS] 1973).

Chukchi Sea Coast

A disproportionate percentage of ornithological investigations in arctic Alaska have centered about Barrow, where ornithologists were attracted because of the propensity of vagrant birds to collect there and because of the above average facilities, conveniences, and transportation afforded first by the whaling station, then by the military, and later by a research laboratory. Recent petroleum development near Prudhoe Bay has resulted in a somewhat commensal eastward shift in ornithological studies.

Bailey (1948), Gabrielson and Lincoln (1959), and Pitelka (1974) reviewed much of the published information on arctic avifauna, including that of the Chukchi coast. Selkregg [1975] mapped various avian habitats, ascribed either relative or absolute values for the population size of certain groups of birds, and included a selected bibliography that did not entirely duplicate those provided by the other reviewers. Watson and Divoky (1975) described the avifauna of Alaska's Beaufort Sea coast, which is much the same as that of the Chukchi coast from Point Barrow south to Cape Lisburne (both coasts are of low relief).

Intensive studies near Barrow have done much to characterize the behavior, productivity, and ecological requirements of calidridine sandpipers (Pitelka 1959; Pitelka et al. 1974; Holmes 1970, 1971) and, to partly explain the cyclical relationships between jaegers (Stercorarius spp.) and their prey (e.g., Pitelka et al. 1955; Maher 1974). Quantitative estimates of certain bird populations at Cape Thompson (Swartz 1966; Williamson et al. 1966), Little Diomede (Kenyon and Brooks 1960), and on the coastal lowlands of the Seward Peninsula (King and Lensink 1971; U.S. NPS 1973), and for black guillemots (Cepphus grylle) throughout the region (Divoky et al. 1974) are among the best data on status of marine birds for any locality in Alaska. Grinnell (1900a) described the birds he observed in the Kotzebue Sound area.

Cursory aerial surveys conducted by J. C. Bartonek, J. G. King, and D. R. Cline (U.S. FWS 1973a; U.S. NPS 1973; this paper) in 1972 and 1973 provided information on the location and relative size of most, if not all, colonies of cliff-nesting marine birds between Point Barrow and the Bering Strait, including those at Cape Lisburne, at Motherhood Point, Nine-mile Point, Cape Deceit, Towalevic Point, Sullivan Bluff, all on the northern base of the Seward Peninsula, and at Fairway Rock. The relative size of populations of most species was probably underestimated because the burrow-and crevice-nesting species were largely unseen.

Bering Sea

Aside from work by Gabrielson and Lincoln (1959) and the early but understandably incomplete accounts by Nelson (1883, 1887) and Turner (1886), no comprehensive description of the avifauna of the Alaskan coast of the Bering Sea exists. Many studies adequately describe local avifauna, and some of them are exemplary assessments of the status of populations.

Most of the coastline suitable for cliff-nesting marine birds and most of the smaller nearshore islands from the Bering Strait south to the tip of the Alaska Peninsula were reconnoitered piecemeal from aircraft between 1970 and 1973 by J. C. Bartonek, J. G. King, D. R. Cline, C. D. Evans, and M. L. Plenert (U.S. FWS 1973a, 1973b; this paper). In late June 1973 Bartonek, Cline, and Plenert made brief reconnaissances on foot of King, Besboro, and Shaiak islands. Bartonek and J. G. Divoky, traveling by boat and occasionally on foot, reconnoitered colonies at Cape Seniavin, a portion of the Walrus Islands group, Shaiak Island, and the coastline from Cape Peirce around Cape Newenham to Security Cove (U.S. FWS 1973a, 1973b; this paper). Although these cursory surveys (especially those from aircraft) tended to identify nesting sites of cliff-nesting birds while missing sites used by burrow-and crevice-nesting species, information was obtained on the location and relative size of many previously unreported colonies.

The mainland and island colonies in Norton Sound have received little notice in the published literature. Bailey (1943, 1948), although working mainly at Little Diomede and in Arctic and Lopp lagoons on the north side of the Seward Peninsula, mentioned the birds at Wales Mountain and Tin City. Nelson (1883, 1887) traveled throughout the region studying the avifauna and the anthropology of Eskimos. Grinnell (1900b) at Nome, McGregor (1902) along the Koyuk River, Hersey (1917) and Turner (1886) near St. Michael, and Cade (1952) at Sledge Island provide fragmentary examples of the area's marine bird populations. Colonies at King, Besboro, Egg, and Sledge islands, near York Mountains, and at Bluff were described in proposals for new National Wildlife Refuges (U.S. FWS 1973a).

Sealy et al. (1971) reviewed the literature and discussed the various zoogeographic relationships among the avifauna of St. Lawrence Island. Fay and Cade (1959) estimated numbers and biomass of all birds on St. Lawrence Island but did not identify locations and sizes of particular populations; consequently, replication of their estimates is precluded. An exemplary study by Bédard (1969) identified the locations and sizes of all populations of crested auklets (Aethia cristatella), least auklets (A. pusilla), and parakeet auklets (Cyclorrhynchus psittacula) on the island. Sealy (1973) identified breeding sites of horned puffins (Fratercula corniculata) there and throughout the species' range. Thompson (1967) listed the birds observed at Northeast Cape and on nearby Punuk Islands.

Annotated accounts have been published on the breeding avifauna of St. Matthew, Hall, and Pinnacle islands by Elliott (1882), Hanna (1917), Bent (1919), and Gabrielson and Lincoln (1959). Klein (1959) presented quantitative data on the birds he observed incidental to his study of reindeer (Rangifer tarandus).

The avifauna of the Yukon-Kuskokwim delta, which is rich both in numbers and diversity, has been treated extensively in the literature. Nelson (1883, 1887), Turner (1886), Conover (1926), Brandt (1943), Gabrielson and Lincoln (1959), Williamson (1957), Kessel et al. (1964), Harris (1966), Dau (1972), and Holmes and Black (1973) all described the avifauna in the same general area of the delta, i.e., the eroding portion in the general vicinity of Hooper and Hazen bays. The avifauna of the aggrading portion of the Yukon delta and of the Kuskokwim's mouth have not been accorded similar attention. Populations of waterfowl nesting on the delta and their wintering affinities were described by King and Lensink (1971) and U.S. FWS (1973c).

Studies of particular species of marine birds on the delta (again, all in the general vicinity of Hooper and Hazen bays) were reported by Hansen and Nelson (1957) and Shepherd (1960) for black brant (Branta bernicla), by Headley (1967) and Eisenhauer and Kirkpatrick (1977) for emperor geese (Anser canagica), by Dau (1974) and Mickelson (1975) for spectacled eiders (Somateria fischeri), by Petersen (1976) for red-throated loons (Gavia stellata), and by Holmes (1970, 1971, 1972) for dunlins (Calidris alpina) and western sandpipers (C. mauri).

Birds of Nunivak Island were reported by Swarth (1934), but the importance of the island to marine birds was not put into proper perspective until the Nunivak National Wildlife Refuge was evaluated for designation as a wilderness area (U.S. FWS 1972).

The Pribilof Islands have served as a focal point for ornithological investigations of the Bering Sea in much the same way that Barrow has for the Arctic. The avifauna of the Pribilofs has been described by Coues (1874), Elliott (1882), Palmer (1899), Hanna (1918), Preble and McAtee (1923), Gabrielson and Lincoln (1959), Kenyon and Phillips (1965), and a host of others that mainly added new species to the record list. Although most of these ornithologists marveled at the numbers of birds, information is lacking from which most changes in populations can be noted. (An exception is the record of common and thick-billed murres, Uria aalge and U. lomvia, which formerly nested in such abundance on Walrus Island that annually several tons of eggs were gathered for consumption by residents of the islands [Palmer 1899], but were greatly reduced in numbers by the summer of 1973, when J. C. Bartonek, J. G. King, G. J. Divoky, and D. T. Montgomery observed only a few thousand murres on a small portion of the island. Most of the suitable nesting sites, especially the flat areas often used by common murres, were occupied by Steller's sea lions, Eumetopias jubata, which, apparently because of reduced hunting pressure, occupied the island and displaced the murres.)

For some unexplained reason the numerous and large marine bird colonies along the north side of Bristol Bay appear to have been largely overlooked until recent years (Bartonek and Gibson 1972). Gabrielson and Lincoln (1959) summarized the few observations by Osgood (1904) and Turner (1886) in this area, but obviously were unaware that, in aggregate, these colonies rival those of the Pribilofs. Dick and Dick (1971) made an exemplary study of marine birds and their numbers at Cape Peirce and on nearby Shaiak Island. Murie (1959) provided annotated remarks on marine birds of Amak Island, but not of nearby Sealion Rocks.

Seven groups of colonies of cliff-, burrow-, and crevice-nesting birds are found on the headlands and islands in the coastal region, each numbering more than 1 million birds; nine colonies range downward to 100,000 birds; and a host of others range downward to 1,000 birds (Fig. 2). Un-estimated numbers of other marine birds nest on the lowlands about Kotzebue Sound, the Yukon-Kuskokwim delta, and Bristol Bay, but are not shown in Fig. 2. The occurrence at colonies of 20 of the nearly 100 species of marine birds is shown in Fig. 3; their relative numbers at these sites are not shown because data are generally lacking.

Chukchi Sea

The largest colonies of seabirds in the Chukchi Sea are those on Little Diomede Island, Cape Lisburne, Cape Thompson, and Fairway Rock. Smaller colonies are in Kotzebue Sound along the northern base of the Seward Peninsula. These colonies are largely dominated by thick-billed and common murres and black-legged kittiwakes (Rissa tridactyla) and on the islands in the Bering Strait also the crested, least, and parakeet auklets. Horned puffins, tufted puffins (Lunda cirrhata), pelagic cormorants (Phalacrocorax pelagicus), and glaucous gulls (Larus hyperboreus) make up the remaining majority. For the whole area there are probably fewer than a hundred birds each of black guillemots and pigeon guillemots (Cepphus columba) occupying colonies. Dovekies (Alle alle) are occasionally sighted in this area, but only as stragglers from their normal range.

Part of the mystery surrounding the nesting location of Kittlitz's murrelet (Brachyramphus brevirostris) was solved when Thompson et al. (1966) discovered a downy chick in the Kukpuk River drainage nearly 45 km by river from salt water. Other nesting sites of the Kittlitz's murrelets in this region were reported for Wales Mountain (Ford 1936; Bailey 1943, 1948) and the Cold Bay area (Bailey 1973) (Fig. 3).

Only the colonies at Cape Thompson have been censused systematically throughout a breeding season. During one of three years of varying census efforts, Swartz (1966) estimated that about 400,000 birds of nine species occupied the cliffs. Whereas the Cape Thompson colonies received considerable attention because of Swartz's efforts, the colonies that extend along nearly 35 km of headlands southward from, but mainly at, Cape Lisburne have received little if any attention by either early or recent ornithologists in the Arctic, even though they support perhaps twice the number of birds. Also perplexing is why Chamisso and Puffin islands with their several thousand nesting horned puffins and lesser numbers of other seabirds were designated as the Chamisso National Wildlife Refuge in the early 1900's when none of the many larger and more species-diverse colonies in the area received comparable recognition by and protection through refuge designation.

The lowlands on the north side of the Seward Peninsula produce fall flights of sea ducks that average 49,200 oldsquaws (Clangula hyemalis), 51,000 eiders (mostly common eiders, Somateria mollissima), and 26,700 scoters (mostly black scoters, Melanitta nigra) (King and Lensink 1971). Small populations of black brant and emperor geese breed in what outwardly appears to be excellent habitat, and King and Lensink (1971) speculated that subsistence hunting by local Eskimos is responsible for suppressing these populations.

Bering Sea

The largest concentration of nesting seabirds in the Bering Sea and perhaps in the entire North Pacific is that on St. George Island. Colonies that rank somewhere below that at St. George are along the coast from Cape Newenham to Cape Peirce, in the Walrus Islands (Round, High, Crooked, and Summit islands, The Twins, and Black Rock), at Cape Mohican on Nunivak Island, St. Matthew Island, Southwest Cape of St. Lawrence Island, and King Island.

The Pribilofs have the unique distinction of being the primary nesting site of red-legged kittiwakes (Rissa brevirostris). They are also interesting from the zoogeographic standpoint in that they are the northernmost stronghold of red-faced cormorants (Phalacrocorax urile); guillemots are conspicuous by their absence, and larid gulls are conspicuously scarce nesters.

St. Matthew Island and associated Hall and Pinnacle islands, and all but Walrus Island of the Pribilofs, are sites of nesting northern fulmars (Fulmarus glacialis). Nesting fork-tailed or Leach's storm-petrels (Oceanodroma furcata and O. leucorhoa) have been found nowhere in this region, although both are commonly observed at sea and both nest throughout the Aleutians.

Most colony sites identified in Fig. 2 are dominated by common or thick-billed murres (or both) and black-legged kittiwakes. Glaucous gulls (generally north of the Yukon-Kuskokwim delta), glaucous-winged gulls (Larus glaucescens) (generally to the south of the delta), and pelagic cormorants occupy almost every rocky prominence along the entire coast (most of these sites are not shown in Figs. 2 and 3). Double-crested cormorants (Phalacrocorax auritus) nest at a few island and inland locations in the Bristol Bay area. The small auklets are largely restricted to islands in the Bering Sea; the parakeet auklet is the only one occasionally found in mainland colonies.

The marine birds of the Yukon-Kuskokwim delta lowlands, although largely uncounted, in their aggregate probably exceed the numbers at any individual site identified in Fig. 2. This is not particularly surprising since the delta has nearly 70,000 km² of habitat (King and Lensink 1971) in contrast to the generally small parcels of habitat occupied at the sea-cliff and island sites.

King and Lensink (1971) estimated that fall flights of sea ducks originating on the delta averaged 292,300 oldsquaws, 51,000 eiders (mostly common and spectacled eiders with lesser numbers of Steller's eiders, Polysticta stelleri), and 157,000 scoters (primarily black scoters). They also estimated that half of the 150,000 black brant and most of the 150,000 emperor geese in Alaska's fall flight originate there. Although no counts have been made, we believe that the delta's lowlands support easily more than half of Alaska's nesting dunlins, black turnstones (Arenaria melanocephala), rock sandpipers (Calidris ptilocnemis), western sandpipers, and substantial percentages of red phalaropes (Phalaropus fulicarius), northern phalaropes (Lobipes lobatus), and red-throated loons.

The north side of the Alaska Peninsula (including the wetlands, uplands, and estuaries) is perhaps more important to marine birds as a staging, feeding, and resting area than as a nesting habitat. The importance of Izembek Lagoon to black brant and emperor geese during fall and spring is a classic example. King and Lensink (1971) estimated that the fall flight of sea ducks originating from the Peninsula averages 53,400 oldsquaws, 1,700 eiders, and 74,400 scoters. Breeding geese are scarce throughout the area.

Most of the major breeding habitats of marine birds in the Chukchi and Bering seas are known, but imprecisely identified as to location and size. With few exceptions, the populations of birds using these habitats are described only by the subjective and ambiguous descriptors of abundance such as "abundant, common, occasional, and rare," which makes measurement of change impossible.

We recommend that first and foremost a catalog of habitats used by birds be developed to aid resource administrators, developers, and biologists (all of whom should be "conservationists") in identifying critical habitats. We believe that such a catalog would preclude many problems because birds and their habitats could be considered at the planning stage rather than only at the operational stage. Such a catalog would also be useful to students of ornithology who are seeking locations suitable for particular studies.

Nowhere in this region have studies of marine birds been of sufficient duration to enable changes in populations (from whatever cause) to be characterized. Since some species of marine bird are known not to breed before at least 3 or more years of age, meaningful information on survival and recruitment in populations cannot be obtained by studies of less than 10 years. We therefore recommend that long-term studies be initiated at as many places as possible, but at least at one site on the Yukon-Kuskokwim delta; at a mainland colony site that has predominantly murres, kittiwakes, puffins, and cormorants; and at an island site that also has small auklets. Although the nesting distribution of the Kittlitz's murrelet remains an enigma, we regard it less of a conservation issue and more of an ornithological challenge. Consideration of logistics and support facilities must, of course, be included in the site selection process. Most of the areas suggested for these studies also merit recognition and protection by being designated as a National Wildlife Refuge, a National Park or Monument, or a State Game Sanctuary.

---

by
Palmer C. Sekora[3]

U.S. Fish and Wildlife Service
Kailua, Hawaii

G. Vernon Byrd[4]

U.S. Fish and Wildlife Service
Adak, Alaska

and

Daniel D. Gibson

University of Alaska
Fairbanks, Alaska

The 1,800-km-long chain of islands known as the Aleutians provides nesting habitat for various species of marine birds, including three species of Procellariiformes and three of cormorants (Phalacrocorax spp.), one species of gull (Larus glaucescens), both kittiwake species (Rissa spp.), two species of terns (Sterna spp.), and at least 13 species of alcids.

Seabird population estimates of known accuracy are lacking for this isolated area. Locations of larger colonies of breeding seabirds are known, however, and sufficient data are available to place colonies in broad size ranges. Published information on the breeding biology of marine birds is also lacking from the Aleutians, but some studies are under way. The distribution of nesting marine birds away from the nesting cliffs is totally unknown.

Introduced predators, primarily arctic foxes (Alopex lagopus), are now found on nearly every island. Breeding marine bird populations have suffered drastic reductions as a result. They have probably also changed because of natural habitat modifications caused by earthquakes, volcanic eruptions, tidal waves, and marine erosion.

The purpose of this paper is to summarize the known present distribution and status of breeding marine birds in the Aleutian Islands.

Description of the Aleutian Islands

The Aleutian Islands form an arc that separates the Bering Sea and the north Pacific Ocean (Fig. 1). The island chain extends from the tip of the Alaska Peninsula to within 483 km of the Commander Islands of Siberia. The chain contains more than 200 islands—the peaks of a submarine volcanic mountain range. Volcanic activity and earthquakes occur regularly.

Weather is characterized by perpetual overcast, dense summer fog, high-velocity winds, and mild temperatures with low annual and diurnal variations. The sea is ice-free year-round except in extremely cold winters, when the arctic ice pack may reach the extreme northern islands.

The Aleutians are treeless except for a few introduced, stunted spruces. Woody shrubs are restricted to the most northern islands on each end of the Chain. Mosses, lichens, club mosses, and heaths are common ground-cover plants, and taller grasses, sedges, and umbellifers constitute the overstory. Hulten (1960) provided a list of terrestrial plants found in the Aleutians. Amundsen and Clebsch (1971) discussed terrestrial plant ecology at Amchitka, central Aleutians. The marine plant communities around the islands are fairly diverse. Lebednik et al. (1971) described marine algal communities at Amchitka.

The easternmost Aleutian island, Unimak, has a mammalian fauna like that of the Alaska Peninsula, including brown bear (Ursus arctos), caribou (Rangifer tarandus), wolf (Canis lupus), and wolverine (Gulo gulo). West of Unimak, red foxes (Vulpes fulva) occurred historically as far as Umnak, and arctic foxes were apparently on Attu when the Russians came in 1741 (Murie 1959). Except for man and dog, no land mammals occurred between Umnak and Aggatu islands. Arctic foxes, introduced before 1930 for fur farming, still roam almost every island. Norway rats (Rattus norvegicus) were introduced on many islands when ships were wrecked or as a result of military activities during World War II.

Sea otters (Enhydra lutris) have repopulated most of the Aleutians after being nearly extirpated by 1900. Rookeries of Steller's sea lion (Eumetopias jubata) are scattered throughout the Aleutians during summer, and numerous harbor seals (Phoca vitulina) haul out on beaches and offshore rocks.

All five species of Pacific salmon (Oncorhynchus spp.) occur near the islands, and at least four of them (all but O. tshawytscha) spawn in Aleutian streams. Dolly Varden (Salvelinus malma) and three-spine sticklebacks (Gasterosteus aculeatus) are found nearly everywhere there is fresh water. The marine environment provides habitat used by at least 77 species of fish (Isakson et al. 1971). O'Clair and Chew (1971) furnished a recent reference to littoral macrofauna at Amchitka.

About 200 species of birds have been recorded in the Aleutians (Aleutian Islands National Wildlife Refuge, unpublished data). Many of these are windblown stragglers from both North America and Asia; only 59 species breed on the islands. Although seabirds make up less than half (26 species or 44%) of the breeding birds, they may compose more than 90% of the breeding avian biomass.

Published ornithological information from the Aleutian Islands is relatively scarce. G. W. Steller, naturalist on Vitus Bering's 1741 expedition to Alaska, was the first person to record ornithological information in the islands (Stejneger 1936). More than a century passed before W. H. Dall (1873, 1874) published the next papers dealing with birds in the Aleutians. In 1878, the U.S. Army Signal Corps sent L. M. Turner to the Aleutians to set up weather stations at several locations. Turner kept notes on birds at various locations in the Aleutians and published two papers (1885, 1886) on his observations. Turner's data (1886) provided the first report based on extended and widespread observations in the area. E. W. Nelson, who replaced Turner, also provided data on birds (Nelson 1887).

In 1906, A. C. Bent came to the Aleutians specifically to look for birds, and he and Alexander Wetmore recorded birds throughout the island chain (Bent 1912). A. H. Clark (1910) provided a valuable record of his observations in the Near Islands. All these workers recorded birds in several locations, but none provided data on more than a very few seabird colonies.

O. J. Murie, U.S. Biological Survey, made the most complete survey of the Aleutians (Murie 1959). He specifically recorded seabird colonies, spending parts of four summers in the area. Murie visited every large Aleutian island and most small ones. He recorded nearly every major colony of cliff-nesting or talus-nesting seabirds known in the Aleutians, but seldom gave sizes of colonies, and separate colonies on a particular island were often not differentiated.

World War II brought several ornithologists to the Aleutians. Cahn (1947), Sutton and Wilson (1946), Taber (1946), and Wilson (1948) provided accounts of birds observed at specific locations. After the war, Fish and Wildlife Service personnel—including I. N. Gabrielson (Gabrielson and Lincoln 1959), K. W. Kenyon (Kenyon 1961), and R. D. Jones (Refuge Narrative Reports 1949-1970)—recorded observations of breeding seabirds at several locations in the Aleutians. Investigations associated with Atomic Energy Commission nuclear testing at Amchitka Island provided the first ecological study of avifauna of an Aleutian island (White et al. 1977). Byrd et al. (1974) provided a list of birds at Adak.

In 1971, the Near Islands were surveyed by U.S. Fish and Wildlife personnel in a Cape Cod dory. In 1972, the Aleutian Islands National Wildlife Refuge obtained a vessel, the Aleutian Tern, which allowed visits to all parts of the island chain. That year, nearly every large island as far west as Buldir was visited, and seabird colonies were mapped. Every island has been visited at least once since 1972.

In estimating the current status of seabirds in the Aleutians, all available data were considered. Most of the information used, however, is from surveys conducted by the U.S. Fish and Wildlife Service (1970-75, unpublished data). Because these surveys only incidentally included Unimak, Akun, Akutan, Unalaska, and Umnak islands, data for these areas are almost totally lacking. Data for Bogoslof, Adak, Amchitka, Buldir, Agattu, Nizki, Alaid, and Attu are most accurate because fairly intensive investigations have been conducted there since 1970.

The available data are of unknown accuracy. The method used by most investigators who have surveyed areas in the Aleutians for seabird colonies has been to circle islands in a ship or small boat; when a colony was encountered, they simply estimated the number of birds they saw at the time. The accuracy of the estimates is affected by weather, distance from the colony, density of birds, ability and experience of the observer, and other variables. Estimates of kittiwakes and cormorants should be the most accurate, since nests were actually counted. Murres (Uria spp.) are readily visible on the cliffs, but the percentage of breeders on the cliffs at a particular time of day during a particular part of the breeding season is not known. Auklet numbers are perhaps hardest to estimate, since swirling "clouds" of birds are encountered.

Even when the estimates of birds seen are assumed to be accurate, data interpretation is complex. Lack of information on diurnal rhythms adds difficulty to data interpretation. Counts of burrow-nesting birds (e.g., puffins) have been inaccurately interpreted because of the lack of understanding of their nesting ecology. Gulls (Larus spp.), terns, and jaegers (Stercorarius spp.) are not well known since shore parties have seldom investigated island interiors. Nocturnal species (e.g., ancient murrelet, Synthliboramphus antiquus, and storm-petrels, Oceanodroma spp.) are perhaps the least known. Since only crude estimates of colony sizes are available, broad limits are used in this paper to describe known colonies.

Even from the sparse literature available, it is apparent that some seabird populations are now drastically different from those in the Aleutians around 1900. Changes in nesting habitat due to volcanic eruptions, tidal waves, marine erosion, and earthquakes have occurred for centuries, and colonial nesting bird populations have fluctuated accordingly. In addition, native Aleuts used marine birds and their eggs for food and their skins for clothing, but the Aleuts were so diminished in numbers by 1900 that they have had little recent effect on the bird populations.

From about 1900 to 1936, arctic foxes were introduced to most of the Aleutians for fur farming. The foxes lived on birds in summer, and some species (e.g., Aleutian Canada geese, Branta canadensis leucopareia) were wiped out wherever foxes were introduced. Ground-nesting and some burrow-nesting seabirds were also drastically reduced or extirpated on many islands.

During World War II the thousands of troops in the Aleutians brought dogs and cats to some of the islands as pets, and many of the animals were set free when the men departed. The military also accidentally introduced Norway rats to some of the islands. Their role in seabird population reductions is unknown.

Figures 2-15 (pages 40-46) present data on the distribution of populations of birds that have survived the foxes and other introduced predators. An annotated list of seabirds breeding in the Aleutians follows.

Annotated List of Species

Northern fulmar (Fulmarus glacialis)

Northern fulmars breed on only three islands: Buldir (200 pairs), Gareloi (1,500 pairs), and Chagulak (more than 100,000 pairs). Fulmars were apparently much more widespread formerly (Murie 1959; Turner 1886). Introduced foxes were probably involved in the decline.

Fork-tailed Storm-petrel and Leach's Storm-petrel (Oceanodroma furcata and O. leucorhoa)

The distribution of storm-petrels is poorly known due to their nocturnal behavior near the nesting colonies. The presence of birds has generally been noted by finding them aboard ships anchored near islands after darkness. Population estimates are not available for any colonies, so symbols used in Fig. 3 indicate probable numbers of breeding birds. In few cases have active burrows or crevices been discovered. Storm-petrels were formerly much more common. Murie (1959) and John L. Trapp (personal communication) found large numbers of storm-petrel remains in fox dens. Most present breeding colonies are probably confined to offshore islets and fox-free islands.

Double-crested Cormorant, Pelagic Cormorant, and Red-faced Cormorant (Phalacrocorax auritus, P. pelagicus, and P. urile)

Double-crested cormorants breed as far west as the Islands of Four Mountains. The colonies vary in size from a few to 25 pairs. Pelagic and red-faced cormorants nest from Amak to Attu on nearly every island. Relative abundance of the two in mixed colonies varies between areas as well as from year to year. Red-faced cormorants tend to nest in colonies mixed with kittiwakes and murres, but pure colonies also occur. Pelagic cormorants occupy isolated, small colonies, but they also nest with kittiwakes and murres and are often found with red-faced cormorants. By far the densest concentration of cormorants occurs in the Near Islands, especially at Attu, where an estimated 77,000 birds were seen in 1970. In the Aleutians as a whole, red-faced cormorants outnumber pelagic cormorants, and double-crested cormorants make up only a very small percentage of the breeding population.

Parasitic Jaeger (Stercorarius parasiticus)

The distribution of jaegers is poorly known because investigators have spent little time ashore on most islands. Murie (1959) found jaegers on a number of islands, and most of the data in Fig. 5 are his. Population estimates are available only for Amchitka (25 pairs; White et al. 1977) and Buldir (30-40 pairs; G. V. Byrd, unpublished data).

Glaucous-winged Gull (Larus glaucescens)

Glaucous-winged gulls no longer nest on islands where foxes occur except where islands in lakes are available. Most colonies are on offshore rocks or islets and range in size from a few pairs to over 200 pairs, and occasionally more. They are found throughout the Aleutians, but the largest known colonies are at Bogoslof (500 pairs) and Buldir (250 pairs).

Black-legged Kittiwake and Red-legged Kittiwake (Rissa tridactyla and R. brevirostris)

Black-legged kittiwakes breed locally in every major island group, usually mixed with murres and cormorants. The large colonies contain over 25,000 birds, but colonies of less than 50 pairs also occur. Red-legged kittiwakes breed only on Buldir and Bogoslof. They are remnants of a previously more widespread population.

Arctic Tern and Aleutian Tern (Sterna paradisaea and S. aleutica)

Terns breed locally in each island group. Both species occur at Attu, Amchitka, Adak, and Umnak, but only arctic terns are found at Nizki. Factors limiting distribution are unknown. Colonies vary in size from less than 10 pairs to 100 pairs.

Common Murre and Thick-billed Murre (Uria aalge and U. lomvia)

Like kittiwakes, murres are abundant locally. A pure colony of either species is almost unknown, although one species often makes up more than 90% of a colony. Common murres may have been reduced by foxes, since they tend to use sites with less slope than those used by thick-billed murres. At Bogoslof and the Baby islands, the birds use inland, gently sloping areas because there are no foxes. The presence of the lichen (Caloplaca spp.), which according to Tuck (1960) is indicative of bird roosts, on several extensive cliff areas suggests that either murres or kittiwakes, or both, formerly used areas they do not use now.

Pigeon Guillemot (Cepphus columba)

This species has been noted near almost every island that has been visited. Nesting under beach boulders and driftwood, the birds only occasionally are found in large concentrations (near Great Sitkin more than 4,000 birds were seen in 1971). Murie et al. (1937) summed up the distribution of pigeon guillemot accurately: "Each island has its meager quota of these birds, nesting unobtrusively among the rocks but never assembled in any really large groups." Estimates of populations may be extremely inaccurate because the diurnal rhythm of the pigeon guillemot is unknown.

Marbled Murrelet and Kittlitz's Murrelet (Brachyramphus marmoratus and B. brevirostris)

Nests of neither species have been located in the Aleutians, but nesting of both is suspected at Adak, Unalaska, and Unimak, where specimens of Kittlitz's with brood patches or eggs in the oviduct have been collected in nearshore waters. Courtship has been recorded in marbled murrelets (Byrd et al. 1974).

Ancient Murrelet (Synthliboramphus antiquus)

The distribution of this species is very poorly known, since it is nocturnal near nesting colonies. Murie (1959) wrote, "This is one of the species that undoubtedly has greatly declined in recent years, as a result of increase of the blue-fox industry." The leading of downy young to sea by the adults is a very noisy process and foxes could easily take large numbers. Also, these murrelets nest in fairly shallow burrows which foxes could dig out easily. Birds were recorded near islands in every group during surveys from 1972 to 1975, but workers seldom went ashore to determine if they were nesting. In Fig. 12, the only basis for designating most of the areas marked as colonies is the presence of birds during breeding season (15 May-1 July).

Cassin's Auklet (Ptychoramphus aleuticus)

This is another species that was more common before the fox was introduced. Cassin's auklet now seems to occur only locally, but these nocturnal birds are probably often overlooked. They are known only from Buldir, Umnak, and the vicinity of Oglodak.

Parakeet Auklet (Cyclorrhynchus psittacula)

This auklet, which nests under beach boulders, in burrows, and in rock crevices, seems to use a greater variety of breeding sites than do the other auklets. The largest known colony is at Chagulak, where an estimated 10,000 were seen in 1972. Smaller colonies are found as far west as Buldir.

Crested Auklet, Least Auklet, and Whiskered Auklet (Aethia cristatella, A. pusilla, and A. pygmaea)

Aethia nest primarily in rock crevices of talus slides. Such habitat occurs locally in each major island group except the Near Islands. Least auklets outnumber crested auklets in the Aleutians, and whiskered auklets are far less common than either. Estimates of populations are probably grossly inaccurate because of the difficulty both in estimating the number of birds in the milling flocks observed and in interpreting the estimates after they are obtained.

Horned Puffin and Tufted Puffin (Fratercula corniculata and Lunda cirrhata)

Horned puffins favor rock crevices in talus slides and cliff faces for nesting, whereas tufted puffins are primarily burrow nesters. The historical distribution of the two species was probably based on availability of nesting sites, so tufted puffins were more widespread and numerous. However, in areas where extensive talus slopes are available, horned puffins reached high densities. Predation by introduced foxes may have altered the distribution of tufted puffins, which now nest primarily on fox-free islets just offshore from the larger islands where foxes occur. The distribution of horned puffins may not have been altered significantly, since they are relatively free from fox predation in their rock crevices.

A complete survey of the Aleutian Islands has not been done. This should be done, by methods that will provide accurate population estimates. Life history information is needed on almost all species, and data should be gathered on selected populations to determine trends. Information on winter distribution should also be compiled. The effects of introduced predators should be evaluated quantitatively, and if control measures are needed, effective, humane methods should be devised and implemented.

The authors are grateful to the following Fish and Wildlife Service personnel who helped collect previously unpublished data used in this paper: E. P. Bailey, C. S. Craighead, C. P. Dau, M. H. Dick, G. J. Divoky, R. Martin, J. L. Trapp, G. W. Watson, and C. M. White. Most of the data were collected from the deck of the Aleutian Islands National Wildlife Refuge research vessel Aleutian Tern. Captain George Putney is acknowledged for his peerless seamanship and constant encouragement; he also contributed observations of birds.

The maps were adapted from a master supplied by Elaine Rhode, Public Affairs Office, U.S. Fish and Wildlife Service, Anchorage; she also suggested the use of squares to display data. C. M. White graciously made his in-press manuscript available. W. B. Emison, R. J. Gordon, and J. L. Trapp kindly made their field notes available, and Trapp helped compile data. Most of the funds for the surveys in 1971-75 were provided by the U.S. Fish and Wildlife Service.

---

by

LeRoy W. Sowl

U.S. Fish and Wildlife Service
1011 East Tudor Road
Anchorage, Alaska 99507

From the perspective of history, 1970 should prove to have been a momentous year for Alaska and its seabirds. Two events, the construction of the Trans-Alaska Pipeline and the passage of the National Environmental Policy Act (NEPA) merged head on in 1970 with the decision that Section 2c of NEPA applied to the proposed pipeline. The systematic appraisal of potential environmental impacts required by Section 2c quickly exposed the inadequacy of the existing data base in many areas. With respect to seabirds in the Gulf of Alaska, it was apparent that there had never been any effort to develop a synthesis of the information accumulated over 230 years. The data gaps which were uncovered were appalling.

While the Trans-Alaska Pipeline impact statement had provided shock therapy, it was not the only influential event on the horizon. Two local disturbances had already preceded the pipeline. These were Project Chariot at Cape Thompson and the Amchitka Island test program. Now in quick succession the Wilderness Act and native land claims added new urgency to the need for solid resource information. More recently, the outer continental shelf minerals leasing program has made the quick development of base-line information even more essential.

All of the new activity in Alaska's coastal waters has the potential to affect seabirds in one way or another. We must remember, however, that man's activities have been affecting seabirds for a long time. We cannot accurately assess the effect of a tanker terminal at Valdez or offshore oil activity without first developing some understanding of the current status of seabirds in the context of the historical record.

Seabird work in Alaska can be divided roughly into three periods. The first is the early historical or exploratory period; it extended from Georg Steller's 1741 visit to Kayak Island to 1935. This was literally a period of exploration and the collection of information was dependent upon interest and opportunity. The second is the reconnaissance period; during this period investigators were dispatched to a particular area to gather general information for management application. This period begins with Murie's extensive investigations of the Alaska Peninsula and the Aleutian Islands; I see it extending from 1936 to 1975. In 1975 the need for data became so acute that it was necessary to enter the third period, one of intensive data gathering. Knowing where the big seabird colonies were located and knowing their general species composition was no longer adequate. The current intensive data-gathering effort in the waters over oil and gas leasing areas is a partial response to the recognition of this inadequacy.

In this paper I draw some tentative conclusions relative to the status of the 26 species of primary seabirds (Fisher and Lockley 1954) breeding in, or which may have bred in, the northern and western Gulf of Alaska area. This area extends from Cape Fairweather, 59°N 138°W, westerly along the coast to Ikatan Bay, 55°N 163°W, at the end of the Alaska Peninsula. These bird species tend to be colonial, but not exclusively so. Two birds which are primary seabirds, the mew gull (Larus canus) and Bonaparte's gull (L. philadelphia), have not been included because they tend to be more riverine than marine in habit. Several marine ducks have been excluded because they are secondary seabirds.

Information from the early exploratory period is summarized under the next section. The more detailed information from the reconnaissance period is discussed in the species accounts.

The history of ornithological field work in the Gulf of Alaska goes back 235 years to 20 July 1741. On that day Bering's surgeon/naturalist, Georg W. Steller, spent a scant 10 h ashore on Kayak Island. He collected a single bird. This bird, later named for Steller, reminded him of a plate of the blue jay by Make Catesby, the colonial-era predecessor of Audubon, in Volume 1 of the Natural History of Carolina, Florida, and the Bahama Islands (Stejneger's annotated translation of Steller's journal in Golder 1925). Collection of the bird confirmed for Steller that the first Russian Expedition had reached America.

Steller was an accomplished naturalist, but his overbearing and superior manner had apparently sorely irritated Bering and his officers long before the expedition reached Kayak Island. The seamen made little effort to go ashore anywhere in Alaska and Steller was blocked from doing so as well. In addition to Kayak Island, he was able to go ashore only on Nagai Island, first with a water party on 30 August and again the next day. He noted that "all sorts of waterbirds in abundance were seen." These included two kinds of cormorants, auks, ducks, gulls, divers, pigeon guillemots (Cepphus columba), tufted puffins (Lunda cirrhata), and horned puffins (Fratercula corniculata).

Stejneger's comment on the identity of the cormorants is interesting because, based on his experience, he assumed them to be pelagic and double-crested cormorants (Phalacrocorax pelagicus and P. auritus). He gave no thought to red-faced cormorants (P. urile) which are now common there.

Steller noted on 6 September off Bird Island in the Shumagin Islands, that "when we were out to sea about half a mile we were especially astonished at the untold numbers of seabirds which we saw on the northern side of the island." These birds were listed as cormorants, auks, horned puffins, fulmars (Fulmarus glacialis), pigeon guillemots, black oystercatchers (Haematopus backmani), and a pied diver which Stejneger assumed was an ancient murrelet (Synthliboramphus antiquus).

On 15 September when Bering's vessel, the St. Peter, was south of Amukta Pass, Steller recorded observing "river gulls." The observation is not as interesting as Stejneger's comment (Golder 1925) concerning it. Stejneger stated that no true river gulls lived in the Aleutians and these must, therefore, have been another small gull with red feet. He thought they must have been the red-legged kittiwake (Rissa brevirostris), which "inhabits the Aleutian Islands from Bering Island to Sannak."

Thirty-seven years after Bering's voyage, Captain James Cook sailed into the Gulf of Alaska, arriving off Kayak Island on 11 May 1778. Cook was not accompanied by an able naturalist. His surgeon, William Anderson, did have some experience gained on earlier voyages in preparing skins and taking notes, but he had contracted tuberculosis and became so ill that even his notes ceased after 8 June, while the expedition was in Cook Inlet.

Cook was under orders to keep a careful record of everything he saw. One of the results was that he had birds collected even though he had no naturalist to do the work. Several birds were collected in Prince William Sound while Cook's vessels were at anchor in Port Etches. These included two marbled murrelets (Brachyramphus marmoratus—type specimens), a black oystercatcher, a surfbird (Aphriza virgata), a surf scoter (Melanitta perspicillata), and a red-breasted merganser (Mergus serrator—type specimen), along with several forest birds (Stresemann 1949).

The watch journals of Cook and his officers provide some additional information. Captain Charles Clerke (Beaglehole 1974) remarked in his log on the passage out of Prince William Sound through Montague Strait on 20 May that "it had almost become tautology to mention whales and seals and innumerable sea fowl that so confoundingly kept their distance."

Between the Trinity Islands and Chirikof Island on 18 June, Cook's men collected a single tufted puffin. Later Cook passed close to the Semidi Islands and the Shumagin Islands and directly through the Sandman Reefs. Beaglehole's version of this part of the voyage makes no mention of seabirds.

There is a gap of 87 years during which there is almost no hint of published material bearing on the status of seabirds in the Gulf of Alaska. In 1865 the Russo-American Telegraph Expedition touched this area. Dall and Bannister (1869) provide us with a few scraps garnered during that expedition, primarily by Bischoff. The glaucous-winged gull (Larus glaucescens) was described as the most common species from California northward. Bischoff's collections at Kodiak indicate that the horned and tufted puffins were collected with ease. He was able also to collect an Aleutian tern (Sterna aleutica—type specimen) along with an egg.

Dall (1873) noted in 1872 that the black-legged kittiwake (Rissa tridactyla) was common at Round Island and Delarof Harbor, Unga Island, in the Shumagins. The inference is that it was more common at these two places than elsewhere. The Arctic tern (Sterna paradisaea) was abundant in the Shumagin Islands and particularly at Range Island in Popoff Strait. Dall expressed the opinion that the horned puffin was very abundant in the Shumagins and appeared to fill the niche of the tufted puffin, which he did not see there. The only other bird which he thought to be very common was the pigeon guillemot. He did not note the common murre (Uria aalge) at all.

In 1908 the second of three Alexander Expeditions conducted field work in the Prince William Sound area. From Dixon (1908) and Grinnell (1910) we can derive some basis for assessing status in a very general way. The most common seabird noted was the marbled murrelet. Glaucous-winged gulls and black-legged kittiwakes were common; the glaucous-winged gull was the more common. Horned puffins were judged to be slightly more common than tufted puffins by both authors. The northern end of Montague Strait appears to have been the center of abundance for puffins. Dixon noted that on 16 July 1908 there were swarms of puffins in the channel along Green Island. Pigeon guillemots were common along the rocky coasts. Parakeet auklets (Cyclorrhyncus psittacula), common murres, and ancient murrelets were noted only in very small numbers.

After the Alexander Expeditions there was another doldrum in which little was done. During this lull in activity, a note by Townsend (1913) appeared which compared the numbers of crested auklets (Aethia cristatella) at Yukon Harbor, Big Koniuji Island, to the least auklets (A. pusilla) of St. George Island, stating that the crested auklets were more numerous. He sailed into the Yukon Harbor anchorage on the evening of 1 August and observed that crested auklets "were present in myriads. The surface of the water was covered with them, and the air was filled with them."

The formal record available to researchers is very shallow for this exploratory period. With a few exceptions it was compiled by non-scientists, primarily explorers and egg and skin collectors.

Setting the Stage

This paper should be viewed as a conceptual model. While I attempted to be as objective as possible, subjectivity was unavoidable. Many of the tentative conclusions are based on very little data. Each improvement will make it a better management tool. Because of the space limitations, it is not possible to go into a detailed tracking of my reasoning for each species. In an attempt to overcome this handicap, I am including some examples of the sorts of reasoning that went into the process.

In 1973 I led a Fish and Wildlife Service (FWS) reconnaissance survey team that was delineating seabird colonies along the Alaska Peninsula. In the Shumagin Islands we entered or crossed Koniuji Strait twice (on 11 and 12 June) without even suspecting the presence of a horned puffin colony. A third passage through the strait (13 June) was not so uneventful. The water and the air were filled with horned puffins. This led to the discovery that the 430-m mountain on the southeastern corner of Big Koniuji was also covered with horned puffins, clear to its top. The minimum estimate of the birds that were visible was 140,000. Even this number of birds would make this the largest horned puffin colony ever discovered. David Spencer (personal communication) had noted similar swarms of horned puffins in this strait in 1956 while flying sea otter surveys in the area. In 1975 a field camp was established at Yukon Harbor, with study of this colony as one of the prime objectives of the investigators. As far as these investigators could tell no such large colony existed there, even though the nesting habitat was still there, unaltered. This sort of event, one of the banes and vagaries of estimating seabird numbers, is not rare.

In 1973, when FWS personnel delineated the colony on the southwestern end of Bird Island in the Shumagins, there were estimated to be 43,000 kittiwakes, 24,000 murres, and 6,000 cormorants present; no tufted puffins were seen about the colony. The last time (in 1970) one of the observers, Edgar Bailey, had visited the colony with Robert Jones, there was an extremely large colony of tufted puffins which Jones (E. Bailey, personal communication) estimated at more than 1 million birds. We made a particular effort to visit Jude Island, between the Shumagin Islands and the Pavlof Islands, because David Spencer (personal communication) had reported once having seen the air over the island filled with an extremely large number of tufted puffins. However, there were no puffins at this colony either.

Let us examine the facts in context. On 8 June we had visited High Island where we had attempted to collect puffin eggs for pesticide analysis, but had been able to find only one egg. Also, there were only 6,000 tufted puffins where George Putney, master/engineer of the Aleutian Tern, had seen much larger numbers in 1972. These two facts could easily be related to explain the current situation because it was still early in the breeding season. The horned puffin observations in Koniuji Strait (11-13 June) were in keeping with this conclusion also—an indication that these birds had not yet settled down to a full breeding effort. The erratic comings and goings of common puffins (Fratercula arctica) early in the season have been well documented (Lockley 1962). It is an easy step to extend this reasoning to the absence of birds at Bird Island on 11 June, even though fresh signs of the characteristic evidence of tufted puffin occupancy were missing. Jude Island provides a different clue, however. There were 3,000 pigeon guillemots, an unheard-of concentration, apparently occupying abandoned tufted puffin burrows on 15 June. Also, on 7 June we had made a very interesting observation that had no special significance at the time: murres on Spitz Island were occupying little parapets created by mashing down the mouths of puffin burrows which filled the slope above the cliff portion of their colony.

After looking at all of the observations cited above, I conclude that tufted puffins were greatly reduced in numbers on these sites in 1973 and that they had been absent from the burrows used by the murres and pigeon guillemots for more than the current breeding season. What causes these sorts of changes? I do not know.

One reason for year-to-year change may be local movements of colonies. Black-legged kittiwakes nest at several places in lower Orca Inlet, Prince William Sound. Counts made at these sites in 1972 and 1974 yielded almost identical totals but the numbers of birds varied between individual sites. This may be an indication that all of these sites are part of one large composite colony and that, at least in this colony and for this species, the birds shift at will.

The best record of population flux involving two species has been summarized by Peterson and Fisher (1955). In 1872 and 1873 the murres observed on Walrus Island in the Pribilofs were almost entirely common murres. In 1890 common and thick-billed murres (Uria lomvia) were evenly matched in number. By 1901 the colony was almost exclusively dominated by thick-billed murres. In 1911 and 1914 the few thick-billed murres present were almost lost among the then dominant common murres. In 1940 thick-billed murres dominated again. When Peterson and Fisher visited the island in 1953, the situation was again reversed and common murres had almost completely replaced the thick-billed murres. These changes are even more impressive because of the number of birds involved, between 1 and 2 million in 1953. There are more tenuous indications that somewhat the same thing may occur between two other congener pairs, the pelagic and red-faced cormorants and the black-legged and red-legged kittiwakes. The causative factor, or factors, is not readily apparent. One possibility is long-term climatic fluctuation.

Dement'ev and Gladkov (1966) provide an example of abrupt and massive change. Before 1876, the pelagic cormorant abounded on the Commander Islands. During the winter of 1876-77, the birds were decimated by an unknown epizootic disease. By spring only a few individuals remained alive. The record shows that by 1882 they were already becoming common again. Red-faced cormorants were apparently not reduced in number because Dement'ev and Gladkov (1966) state that they were common in "the second half of the last century and the beginning of this." Did they flourish only while the pelagic cormorants were reduced in number?

Bowles (1908) gives another indication of naturally induced population impact. He noted large numbers of dead seabirds on Washington beaches and the ocean "rather plentifully dotted with sick birds ..." He examined some birds and found "many hundreds" of tapeworms in every bird. His conclusion was that their intestines were so solidly packed with tapeworms that starvation was "an absolute certainty."

Some apparent disruptions are long term. In the Gulf of Alaska there is a hiatus in the distributions of a number of small seabirds that are active around their colonies only at night. Repeatedly, the northern Gulf of Alaska shows up as an area of reduced population, as a boundary between subspecies, or as a limit to a range. This same area has a noticeable lack of total darkness during a substantial portion of the breeding season.

The nocturnal habit no doubt evolved because it was advantageous to concentrate on the breeding grounds only under the cover of darkness, when diurnal predators were at a great disadvantage. Cody (1973) states that Cassin's auklet (Ptychoramphus aleuticus), which is strictly nocturnal around its colonies, avoids these colonies on brightly moonlit nights. He sees this as an apparent response to gull predation. At higher latitudes the small alcids have overcome this disadvantage by swamping predators through their sheer numbers. In the Gulf of Alaska I suspect that few of the small seabirds, except possibly the fork-tailed storm-petrel (Oceanodroma furcata), have ever achieved great enough numbers to offset the impact of extended daylight.

Past disruptions of seabird populations are both natural and man-induced; however, the documentary record is much too fragmentary to allow us to fully appreciate what has occurred or what the long-term effect has been. To give some perspective to the problems associated with assessing change and attempting to understand it, some of the indicators of natural and unnatural change and flux in seabird populations are reviewed here.

The flux in bird numbers can be related to the time of day, season of the year, and atmospheric conditions on a short-term basis. This sort of flux or apparent flux can easily be explained. The underlying cause of some of the longer term flux is not so easily arrived at. Murie (1959), Gabrielson and Lincoln (1959), and Sowl and Bartonek (1974) have noted some of the man-induced changes. These are also explored to some extent in the species accounts as they are found to apply.

I sometimes refer to a colony size class when discussing the existing data rather than to an actual population estimate. The size classes used are defined as follows:

• Class I—less than 100 birds
• Class II—100-1,000
• Class III—1,000-10,000
• Class IV—10,000-100,000
• Class V—100,000-1,000,000
• Class VI—more than 1,000,000

The Dictionary of Alaska Place Names (Orth 1967) is the reference for those who wish to locate some of the less obvious sites. The Coast Pilot, No. 9 (U.S. Department of Commerce 1964) is another useful reference.

Species Accounts

Northern Fulmar (Fulmarus glacialis)

Petrels of a number of species can be found in the Gulf of Alaska, some of them in great numbers. Only the northern fulmar breeds there.

The fulmar is common in the offshore waters of the northern Gulf of Alaska throughout most of the year (Isleib and Kessel 1973). Most authors, including Clark (1911), one of the earlier ones, who commented on the distribution of fulmars farther out in the Gulf, have considered them to be abundant. Nichols (1927) raised one of the few voices of apparent dissent; he noted that in 1926 he encountered the largest number of fulmars (about 800) on 11 July in Shelikof Strait after he had left the Gulf. During the summer, fulmars are very common seaward of Montague Island, particularly to the northeast of Patton Bay and in the approaches to Montague Strait. Data derived from FWS surveys in July and August 1972 showed an estimated 10,000 fulmars in a stretch of waters 19 km wide along the east side of Montague Island (Isleib and Kessel 1973).

Over the Portlock Banks and in Stevenson Entrance, fulmars sometimes concentrate in very large numbers, either by themselves or in company with sooty shearwaters (Puffinus griseus). In August 1973, FWS observers crossing Perenosa Bay saw large numbers of tube-nosed birds moving northeastward across the Bay. Although these appeared to be predominantly shearwaters, there were also many fulmars. There was a general movement of birds through Shuyak Strait from Shelikof Strait into the Gulf of Alaska. It was not determined whether the fulmars were moving with the shearwaters or on a regular feeding flight. Fulmars are often found close to Afognak Island in the area between Sea Lion Rocks and Sea Otter Island. Gabrielson and Lincoln (1959) reported seeing swarms of fulmars in Marmot Strait and around the small islands on the north side of Afognak in early August. Murie (1959) noted fulmars in Shelikof Strait and again around the Shumagin Islands. There is nothing in this record to indicate any change in their distribution at sea recently.

The Semidi Islands support the Gulf of Alaska's largest fulmar breeding population, a Class V colony (U.S. Bureau of Sport Fisheries and Wildlife 1973). Gabrielson and Lincoln (1959) considered it to be one of the four largest colonies in Alaska.

Gabrielson (1940) was told by Captain Sellevold of the marine vessel Brown Bear that he thought the birds nested on Sea Otter Island in Perenosa Bay. Gabrielson also learned that they probably nested on Sea Lion Rock at the head of Marmot Strait. In August 1973 I observed fulmars in close proximity to Sea Lion Rock. More recently, small numbers of apparently breeding fulmars have been found in the Barren Islands (L. W. Sowl, personal observation and Edgar Bailey, unpublished FWS report, Anchorage, Alaska). Although no other colonies are known or suspected, the evidence suggests the possible existence of some.

Peterson and Fisher (1955), on noting dark fulmars between St. Paul and St. George when only the light morph was present on any of the colonies in the Pribilofs, expressed no surprise. They offered the opinion that a round trip of 960 km to one of the dark morph colonies in the Aleutians just might be within the operating range of a fulmar on a 4-day vacation from nest-tending duties. Using this as a general yardstick, it appears that the rich foraging grounds over the Portlock Banks might also be within the range of breeding fulmars from the Semidis. The trip up Shelikof Strait and on to Portlock Bank by way of Shuyak Strait is only slightly longer than the one from Chagulak to St. Paul. The feeding grounds off Montague Island would require a 1,600-km round trip from the colonies in the Semidi Islands. Birds from the Barren Islands and any colonies around Shuyak Island could easily reach the Montague Island grounds, but why would they cross the Portlock Banks to do so?

Fulmar colonies may be found in the Chiswell Islands. It is also a possibility that the existence of colonies on islands along the north coast of Afognak Island will be verified and that others will be found in the vicinity of Shuyak Island. Gabrielson and Lincoln (1959) expressed the opinion that there is almost certainly a colony on Sutwik Island. If there is one, however, I did not see it on one quick trip around the island in 1973.

Gabrielson (1940) expressed surprise at the size of the Semidi Island breeding colony. Gabrielson and Lincoln (1959) considered 1911 to be the first time breeding fulmars were found in the Shumagins. They apparently based this on two eggs collected there that year and documented in a plate in Bent (1964). Other than Gabrielson's opinion, there is nothing to indicate a major change in fulmar status during this century. If there has been a change in status, it has probably been in the direction of increasing populations.

Fork-tailed Storm-petrel (Oceanodroma furcata)

The fork-tailed storm-petrel probably breeds throughout the Gulf of Alaska. It is abundant at sea during the summer in most offshore waters. Murie (1959) described it as the dominant petrel in the Bering Sea and the North Pacific.

In view of its wide distribution and apparent abundance very little is known about the fork-tailed storm-petrel's breeding colonies. Friedmann (1935) recorded specimens and eggs from Kodiak dating back to 1843. Murie (1959) noted them as nesting on Sanak Island and stated that they almost certainly nested in the Shumagins and on other islands along the Alaska Peninsula. David Roseneau (Isleib and Kessel 1973) found this storm-petrel "breeding by the 10,000's" on East Amatuli Island in the Barren Islands in June 1965. This was subsequently verified in 1974 by Edgar P. Bailey (unpublished report, FWS, Anchorage, Alaska).

On 2 July 1972, responding to a tip by James W. Brooks (personal communication), M. E. Isleib and I anchored at Fish Island in the Wooded Islands. We did not locate any storm-petrel burrows, but a steady flow of storm-petrels passed over the boat throughout the darkest part of the night. Surveys conducted at about that time provided an estimate of 19,000 fork-tailed storm-petrels in Prince William Sound, primarily in or close to Montague Strait, and in coastal waters on the east side of the Sound's outer islands. In this area Isleib (personal communication) has noted a general movement of fork-tailed storm-petrels westward around Montague Island and into Prince William Sound through Montague Strait each morning and a corresponding countermovement each evening. I conclude that in 1972 there was a Class IV colony in the Wooded Islands, numbering between 19,000 and 38,000 birds. Additional colonies will be discovered in a similar manner as more systematic searches are made.

No colonies were discovered during the 1973 reconnaissance survey of the islands south of Alaska Peninsula. Working primarily inshore, FWS investigators encountered very few storm-petrels during the day. On the night of 14 June, the FWS vessel, Aleutian Tern, responded to a Mayday call and was either in transit or participating in rescue operations from 2245 to 0420 h on the morning of 15 June. During this period numerous fork-tailed storm-petrels were encountered, particularly off Cape Wedge on Nagai Island. After we anchored in Eagle Harbor on Nagai, more storm-petrels were heard about the vessel.

At about this same date, National Marine Fisheries Service enforcement officers flying fisheries patrols observed storm-petrels in abundance south of the Shumagin Islands (James Branson, personal communication). These observations support the belief that there are probably substantial undiscovered colonies in the Shumagin Islands.

Fork-tailed storm-petrels are abundant summer residents in the northern Gulf of Alaska and the estimate by Isleib and Kessel (1973) is that populations using the waters off the North Gulf Coast probably number in the millions. Certainly the same estimate is valid for the rest of the Gulf area west of the Chugach Islands.

The status of these birds relative to their historical abundance cannot be derived from the existing information. There is strong suspicion that the introduction of fox on many of the islands in the area during the early part of this century probably caused a reduction in their numbers. Murie (1959) said that experience taught him that wings left from fox kills or remains of storm-petrels in fox droppings could be accepted as evidence of the presence of a colony. Gabrielson and Lincoln (1959) reported that E. P. Walker visited the Wooded Islands in 1922 searching for a storm-petrel colony that had been reported to exist there in 1918. He could not find it even though he searched diligently. This apparent disappearance was attributed to the introduction of fox.

There is another factor to consider, however. The limited number of specimens now available from the Gulf of Alaska indicates that separate subspecies occupy the eastern and western Gulf of Alaska. The accepted boundary is somewhere in the vicinity of Prince William Sound. This is an indication that there has been a hiatus in this area of rather long duration. I have speculated that this sort of break may be in some way related to the length of day and a period during the summer when there is little darkness to cover activities near the colony. Thoresen (1964) and Cody (1973) have both reported that western gulls (Larus occidentalis) assemble in Cassin's auklet colonies on moonlit nights to prey on arriving adults. It is likely that other nocturnal species would provoke the same sort of hunting tactic. A light-related predation factor implies that the predators rely on sight. Avian predators are indicated.

Leach's Storm-petrel (Oceanodroma leucorhoa)

Even less well understood than the breeding distribution of the fork-tailed storm-petrel is that of Leach's storm-petrel.

Bendire (1895) quotes notes from Chase Littlejohn, who found Leach's storm-petrel to be an abundant breeder on unspecified small islands near Sanak in 1894. It greatly outnumbered the fork-tailed storm-petrel. On his visit in 1937 Murie (1959) learned that all of the large colonies of seabirds that had once existed there were gone. He attributed this to overfishing and associated perturbation and to the introduction of fox. No systematic assessment of seabirds on Sanak has been attempted since Littlejohn's time.

No Leach's storm-petrel colonies have been encountered during reconnaissance surveys of the Gulf of Alaska. Small numbers have been reported from time to time and while it is very much less abundant than the fork-tailed storm-petrel, I expect that it will be found in small numbers at various places in the Gulf of Alaska when it becomes possible to make more thorough searches. It may occur in remote areas like the smaller islands scattered throughout the Sandman Reefs—possibly even in large numbers. On the basis of the Sanak record, we must assume that this storm-petrel has been greatly reduced in numbers, at least in the western portion of the Gulf.

Double-crested Cormorant (Phalacrocorax auritus)

The white-crested cormorant, the race of the double-crested cormorant residing in the Gulf of Alaska, is principally an inhabitant of the marine environment. This cormorant is a common, but apparently patchily distributed, resident throughout the northern and western Gulf of Alaska.

Gabrielson and Lincoln (1959) thought that it nested only from Kodiak Island westward into the Aleutians. However, it probably breeds from Yakutat Bay westward. Isleib and Kessel (1973) estimated the abundance of the double-crested cormorant along the North Gulf Coast as several thousands, about one-tenth as common as the pelagic cormorant. It is the third most abundant of the four cormorant species nesting in the area. It occurs as scattered inclusions in many colonies throughout the area, and at least in the Shumagin Islands, even occurs in some colonies by itself.

There are no data on which to base an estimate of any change in status. It probably is not much affected by many of the naturally occurring perturbations.

Brandt's Cormorant (Phalacrocorax penicillatus)

On 22 July 1972, 13 Brandt's cormorants (4 sitting on nests) were found at Seal Rocks in Hinchinbrook Entrance, Prince William Sound (Isleib and Kessel 1973). Two years later I positively identified two individuals in breeding plumage among a mixed group of cormorants in the Chiswell Islands west of Seward. Are these recent range extensions? Possibly, but I propose an alternative explanation.

Palmer (1962) showed the distribution of this cormorant as breeding north to Puget Sound and as a straggler north to Forrester Island, Alaska. This viewpoint is shared by the American Ornithologists' Union (1957), which regards the bird as casual as far north as Forrester Island, where this species was collected by Willet (1918).

Let us look at the other record, the one that is not supported by specimens. Bent (1964) thought of Brandt's cormorant as a breeding resident of Forrester Island. Gabrielson and Lincoln (1959) admonished bird observers to be on the lookout for this particular cormorant in the vicinity of Ketchikan and Prince of Wales Island. Brandt's cormorant also appears on the bird list for the Kodiak National Wildlife Refuge as an accidental visitor.

Early observers like Bent were explorers. They carefully examined and made notes on all the birds they saw because there was always a chance of a new discovery. It is also very probable that Bent paid particular attention to the cormorants when he was at a place like Forrester Island. He would have undoubtedly been very interested in trying to confirm the presence of the now extinct Palla's cormorant (P. perspiculatus), as he must have been aware of Schlegel's (1862-64) list of the birds in the Dresden Museum since Willet (1914) had recently referred to it. The staffs for the Kodiak and Aleutian Islands National Wildlife refuges have included some very careful observers, such as Frank Beals. These men would have noticed the difference if a new bird such as Brandt's cormorant was seen, verified the sighting visually, and then noted it in their field diaries. They would not have bothered to develop the type of proof needed for an undisputable record, but the bird would have appeared in the refuge bird list (as it does).

The outside coasts of the Alexander Archipelago, Kenai Peninsula, and the Islands of the Kodiak Archipelago impose some logistical requirements which discourage all but the most determined birders. Not many have been able to reach more than very limited segments of the entire coast. Given the vast distances involved, few of the FWS vessels passing through the area have had the time to thoroughly examine any cormorant colonies or roosts bird by bird. Even for those who pause, the ever present swells and the constant chop of the summer westerlies make positive identification difficult.

It is possible that Brandt's cormorant has been in the area in small numbers for a long time, either regularly or intermittently. It could have escaped observation because of the conditions described above. This species may be there as a relict, as a pioneer, or only because surplus birds are being pushed into marginal habitat by population pressures on their main range to the south.

Pelagic Cormorant (Phalacrocorax pelagicus)

The pelagic cormorant is the most abundant of the four cormorants residing in the Gulf of Alaska. It is found throughout coastal Alaska south of the Bering Strait and even in some colonies in the southern Chukchi Sea.

Cormorants have a certain invisibility which is brought about by their universal presence. This blindness appears to have affected everyone, even the earliest observers.

The earliest accounts provide a composite picture of the distribution and abundance of the pelagic cormorant which is very similar to that encountered today. In southeastern Alaska, beginning at the eastern edge of the area under discussion, the pelagic cormorant was pictured as the sole resident cormorant. However, we know from Willet's collection of a Brandt's cormorant at Forrester Island that this might not be quite true. From Yakutat Bay westward into the Aleutians this species coexisted with the double-crested cormorant. In the Western Aleutians there is some disagreement, but in general it appears to have been accepted that the red-faced cormorant occurred there along with pelagic and possibly double-crested cormorants. In the Bering Sea this species coexisted with the red-faced cormorant.

A number of recent authors (Gabrielson 1940, 1944; Murie 1959; and others) have considered the pelagic cormorant to be the most widely distributed and abundant of the four species found in Alaska. Since the modern picture fits, in a general way at least, it would be easy to conclude that the species enjoys an unchanged status. There is just a faint suggestion that this may not be true.

Dement'ev and Gladkov (1966) refer to a great die-off of pelagic cormorants referred to earlier, in the Commander Islands. Stejneger (1885) enlarges on this disaster. It is true that Stejneger visited these islands a relatively short time after the die-off, but he reported that even though the pelagic cormorants were increasing, "people having seen their former multitude think that there is no comparison between the past and the present." Murie (1959) thought that the pelagic cormorant, while numerous, was outnumbered by the red-faced cormorant in the Aleutians. More recently there has been the rapid eastward expansion of the red-faced cormorant. Although it is not possible to determine what the real status of the pelagic cormorant is relative to its past status, I conclude that during this century its status relative to that of the red-faced cormorant has declined.

Red-faced Cormorant (Phalacrocorax urile)

The red-faced cormorant, in spite of superficial similarities to the pelagic cormorant, just does not look the same to an experienced observer. However, it would have been possible for inexperienced observers in the days before modern optics to overlook the differences. The problem was further compounded by the "invisibility" of the ubiquitous cormorants referred to earlier. Apparent absences or blank spots in their range may not have been real.

Dement'ev and Gladkov (1966), reporting on the Russian record, stated that the red-faced cormorant was common in the Commander Islands during the last part of the 19th century and into the early part of the 20th. Older authors had also reported it from Kamchatka and the Kurile Islands. Now, according to Dement'ev and Gladkov, it is an uncommon breeder on Mednyi Island in the Commander Islands and occurs only as an autumn visitor to some of the southern Kurile Islands.

Turner (1885) reported that the double-crested cormorant was abundant in the Near Islands and that the pelagic cormorant was common, but makes no reference to the red-faced cormorant. One specimen of the latter in the Leningrad Academy of Science was taken at Attu on 16 September 1844 (Gabrielson and Lincoln 1959), which indicates that they were probably present during the period reported on by Turner and, therefore, relatively uncommon. Clark (1911) identified red-faced cormorants only a few times and in the Aleutians only once, near Agattu. Dall (1874) noted two red-faced cormorants collected at Amchitka but he (Dall 1873) apparently did not see any east of Unalaska.

Nelson (1887) apparently found red-faced cormorants breeding on the Siberian and Alaskan mainlands at either side of Bering Strait, but Bailey (1948) searched for some sign of their presence and found none. Nelson (1887) also reported the red-faced cormorant from St. Matthew and St. Lawrence islands in the northern Bering Sea and from St. Michael and Nelson Island on the Alaskan coast. Gabrielson and Lincoln (1959) pointed out that it has not been found breeding north of the Pribilofs since then. Friedmann (1934) provides support for Nelson by reporting red-faced cormorant bones from archeological sites on St. Lawrence. Gabrielson and Lincoln (1959) cited two red-faced cormorants in the Leningrad Academy of Science which were collected in the Pribilofs in 1843. Dall and Bannister (1869) reported them to be plentiful on St. George Island. Baird (1869) also noted their presence in the Pribilofs.

Bent (1964) makes no mention of seeing the red-faced cormorant in the Aleutians. He gives their breeding range as the Bering Sea region, the Pribilof Islands, and perhaps the western Aleutians, the Commander Islands, and the coast of Siberia north of North Cape. The American Ornithologists' Union (1931) gave their breeding range as the Pribilof Islands, the Commander Islands, and Siberia north to North Cape.

Murie (1959) found a colony of between 4,000 and 5,000 red-faced cormorants nesting on Amak Island in 1925. In 1936 he was surprised to find that the red-faced cormorant was the most abundant breeding cormorant in the Aleutian Islands. Pelagic cormorants still appeared to be most numerous, but there were large numbers of nonbreeding birds. In 1936 he located "a good sized colony" of red-faced cormorants at Unga in the Shumagin Islands. He found about 300 birds starting their nests on 16 May.

In August 1946 Gabrielson (Gabrielson and Lincoln 1959) visited the colony at Delarof Harbor, Unga, where several thousand cormorants were observed. From a number of small samples he estimated that the red-faced cormorants outnumbered pelagic cormorants five to two. In 1973 I observed about 2,000 cormorants, mostly red-faced, in this colony. Gabrielson also located them at two other sites in the Shumagins and at Aghiyuk Island in the Semidi Islands.

Howell (1948) noted only double-crested cormorants at Double Island, Kodiak. Shortly after that the leaflet, Birds of the Kodiak Island National Wildlife Refuge (first issued in 1955), listed red-faced cormorants as common summer residents. The red-faced cormorant was next found at Katchemak Bay about 1963. Isleib (Isleib and Kessel 1973) first noticed red-faced cormorants wintering in Prince William Sound in 1969. In July 1972 Isleib and Sowl had found a colony containing 75 nests at Point Elrington at the western approach to Prince William Sound. By 1974 Isleib and Haddock (unpublished data, FWS, Anchorage, Alaska) found them east of the Copper River Delta at Wingham Island.

The relatively rapid expansion of the range and apparent population size of the red-faced cormorant is remarkable. But has this been a real expansion into vast stretches of new territory? The record in the literature which I have summarized shows, I think, something else. We can demonstrate a historical range for the red-faced cormorant that extends on the Asiatic Coast from North Cape, Siberia, south to the Kurile Islands, the entire Aleutian Arc including the Commander Islands, all the Bering Sea islands north to Bering Strait, Norton Sound, Nelson Island, and the islands south of the Alaska Peninsula at least as far east as Kodiak Island. The recently occupied coast from Cook Inlet to the Copper River may represent a real range extension. The breeding range of this species at the present time does not include parts of its historical range west of the Commander Islands or north of the Pribilof Islands.

The fragmentary record appears to show a long-term perturbation in the range and populations of the red-faced cormorant that covers at least 100 years. I believe that we are probably seeing a recovery of lost range and a return to something resembling a former distribution and abundance.

What caused the perturbation? I am not prepared to answer this question, but there are two occurrences which I find suggestive.

It is interesting to note (Dement'ev and Gladkov 1966) that on the Commander Islands the red-faced cormorant was most abundant during the first 50-odd years after the pelagic cormorants had been wiped out in the winter of 1876-77. Perhaps some clues are to be found in the interactions between these similar species.

It does not appear that the introduction of fox could have been a causative factor. The first observations of population expansion were noted almost concurrently with the heyday of the fox-farming industry. Because of its choice of nesting habitat (very steep cliffs), this cormorant would not have been affected by predators except for the one that went into a very rapid population decline at a time that would fit—the Aleut.

Jochelson (1968) and Hrdlicka (1945) summarized references to Aleut clothing in the diaries and reports of early Russian visitors to the Aleutian Islands. Evidently Aleut women sometimes wore a long, robe-like parka made of harbor seal (Phoca vitulina) skins or, for women of high rank, parkas made of sea otter (Enhydra lutra). The men in almost all reports were said to have worn bird-skin parkas; puffins and guillemots appear to have been preferred, but cormorants were sometimes used. It took about 40 puffin skins to fabricate a parka and a man evidently needed from one to three of these garments each year.

Sea otter populations were drastically reduced by Russian hunters. Rats were introduced to the Aleutians very early during the Russian period and must have had a substantial impact on populations of tufted puffins and guillemots. The introduction of fox would have had a further impact on burrow-nesting birds. Turner (1885) noted that Aleuts in the Near Islands kept the fox confined to Attu so that they could keep the fox away from the birds on Agattu. This is evidence of an Aleut recognition of serious competition. Could cormorants, particularly red-faced cormorants, have been preferred sources of fiber? Were Aleuts forced to rely more heavily on cormorant skins as puffin and guillemot numbers were reduced by rats and fox and sea otters by men?

Whatever the cause and effect, the status of red-faced cormorants now appears to be better in the Gulf of Alaska than for at least the last 100 years.

Glaucous-winged Gull (Larus glaucescens)

The glaucous-winged gull is apparently one of the more successful seabirds breeding in the Gulf of Alaska. While it is outnumbered (both locally and in total abundance) by the black-legged kittiwake, it is generally the most commonly seen and most uniformly distributed gull in the Gulf of Alaska. Murie (1959) called it the common breeding gull about the Alaska Peninsula. Cahalane (1943, 1944) considered it to be numerous to abundant around Kodiak and in the Shelikoff Strait area. Gabrielson (1944) reported that it could be seen in small numbers everywhere. Most recently, Isleib and Kessel (1973) reported it to be an abundant resident in the north Gulf Coast area. My own experience would confirm these observations.

This gull appears to use a wider variety of nesting sites than some others (Gabrielson and Lincoln 1959). Except where man's activities have created new food sources, there appears to be a close link between the location of glaucous-winged gull colonies and those of murres, kittiwakes, and cormorants. Swartz (1966) found that during the breeding season glaucous-winged gulls at Cape Thompson derived almost all of their food from murre eggs and chicks. I have noted small numbers of these gulls nesting, usually on turf near the tops of cliffs, in most colonies of favored prey species.

The glaucous-winged gull is the principal scavenger throughout much of coastal south-central Alaska. This has sometimes resulted in the development of large concentrations near canneries and, more recently, near dumps.

Two glaucous-winged gull concentrations stand out in the northern Gulf of Alaska. One of these is on Egg Island at the western end of the Copper River Delta. Patten (1976) estimated that this colony contained 10,000-12,000 gulls in 1975. At times it appears to spread onto nearby Hinchinbrook Island. M. E. Isleib (personal communication) has estimated its size as high as 25,000 gulls. The other large concentration is on the Susitna Flats across Cook Inlet from Anchorage. This colony, or colony cluster, may be larger than the one at Egg Island. There are no other known colonies even approaching these in size. Most colonies range between a few pairs and 2,000-3,000.

Glaucous-winged gulls do not appear to have had any great changes in population that can be detected from the literature. There have almost certainly been local fluctuations in the number of breeding birds as food supplies, such as canneries and dumps, have appeared or disappeared in an area. Long-term changes in salmon runs have undoubtedly had an impact as well. One other change, the reduced level of egging, has undoubtedly had an effect also. Along the Alaska Peninsula and in the Shumagin Islands, cannery workers of Filipino heritage and fishermen who have a strong Aleut heritage still harvest gull eggs for food. However, this activity is much reduced from what it must have been.

Herring Gull (Larus argentatus)

The herring gull is a resident of Upper Cook Inlet and is found up and down the coast from Prince William Sound to the Alaska Peninsula. Not too much was learned about it during the recent FWS reconnaissance. Williamson and Peyton (1963) reported the interbreeding of herring gulls and glaucous-winged gulls in this area. This interbreeding has resulted in a situation in which assignment of these gulls to one group or another in the field can be rather arbitrary. The result has most often been that field observers tend to lump them with glaucous-winged gulls unless their herring gull characteristics are obvious. Specimens collected by Williamson and Peyton (1963) indicate that herring gulls have the edge in numbers in Upper Cook Inlet.

Black-legged Kittiwake (Rissa tridactyla)

The black-legged kittiwake is the most abundant gull in the northern and western Gulf of Alaska. Colonies of this species can be found throughout the entire area, and range in size from a few pairs (Class I) to more than 100,000 birds (Class V). They may be found in essentially pure colonies, but are often found sharing colonies with murres.

The center of abundance for breeding black-legged kittiwakes in the Gulf of Alaska is in the Semidi Islands, where Palmer Sekora (U.S. Bureau of Sport Fisheries and Wildlife 1973) estimated that there were 426,000 breeding kittiwakes in 1972. He located kittiwake colonies at eight sites, ranging in size from 1,000 to 109,000 nesting birds. The size of the average colonial site was 27,000 birds. Ten sites were Class IV in size and one was a solid Class V.

The easternmost known colony in the northern Gulf of Alaska is at Wingham Island. Up to 1973, 22 colonies had been located in Prince William Sound. The largest of these contained only 5,636 nests in 1972 (Isleib and Kessel 1973). Class IV or larger colonies are found at Cape Resurrection, the Barren Islands, Chisik Island, Boulder Bay and Cape Chiniak on Kodiak Island, and at Delarof Harbor and the Haystacks in the Shumagin Islands. It is interesting to note that Gabrielson (1940) considered Whale Island to be one of the largest known kittiwake colonies in Alaska. He stated that there were many thousands of pairs extending over a mile or more of cliff. He saw a second site which he did not visit but looked equally large. A photograph in an article by East (1943) also indicated the presence of a large colony. C. J. Lensink (personal communication) estimated that there were about 100,000 kittiwakes in the colony in 1956. When last visited by Vernon Berns (personal communication), this colony contained only 3,000 birds. It is also of interest that Gabrielson (1940, 1944) did not notice either the kittiwakes or the murres now breeding on Nord Island in the Barren Islands or the kittiwakes on East Amatuli Island.

Whale Island and possibly the colonies in the Barren Islands give evidence of local population fluctuations, but for the most part I have not found an indication of a major perturbation over the past 40 years. Before 1936, the record is too fragmentary to allow an assessment.

One of the interesting aspects of kittiwake ecology in the Gulf of Alaska is the common occurrence of breeding failure. David Snarski (December 1943 Quarterly Progress Report, Alaska Cooperative Wildlife Research Unit, University of Alaska) observed breeding failure on colonies in the Tuxedni National Wildlife Refuge in 1970 and 1971 and obtained circumstantial evidence of another failure in 1972. In 1973 all of the breeding cliffs were occupied and nesting was successful. Whatever the cause of these periodic failures, they do not yet appear to have had a permanent impact that we are able to measure.

Red-legged Kittiwake (Rissa brevirostris)

Red-legged kittiwakes are not now known to breed in the western Gulf of Alaska. Turner (1886) stated that he saw a few at Sanak in 1878. We also have Stejneger's (1885) statement, that "red-legged" kittiwakes nest from Bering Island to Sanak. Friedmann (1937) reported two humeri from Kodiak Island middens. During the summer of 1976, two birds were observed off Kodiak Island by Irving M. Warner (personal communication), and one at 158°W and 54°30'-54°20'N south and east of the Shumagin Islands (Patrick J. Gould, personal communication).

Turner (1885) listed the red-legged kittiwake as abundant and breeding in the Near Islands. Turner (1886) also stated that he had seen quite a number about a cliff back of the village on Akutan Island in 1878. He added that to the westward this kittiwake was more abundant than the black-legged kittiwake. Murie (1959) expressed the opinion that Turner had confused the short-billed gull with the "short-billed" kittiwake. Clark (1911) also reported that he had seen the red-legged kittiwake in small numbers near Unalaska and that they became progressively more common west to the Near Islands. Nelson (1887) reported seeing large numbers of red-legged kittiwakes at Unalaska. Murie (1959) and Gabrielson (1940, 1944) did not see any red-legged kittiwakes in the Aleutian Islands. The species has recently been discovered breeding at Buldir and Bogoslof islands (G. Vernon Byrd, personal communication).

Is it possible that we have here another species which is exhibiting a response to some unknown long-term perturbation? The suggestion that such an event has occurred is faint, but it is there. Do we have in the red-legged and black-legged kittiwakes an example of yet another congener pair that has been affected by some perturbation in which one was affected positively and the other negatively? Clark (1911) reported small numbers of black-legged kittiwakes to go with large numbers of red-legged kittiwakes in the Near Islands, which is the reverse of the current situation.

Arctic Tern (Sterna paradisaea)

Gabrielson and Lincoln (1959) attribute to the Arctic tern the most extensive range of any Alaskan water bird. It is found in suitable habitat everywhere north of Tracy Arm in Southeastern Alaska. Murie (1959) stated that he found it nesting at suitable sites everywhere he went. Isleib and Kessel (1973) considered it to be an abundant breeder in Prince William Sound and along the northern Gulf Coast.

The Arctic tern was observed in FWS aerial surveys in Prince William Sound, and surveys in July and August 1972 provided an estimate of 45,000 terns in the Sound (Isleib and Kessel 1973). On the other hand, tern colonies were located only rarely in the FWS colony surveys before 1975. This is, however, a reflection of the equipment and methods used and not of the abundance of terns.

From the fragmentary data available, it is not possible to detect changes in Arctic tern status at the present time. We have to assume that the widespread introduction of fox had at least local impact. Although this tern uses a wide variety of nesting sites, it tends to nest on flat sites where access by mammalian predators is easy.

Aleutian Tern (Sterna aleutica)

No Aleutian tern colonies were discovered in the Gulf of Alaska area during FWS colony surveys in the early 1970's. This is again a reflection of the fact that surveys were not designed to locate tern colonies. Aleutian terns were encountered at least twice, once during late March 1972 in Hawkins Cutoff, Prince William Sound, and again when two birds were noted offshore from the Katmai National Monument on 30 May 1973 (L. W. Sowl, personal observations).

The type specimen of the Aleutian tern and a single egg were collected at Kodiak Island on 12 June 1868 by Bischoff (Coues 1874). Fisher (Gabrielson and Lincoln 1959) collected four more eggs in 1882. The bird was not found breeding there until Howell (1948) found a colony of 50 pairs at Bell's Flats in 1944. Walker (1923) found them nesting on the Situk River, Yakutat, in 1917 and shortly thereafter saw them at the Alsek River Flats. He also reported that D. H. Stevenson of the Bureau of Biological Survey had told him that they nested on the Isanotski Islands at the end of the Alaska Peninsula. This latter report was the only one from the Aleutian Island chain for many years. Isleib and Kessel (1973) considered it an uncommon local breeder in the northern Gulf of Alaska. Isleib estimated its population at a few hundred pairs on the Copper River Delta in May 1973 and 300-500 birds in June 1970. He also reported that they appeared more or less regularly near Controller Bay and off the Situk River.

In recent years Aleutian terns have been seen with increasing frequency in many places in western Alaska and the Aleutian Islands. This is probably partly due to the increasing level of field work. At Amchitka Island the several colonies that have been found in recent years are almost certainly exhibiting a response to the removal of fox from the island.

Although there is no way of determining what the past status of the Aleutian tern has been in the Gulf of Alaska area, it has been there in small numbers since it was first discovered on Kodiak. It has probably not been abundant at any time and may have suffered a long-term decline brought about by the introduction of fox.

Common Murre (Uria aalge)

The common murre is resident in the northern and western Gulf of Alaska from Pinnacle Rock, Kayak Island, westward. East of Cook Inlet colonies are located at Wingham Island, the Martin Islands, Middleton Island, Porpoise Rock in Hinchinbrook Entrance, Barwell Island/Cape Resurrection, the Chiswell Islands, the Barren Islands, and Chisik Island.

For some reason, the islands of the Kodiak-Afognak Archipelago do not host any known major murre colonies. There is also a rather large gap between the Chisik Island colony and the next major colony at Oil Creek west of Puale Bay. Directly west of Oil Creek is another colony at Cape Unalishagvak. Both of these latter colonies are Class V and they are the first colonies of this size to be encountered in the Gulf of Alaska. West of these colonies the next large colony is at Atkulik Island. To the south, midway between the last-named colonies, lies the major composite murre colony in the Semidi Islands. These sites make up the only Class VI colony in the Gulf of Alaska. Westward, the next major colony, a Class V, is at Spitz Island south of Mitrofania Island. In the Shumagin Islands one Class V colony is at Karpa Island, and lesser colonies with large murre components are found at the Haystacks, Castle Rock, and Bird Island. Only minor murre colonies are found between the Shumagin Islands and the end of the Alaska Peninsula.

Gabrielson and Lincoln (1959) were aware only of the colonies at Cape Resurrection (which Gabrielson considered to be large), at the Chiswell Islands, and at Chisik Island for the area from Cook Inlet east. Gabrielson visited the Barren Islands on 13 June 1940 and apparently did not notice the present murre colonies, both Class IV, at East Amutuli (an island which he visited) and Nord Island.

Gabrielson (Gabrielson and Lincoln 1959) found a few small colonies at Kodiak, mostly on small offshore islands. Gabrielson found common murres to be abundant in the Semidi Islands and stated that there were no notable colonies in the Shumagins, although on his return to the Shumagins in 1949 he did find a fairly large colony at the Haystacks. That size description would fit the colony that is there now. He obviously did not see the other colonies. Rausch (1958) reported murres from Middleton Island.

There is quite a difference between the distribution of murres as we know it today and the way Gabrielson and Lincoln pictured it. Why does this difference exist? There are two possible answers: either the number of colonies has increased, or the coverage of colony locations has improved. The latter case, at least, is established. I must confess to being puzzled by the way Gabrielson was able to move about close to what are now known to be sizeable colonies without seeing them, those in the Barren Islands and the Shumagin Islands in particular. Perhaps this represents the vague outlines of yet another population change.

The center of abundance for murre distribution in the Gulf of Alaska today is from Paule Bay west to eastern Shumagin Islands. The Semidi Islands are the heartland of this area of maximum abundance. We have no definitive data on species composition of these colonies. Common murres undoubtedly dominate in most of the colonies; the only ones where we know of a sizeable thick-billed murre component are in the Shumagin Islands.

Thick-billed Murre (Uria lomvia)

Thick-billed murre population information cannot be separated from that of the common murre on the basis of existing data. A direct assessment of present-day status is not possible. After reviewing what we know about their distribution, I suggest a way to examine the question indirectly.

The thick-billed murre is found in colonies with the common murre from Middleton Island westward; Rausch (1958) noted about 400 murres at Middleton Island and observed that the thick-billed murre outnumbered the common murre by several times. Isleib and Sowl (FWS, unpublished data) saw a thick-billed murre mixed with common murres at Porpoise Rock in July 1972. Isleib and Kessel (1973) expressed the opinion that small numbers of thick-billed murres will be found in most common murre colonies in the northern Gulf of Alaska when it is possible to survey these colonies in detail. Karpa Island had a significant component of thick-billed murres in June 1973, and they constituted 40% of the colony at the Haystacks (L. W. Sowl, unpublished data).

Bent (1963) reported that many thick-billed murre eggs have been taken by collectors at Round Island in the Shumagin Islands. Dall and Bannister (1869) reported a thick-billed murre that was taken at Kodiak in 1867.

The Gulf of Alaska is at the periphery of the breeding range of the thick-billed murre. While it probably occurs in mixed colonies with the common murre throughout this area, the thick-billed murre is much less abundant. Occasionally in the Gulf of Alaska, a colony will be occupied predominantly by the thick-billed murre. Gabrielson and Lincoln (1959) noted that the thick-billed murre outnumbered the common murre in many colonies in the Aleutians and that it became progressively more common at higher latitudes.

We have almost no data relative to the species composition of murre colonies in the Gulf of Alaska. Until we do it will not be possible to fully understand the population status of the thick-billed murre. It appears that changes in the species composition of murre colonies in the Bering Sea may be an indicator of perturbation. The data for the Gulf of Alaska are still too fragmentary to provide any indication of whether or not the same indicator would work there. Close monitoring of the Shumagin Islands colonies over a number of years might produce the answer.

Earlier in this paper I noted the dramatic changes in species composition of murre colonies on Walrus Island. Gabrielson and Lincoln (1959) also commented on this well-documented and anything but static situation. Investigators who visited this island during 1976 reported seeing no murres on the island and only small numbers on offshore rocks. James Bartonek (personal communication) said that this situation has prevailed for several years.

There is an indication that a similar population fluctuation and change in species composition of murre colonies have also occurred on St. Matthew Island. Bent (1963) found mostly common murres and few thick-billed murres at St. Matthew. Hanna (1916) saw only thick-billed murres. Later, Gabrielson (1941) found this to be true in 1940.

Dramatic fluctuation in murre populations may be common and, at least in some cases, the two species may be affected differently. Perhaps this phenomenon has potential for providing us with an indicator of some natural perturbations.

Peterson and Fisher (1955) expressed the opinion that thick-billed murres arrived at the nesting ledges later than the common murre and had to take the sites that were left. Tuck (1960) reported data from the western Atlantic showing that thick-billed murres arrive later than common murres. On the other hand, Belopol'skii (1961) reported data showing that the two species arrive on breeding colonies in East Murman simultaneously. At Cape Thompson, Swartz (1966) found that thick-billed murres arrived about a week before common murres. The date of arrival, while perhaps a contributing factor, is probably not decisive. Interspecific competition of another sort is indicated.

In mixed murre colonies where there are large numbers of common murres, this species occupies the choice nesting sites. Thick-billed murres are usually left with the narrower ledges while the common murres occupy the longer, broader ledges (Belopol'skii 1961). The broader ledges have lower chick and egg mortality (Spring 1971). Spring also noted that thick-billed murres are excluded from the centers of mixed colonies. Johnson (1938) found that this contributes to higher losses of eggs to predators and to the loss of other social benefits of occupying the colony center (Johnson 1941).

Kozlova (1961) said that during the occupation of a colony there is a sharp competitive struggle between the two species. In the end thick-billed murres are pushed out to the periphery of the colonies or left with narrow ledges or other equally unfavorable sites. Spring (1971) studied the functional anatomy of both species and concluded that the common murre is more successful in these encounters because it has a more upright gait and greater agility than the thick-billed murre.

It follows that in a portion of their respective ranges, where the two species overlap and where there is an equal chance that either common murres or thick-billed murres will dominate a given colony, the common murre dominates. I conclude from this that where there are dramatic changes in species composition of murre colonies, such as at Walrus Island, it is probably because the common murre has been greatly reduced in numbers at the colony.

Spring (1971) concluded that the common murre is well adapted to pursuit and capture of pelagic fishes and that the thick-billed murre is better adapted for deep diving and the capture of benthic fishes and pelagic and benthic invertebrates. Having greater latitude for food selection, the thick-billed murre would have a greater tolerance for ecological perturbations affecting the available food supply. The common murre has an advantage when pelagic fishes are available but cannot switch to the other foods as readily as can the thick-billed murre. The low density of pelagic fishes in high arctic areas probably also accounts for the greater success of the thick-billed murre at higher latitudes relative to common murres.

Belopol'skii (1961) presented data from East Murman which indicates that the common murre restricts its diet almost entirely to a small number of fish species. Swartz (1966) found strong indications that there were significant differences in the food preferences of the two species of murres. Thick-billed murres made much greater use of invertebrates. Bédard (1976) asserted that it is well known that the common murre is quite partial to zooplankton. So again the issue is not clear-cut.

The situation is, of course, much more complex than I have portrayed it. Nonetheless, I think that it offers potential for use as a tool in assessing population change and perturbations in the food supply which should be studied quite closely.

Pigeon Guillemot (Cepphus columba)

Gabrielson and Lincoln (1959) noted that the pigeon guillemot was one of the most regularly observed birds in Alaskan waters. It is found everywhere throughout the northern and western Gulf of Alaska area, with only a few understandable and relatively small blanks, such as in the silty waters of Upper Cook Inlet. Because it obviously lacks the breeding murres' need for close contact with its nearest neighbors, it is able to exploit the available nesting habitat to the fullest. It seems that literally every bit of suitable nesting habitat is normally occupied.

Because of the dispersed way in which it breeds and because it does much of its feeding in the onshore zone (which is hazardous for boats) the pigeon guillemot is an almost impossible species to inventory by standard methods.

There is no evidence that the pigeon guillemot has been greatly affected by any major perturbation. Because of its choice of nesting habitat, it is probably subject to the attack of only one egg predator, the rat. Because of its loose social structure and the way it selects nesting sites, eggs and young do not sustain loss from panic flights. Its dispersed distribution should insure that man-made impacts such as oil spills will have limited impact.

The population levels of the pigeon guillemot are probably relatively very stable. The widespread introduction of the rat to most of its nesting range undoubtedly had impact, but this impact has gone undocumented. It would be interesting to follow the response of guillemot populations on islands where rats had been totally removed, if that ever becomes more than a dream.

Marbeled Murrelet (Brachyramphus marmoratus)

The marbled murrelet apparently breeds throughout most of the northern and western Gulf of Alaska. This apparently is a necessary condition because to date, at least in this part of Alaska, we can only guess where and under what conditions this murrelet breeds.

In some relatively sheltered waters like Prince William Sound, where marbled murrelets were estimated to number about 250,000 in 1972 (Isleib and Kessel 1973), they are the most abundant seabirds. We know from Dixon (1908) and Grinnell (1910) that this has been so in Prince William Sound since the beginning of the century. We know also that the type specimens came from there as well (Stresemann 1949), which is not necessarily an indication of abundance but is suggestive of their abundance relative to species not collected.

Gabrielson (Gabrielson and Lincoln 1959) found marbled murrelets common near Yakutat, in Prince William Sound, in Resurrection Bay, and at Kodiak, and reported seeing them at the Chiswell Islands and at Chignik and Pavlof Bay on the Alaska Peninsula. Cahalane (1943, 1944) found them to be common in Kupreanof Strait, and along the Alaska Peninsula north of Katmai Bay. Murie (1959) found them all along the Alaska Peninsula. My own field notes from 1973 indicate that the only place where they were common along the Alaska Peninsula was at Wide Bay.

We can sample marbled murrelet numbers by using standard transect methodology; however, I have some very serious reservations about our ability to convert these data into a population estimate. This is not an unusual assessment for Alaskan seabirds in general, but I think it is particularly apropos to this species.

We are still able only to guess at where the marbled murrelet nests and we have not a clue as to what sort of nesting strategy they pursue. I am not prepared to accept, on the basis of one North American record (Binford et al. 1975), that tree nesting is its habit throughout its range. What has been proved is that the marbled murrelet nests in trees and not, as these authors would have us believe, that it does not nest on the ground. It has become rather fashionable to ignore the Chichagof Island record (a ground nest), but it has not been discredited. The color of the Chichagof egg differs from that of the Big Basin egg, but does agree with the one taken from an oviduct by Cantwell (Gabrielson and Lincoln 1959). My own experience leads me to believe that tree nesting, if it occurs, is not the common habit of marbled murrelets nesting in the Prince William Sound region.

After many hours of observing marbled murrelets over a period of several years, I am intrigued by a number of things. These birds, as often as not, appear to be clustered in "pairs" as they feed. This occurs even at what should be the height of the breeding season. On several occasions I have noted a very pronounced evening flight of these birds from gathering areas on the water up into the surrounding mountains at sunset. This has moved me to wonder if their nesting strategy includes incubating at night but less than full-time attendance on days when the eggs can be warmed by the sun. We know that periodic egg-neglect is an aspect of storm-petrel behavior (Pefaur 1974). Is this behavior also possible on a more regular basis in an alcid? If so, it would certainly help explain why nests are hard to find.

It is apparent that more needs to be known about the population dynamics and life history of the marbled murrelet before we can make a proper estimate of its abundance. In spite of the fragmentary record, I conclude that the marbled murrelet probably enjoys the same relative abundance and distribution that it did at the beginning of the century.

Kittlitz's Murrelet (Brachyramphus brevirostris)

The Kittlitz's murrelet is not as abundant as the marbled murrelet, but locally it is sometimes found in large numbers. FWS surveys conducted during July-August 1972 provide an estimate of 57,000 murrelets of this species in Prince William Sound. Almost a fifth of these were concentrated in Unakwik Inlet above Unakwik Reef. Even more interesting, about 2,500 of these birds were concentrated in one loose flock.

In addition to Unakwik Inlet, Kittlitz's murrelets concentrate in College Fjord in Prince William Sound and in the waters fronting the Bering-Malaspina ice-fields (Isleib and Kessel 1973). Common as they are in these waters, this species is supposed to be even more abundant at Glacier Bay. The common feature of these waters is the amount of ice that can be found below their tributary glaciers.

The Kittlitz's murrelet is apparently distributed from LeConte Bay, east of Petersburg, Alaska, north to Point Barrow and west across the Aleutians to Attu, where Murie collected a pair (Gabrielson and Lincoln 1959). I once flushed a murrelet from an area of tread and riser topography near the top of the highest point on Kiska Island in heavy cloud cover, and although I could not see this bird well, I thought it to be of this species. From the range description in Gabrielson and Lincoln (1959) and Udvardy's (1963) range map, it is apparent that the distribution of this species is rather patchy, but I suspect that for the more mountainous part of its range this is more apparent than accurate. The record is too fragmentary to allow an assessment of any change in status during the historical period.

Ancient Murrelet (Synthliboramphus antiquus)

Chase Littlejohn (Bendire 1895) spent the spring and summer of 1894 collecting eggs on islands south of the Alaska Peninsula. He has left us a detailed record of what he saw but not where he saw it. Bent (1963) stated flatly that the site of his collecting was Sanak Island and this has common acceptance. Several things in his account point to a site which was a small island with several peers close by, but this could not have been Sanak. It could have been an island in the Sanak Island group or it could equally well have been somewhere in the Sandman Reefs. Unfortunately, because of this the record is clouded. There has never been anything approaching a survey of the southern half of the Sandman Reefs. We do not know what breeding colonies are there.

At any rate, Littlejohn told of the large numbers of Leach's storm-petrels, fork-tailed storm-petrels, auklets (of which only Cassin's is specifically identified), and ancient murrelets which occupied a large number of small islands. He could not calculate the number of breeding murrelets on his small island, the size of which I interpret to have been of the same order of magnitude as two others which he estimated were about 2 acres. He does say that the murrelets must have numbered several thousand and could, if left alone by the Aleuts, have quickly grown too numerous for the island to accommodate.

Murie (1959) made a brief visit to Sanak in 1937 and learned that there were no longer any large colonies of seabirds. He attributes this to exploitation of the fisheries and to the fox-farming industry. Littlejohn told of the repeated visits of Aleuts to his small islands, where they took hundreds of birds each time and all of the eggs they could find. This kind of activity could not help but disrupt the breeding on these islands.

Littlejohn's description of the ancient murrelet's nest leaves little doubt that the birds could be reached by fox or rats with ease. The birds showed no particular care in selecting a nest site and often worked their way back no more than about a meter into the dead vegetative cover from preceding years, where they scratched out a shallow nest.

There are few records of the ancient murrelet from the northern and western Gulf of Alaska. Friedmann (1935) reported the collection of a series of eggs in 1884 on Kodiak Island. Chase Littlejohn (Bendire 1895) collected eggs from somewhere in the Sanak Group in 1894. In 1908 Dixon (Grinnell 1910) saw a bird in Port Nellie Juan. Several were seen by Jaques (1930) near Belkofski in May 1928. Gabrielson collected one bird at Cordova in September 1941 and another at the Chiswell Islands in July 1945 (Gabrielson and Lincoln 1959). He saw numerous flocks in the Gulf of Alaska on 30 July of an unnamed year. In 1943, he would have been near Cape Spencer on that date. In 1945 he would have been near the Chiswell Islands. In either case, he was probably somewhere in Blying Sound.

The ancient murrelet is relatively uncommon but regularly observed in the inshore waters along the outer coasts of the islands fronting Prince William Sound. FWS surveys in July-August 1972 provided an estimate of almost 1,000 birds, mostly in nonbreeding plumage, along the outer coast of Prince William Sound (Isleib and Kessel 1973). Small numbers were found feeding close to the Wooded Islands on 24 July (my personal observation). Rausch (1958) saw a few off Middleton Island in 1956. Isleib (Isleib and Kessel 1973) saw 400-500 widely distributed at the mouth of Yakutat Bay in July and August 1968. The only large numbers of ancient murrelets encountered on the FWS survey of the Alaskan Peninsula in 1973 were in the Shumagin Islands. They were very common in East Nagai Strait on 9 June and more than half of the 1,300 seabirds per square nautical mile encountered between Little Koniuji and Chernabura Islands on 11 June were ancient murrelets. At Nagai Island an estimated 5,000 ancient murrelets were observed in the west bay at Pirate Shake, and later (on 19 June) several were observed in the vicinity of Midun Island (FWS, Anchorage, Alaska, unpublished data).

On the basis of the observations recounted above, I have to conclude that ancient murrelets are fairly regularly, if patchily, distributed throughout the northern and western Gulf of Alaska. I do not believe that the void in their range shown for the northern Gulf of Alaska by Udvardy (1963) is correct. Several colonies are there, awaiting discovery.

Ancient murrelets are not abundant in the Gulf of Alaska but they are certainly more numerous than we have been able to prove. It is not possible to tell from the existing data whether they were once more abundant than they are now. I suspect, on the basis of the Sanak Island experience, that we can conclude that this species has been reduced in number by various of man's activities.

Cassin's Auklet (Ptychoramphus aleuticus)

Cassin's auklet is a very uncommon bird in the northern Gulf of Alaska. In the western Gulf it is more common, particularly from the Shumagins west.

This auklet apparently once bred in great numbers on islands in or near the Sanak Group where Chase Littlejohn (Bendire 1895) found them to be twice as numerous as the ancient murrelets. Murie (1959) did not find them there.

Littlejohn began encountering Cassin's auklets at sea some 290 km southeast of Unga, Shumagin Islands. Murie (1959) encountered them near the Shumagins in May 1937. During the FWS 1973 reconnaissance survey of the Alaska Peninsula, these auklets were not encountered (or at least not identified) until we reached the vicinity of Unga Strait where we saw a few in mixed flocks with other murrelets and auklets. They were most numerous in East Nagai Strait. We encountered them only twice in a situation which indicated they might be breeding—on Hall and Herendeen islands on the north end of Little Koniuji Island.

Murie (1959) considered Cassin's auklet to be no longer common west of Kodiak. In Gabrielson's many voyages through the northern and western Gulf of Alaska he encountered them only twice, once off Cape Spencer and once in the Chiswell Islands.

Thoresen (1964) commented that throughout the northern part of its range the Cassin's auklet has become gradually less frequent. Although there are no data to dispute this, I believe, as do Isleib and Kessel (1973), that they are more numerous than observations would indicate, and I would apply this to the entire area. There are certainly colonies remaining in the Shumagin Islands, and quite probably along the south coast of the Kenai Peninsula. When it is possible to fully explore the Sandman Reefs there is a good probability that they will be found there.

We can only guess at the reasons for their decline. Bendire (1895) and Murie (1959) have described some contributing factors.

Parakeet Auklet (Cyclorrhyncus psittacula)

Gabrielson and Lincoln (1959) described the parakeet auklet as the least colonial of any of the Alaskan auklets. They also considered the Aleutian Islands to be its principal nesting grounds. There are old records of breeding parakeet auklets from Kodiak (Friedmann 1935) and Little Koniuji (Bean 1882). Grinnell (1910) reported two that were seen on Green Island, Prince William Sound, and several more that were seen near Knight Island.

Murie (1959) did not see any parakeet auklets near Kodiak and Afognak islands which he considered to be the eastern part of their range. He did not think they were abundant anywhere along the Alaska Peninsula. He found a few near Sutwik Island in May 1936 and then noted that they were fairly common near the Shumagins in May 1937.

Gabrielson found this species to be quite numerous on the north side of Chowiet Island in the Semidi Islands in 1945 (Gabrielson and Lincoln 1959). He also saw numerous individuals in Marmot Strait and saw one in the Chiswell Islands during the same year. David Roseneau (Isleib and Kessel 1973) found hundreds close to East Amatuli Island in the Barren Islands in 1965.

During FWS colony surveys, parakeet auklets have been found in close proximity to six seabird colonies in Prince William Sound. During the July-August 1972 surveys, they were estimated to number about 3,000 in the Sound. They have also been found closely associated with Chisik Island (David Snarski, personal communication), the Chiswell Islands, Nord and Sud islands in the Barrens, Sea Otter Island, and Central and Long islands along the Alaska Peninsula. They were most numerous in the Shumagin Islands, where they were found near Castle Rock, Hall (9,000), Herendeen (3,000), Atkins (more than 5,000), and Little Koniuju islands. They were again encountered south and west of Cold Bay at High, Fawn, Let, Amagat, Umga, and Patton islands. Many of these islands are in the north half of the Sandman Reefs, the only portion where any attempt has been made to survey seabird colonies.

The parakeet auklet may not be abundant anywhere in the Gulf of Alaska but, based on the numbers of places it has been seen in recent years, its population appears to be well dispersed and probably doing very well. This auklet is most abundant from the Shumagin Islands westward. It is almost certainly more numerous than has been thought. Its habits are secretive enough so that it could easily escape notice.

Because the parakeet auklet nests predominantly under boulders, it probably was not much affected by fox. Rats would certainly have reduced its numbers wherever these were introduced into its breeding habitat. We have no data to tell us whether there may have been population fluctuations in the past, but there undoubtedly were at least minor ones locally after rats were introduced.

Crested Auklet (Aethia pygmaea)

Udvardy (1963) shows the breeding range of the crested auklet as extending from southern Kodiak Island westward. Within the northern and western Gulf of Alaska, it is certainly most abundant in the eastern Shumagin Islands.

Isleib saw this auklet in Prince William Sound 3 times during the winter of 1972-73. These are the only records he was aware of for that area (Isleib and Kessel 1973). David Roseneau (Isleib and Kessel 1973) saw several in Amatuli Cove, Barren Islands, in June 1965. I observed one in the vicinity of Cape Spencer in August 1973.

Friedmann (1935) listed the crested auklet as a breeding bird at Kodiak, but considered it to be much more abundant as a wintering bird. Townsend (1913) has provided us with a vivid description of the myriads of crested auklets he encountered at Yukon Harbor, Little Koniuji Island. Gabrielson and Lincoln (1959) noted large numbers of crested auklet around Simeonof and Bird islands in the Shumagin Islands in 1946 and stated that the Yukon Harbor colony was still thriving.

Crested auklets were not encountered on the 1973 FWS reconnaissance survey until we reached the Shumagin Islands. They were abundant only at the southeastern end of Little Koniuji, where we encountered perhaps 10,000 in Yukon Harbor and more than 50,000 in a small cove directly south of Yukon Harbor on the opposite side of the island. As numerous as they were, they did not match Townsend's myriads or even come close to his assessment that they "were here more numerous than the 'choochkies' at St. George." St. George Island in the Pribilofs is famous for its least auklets which, in the past, have been estimated to number as high as 36 million (Peterson and Fisher 1955). The numbers there today do not even approach this level and we have no way of knowing how abundant they were when Townsend visited the Pribilofs, but I think it is safe to say that they probably numbered in the millions. There are probably more crested auklets than we observed on Little Koniuji, but there is certainly no longer anything approaching millions of birds. Properly pronounced, Koniuji is the Aleut name for the crested auklet, so we can assume that the original inhabitants were impressed by its numbers.

During the 1973 FWS survey we did not see crested auklets at either Simeonof or Bird islands. On the overgrazed and cattle-trampled Simeonof it does not seem possible that any could still exist.

I suspect that a cattleman's greed has been the undoing of any crested auklets that may have nested on Simeonof Island. This would not account for the loss of any colonies that may have been on Bird Island, but the decaying fox-trapper's cabin on that island undoubtedly tells the story. Churnabura, with its feral cattle, presents much the same problem as Simeonof. As for Little Koniuji, have horned puffins been partly responsible for the decrease in crested auklets? The puffin colony at the south end of Little Koniuji must be exactly where Townsend's millions of crested auklets once nested.

Least Auklet (Aethia pusilla)

No least auklets were encountered in FWS surveys in the Gulf of Alaska in the early 1970's. Udvardy (1963) shows their breeding range as starting well west in the Aleutians. Gabrielson and Lincoln (1959) give the eastern limit of their breeding range as the Shumagin Islands. Bent (1963) listed their breeding range as extending east to Kodiak Island, and Friedmann (1935) knew of only a few specimens taken in the winter from Kodiak. Perhaps least auklets nested somewhere in the western Gulf of Alaska, and they may still, but at the moment we have no evidence to prove that they do.

Rhinoceros Auklet (Cerorhinca monocerata)

Udvardy (1963) would have us believe that the rhinoceros auklet did not nest between southeastern Alaska and the southern Kurile Islands. Bent (1963b), on the other hand, lists their breeding range as extending from Washington to Agattu. Clark (1910) noted this species in small numbers at Atka and Agattu. Because of the lack of proof, Udvardy probably had no options. I believe that Bent was probably closer to describing their original range. I base this assumption on recent observations and on the additional fragments of information reported by Gabrielson and Lincoln (1959). Murie (1959) failed to find this species anywhere in the Aleutians, but his primary reason for being there, the fox-farming industry, may have had a lot to do with his not being able to find any.

The FWS surveys in Prince William Sound in July-August 1972 located small numbers of rhinoceros auklets in breeding plumage at the Wooded Islands and at Stoney Island and Channel Island in Montague Strait. These birds gave every impression of being local breeders. David Roseneau (Isleib and Kessel 1973) encountered two at the Barren Islands in June 1965. Isleib and Kessel (1973) list a few other records from this area.

My own experience leads me to believe that there is a large colony somewhere on Afognak Island, probably on or near Tonki Cape. On 30 May 1973 I noted a lone bird north of Afognak Island. Later, on 8 and 9 August, I saw several in the same area. On 13 August in Marmot Strait I observed a number of rhinoceros auklets, either singly or in groups of up to 12. Some of these had small fish in their beaks. As they flushed, they all flew off toward Tonki Cape. This observation was made just at last light, and I believe that there were many others about that could not be seen in the dying light. We did not encounter this species along the Alaska Peninsula during the FWS survey in 1973 until we reached the end. There I had one quick glimpse of what I was certain was a rhinoceros auklet at Amagat Island.

Horned Puffin (Fratercula corniculata)

The horned puffin is one of the most abundant breeding birds in the Gulf of Alaska. There are only a few really large colonies but these birds breed just about anywhere there is a cliff (even a low one) with suitable fractures and crevices. During the Alaska Peninsula surveys in 1973, I estimated that the frequency with which these birds were seen on the water was about half that of the tufted puffin. They have been recorded in so many places that there is nothing to be gained by a reiteration of the record in the literature.

The horned puffins reach their greatest density in the Gulf of Alaska west of Kodiak Island. Murie (1959) estimated that the colony at Amagat Island, Morzhovi Bay, contained 15,000 birds, one of the largest he had seen. It contained at least 50,000 in 1973. Even at that, it was no match for the colony on Little Koniuji Island with its minimum 140,000 horned puffins. Other colonies with large horned puffin components were at High Island (40,000), Castle Rock (20,000), Mitrofani Island (35,000), and Sosbee Bay (15,000).

Earlier in this paper, I commented at length on the great and often rapid fluctuations in populations of tufted puffins. The same phenomenon affects horned puffins. In 1975 there were relatively small numbers of horned puffins at Little Koniuji where they had flourished 2 years earlier (James Bartonek, personal communication). Because they are apparently subject to erratically oscillating populations, it is hard to tell how they have fared over the years.

Tufted Puffin (Lunda cirrhata)

The tufted puffin, as previously indicated, is also a bird with widely fluctuating populations. Until we develop an understanding of their population dynamics and can understand the underlying cause of these fluctuations it will not be possible to assess trends in their populations or understand the implications of such trends.

Tufted puffins are abundant throughout the Gulf of Alaska. Small colonies can be located almost anywhere. Along the Alaska Peninsula there are a number of colonies with an estimated breeding population in 1973 of more than 15,000 birds. These are: Ashiiak Island (20,000), Central Island (90,000), the Brother Islands (45,000), The Haystacks (19,000), Castle Rock (85,000), Bird Island (none, but may contain 500,000-1,000,000 at times), Peninsula Islands (35,000), the Twins (18,000), Amagat Island (40,000), and Umga Island (22,000). These colonies correspond to the area where colonies were listed for the horned puffin.

Tufted puffin populations respond readily to some undetermined short-term perturbations. This is clearly demonstrated by their rapid population fluctuations. Because of their numbers and because of the apparent rapidity with which their numbers rebound, it is not so apparent that they have been affected by long-term perturbations, as so many other seabirds apparently have.

There is much unused or underused nesting habitat suitable for this species. In some cases there are very strong clues pointing to why this habitat is vacant. On many islands along the Alaska Peninsula, which have very good-looking tufted puffin nesting habitat and no puffins, there are visible signs of the presence of fox—either fox trails or abandoned trappers' cabins. I also suspect that the brown bear (Ursa arctos) is another possible contributing factor to population declines of burrow nesters along this coast. I have seen brown bears swimming from island to island on foraging expeditions. George J. Divoky (personal communication) has found brown bears visiting Ugaiushak Island, which is 13 km from shore. There are other islands between Ugaiushak and the mainland but the shortest route from shore would require one swim of 7 km. The motivation must be strong.

Tufted puffins may shift from colony to colony. This could be an explanation for apparent local population fluctuation, but if so, I am puzzled by the apparent tenacity with which puffins cling to some sites. Their constant occupancy of sites where the vegetative mat is breakaway tundra (Amundsen 1972) or is underlain by sand results in the destruction of these sites. Tufted puffins often cling to them in spite of the fact that they have been reduced to "slums."

My conclusion is that in spite of their large numbers it appears that tufted puffin populations in the Gulf of Alaska probably have been reduced to a level below that of their undisturbed state.

Seabird numbers in the Gulf of Alaska are not static. Generally, they are probably much less abundant than they were when Bering made his voyage of discovery. There are, nonetheless, considerable numbers of seabirds breeding along the coasts of these waters. Some species show signs of recovery from past insults by man. With enlightened management there is still time to preserve the vast natural heritage that they represent and, in many cases, to improve their status.

In attempting to address a complicated subject in short space and a relatively narrow frame of reference, I have certainly erred a number of times. I would like to see the wealth of new data that will be derived from current work applied to this concept. An understanding of past population fluctuations and the underlying perturbations that they reflect is essential for managers faced with the problem of making good decisions on measures to mitigate the potential adverse impact of development.

---

by

David A. Manuwal

College of Forest Resources
University of Washington
Seattle, Washington 98195

and

R. Wayne Campbell

Provincial Museum
Victoria, British Columbia, Canada

The purpose of this paper is to discuss the known distribution, habitat, abundance, and status of breeding seabirds of the Alexander Archipelago in southeastern Alaska, the Province of British Columbia, Canada, and the State of Washington.

Even though several studies of the breeding biology of several seabird species in this area have been published, there have been few published surveys of known breeding colonies. In British Columbia the most extensive work has been done by the British Columbia Provincial Museum and the University of British Columbia (Drent and Guiguet 1961). Gabrielson and Lincoln (1959) summarized the available literature on Alaskan birds up to about 1958. Since then, no extensive surveys have been conducted in southeastern Alaska. The U.S. Department of the Interior (1972), in its environmental impact statement for the Trans-Alaska Pipeline, presented additional information on the seabirds of other parts of Alaska. In Washington, there are no published comprehensive surveys except those of Kenyon and Scheffer (1961) and unpublished surveys by the U.S. Fish and Wildlife Service and the University of Washington.

There are 19 species of seabirds that breed along the Pacific coast of southeastern Alaska, British Columbia, and Washington (Table 1). Southeastern Alaska has the largest number (17) of species. Errors in species identification are most likely with the Larus gulls, particularly in southeastern Alaska where the herring gull (L. argentatus) and glaucous-winged gull (L. glaucescens) breed in mixed colonies (Patten and Weisbrod 1974). A similar situation exists in Washington where the western (L. occidentalis) and glaucous-winged gulls intergrade (Scott 1971). Brandt's (Phalacrocorax penicillatus) and double-crested cormorants (P. auritus) are often difficult to identify from the air. This would be a problem in Washington and the southwest coast of Vancouver Island, where the two species are locally sympatric.

The area under consideration is the 400-km-long Alexander Archipelago (Fig. 1). This complex pattern of islands, bays, and inlets is characterized by extremely high precipitation and typical cool marine temperatures. Average annual precipitation in the Sitka area is 245.4 cm (1931-60), and the average annual temperature is 6.3°C (U.S. Weather Bureau 1974). As a consequence of this cool, humid environment, most of the islands are densely covered with conifers, chiefly Sitka spruce (Picea sitkensis) and hemlock (Tsuga heterophylla), and an almost impenetrable shrub cover composed of salmonberry (Rubus spectabilis), elderberry (Sambucus callicaipa), devil's club (Echinopanax horridus), and three species of Vaccinium (Heath 1915).

There are 16 species of marine birds breeding in the Alexander Archipelago. The major seabird breeding colonies are located at Glacier Bay and at St. Lazaria, Hazy, and Forrester islands (Fig. 1; Table 2). Published surveys of these colonies are available only for St. Lazaria (Willett 1912) and Forrester islands (Heath 1915; Willett 1915). Several authors have reported on seabirds from surrounding areas (Grinnell 1897, 1898, 1909; Swarth 1911, 1922, 1936; Patten 1974). There have been no surveys of seabirds of southeastern Alaska since before the 1940's (J. G. King, Jr., personal communication). However, since census data are available for only two colonies, we discuss them in more detail.

The studies by Willett (1912, 1915) and Heath (1915) provide some base-line information on species composition and abundance with which future studies on St. Lazaria and Forrester islands can be compared (Table 2). The somewhat greater species diversity on Forrester Island is primarily due to its greater size and more suitable soil type for ancient murrelets (Synthliboramphus antiquus) and Cassin's auklets (Ptychoramphus aleutica), species that are absent on St. Lazaria. Storm-petrels (Oceanodroma spp.) are the most numerous species on both islands, but there are proportionately more storm-petrels (88%) on St. Lazaria than on Forrester (36%). On the other hand, there are many large, burrowing alcids on Forrester Island. Nearly a third of the birds on Forrester are rhinoceros auklets (Cerorhinca monocerata), tufted puffins (Lunda cirrhata), and horned puffins (Fratercula corniculata).

The species composition of seabirds breeding on other islands is similar to that found on Forrester and St. Lazaria islands but less abundant. In Glacier Bay, for example, the only population data available are those provided by Patten (1974) for North Marble Island: pelagic cormorants, Phalacrocorax pelagicus (30 pairs); black oystercatchers, Haematopus bachmani (8); herring gulls (7); glaucous-winged gulls (500); common murres, Uria aalge (18); pigeon guillemots, Cepphus columba (60); horned puffins (4); and tufted puffins (30).

At the present time, it is impossible to draw any conclusions about changes in population density and distribution for most of the seabirds breeding in southeastern Alaska. Adequate data are available only for St. Lazaria and Forrester islands where Willet and Heath provided the only early extensive census data for this part of Alaska.

The rugged British Columbia coastline is characterized by 930 km of islands and inlets (Figs. 2, 3). With the exception of the inner southern portion, this coast is mostly uninhabited. The physical characteristics of the offshore islands are similar to those found off the Washington coast. Descriptions of some of these islands and the 15 species of breeding seabirds on them have been given by Drent and Guiguet (1961), Guiguet (1971), and Summers (1974).

A detailed analysis of British Columbia seabirds is not presented here since a more thorough analysis is in preparation by R. W. Campbell and R. H. Drent (manuscript). Instead, we present seabird population estimates available for the Province up to the summer of 1975; Tables 3 and 4 summarize these estimates for the five major portions of coastal British Columbia. The coast of British Columbia contains a myriad of small islands where there may be small numbers of breeding seabirds. Many of these have not been censused and are too numerous to include in Tables 3 and 4.

More than half of the breeding seabirds in British Columbia are found on the east coast of the Queen Charlotte Islands, and the fork-tailed storm-petrel (Oceanodroma furcata) comprises more than half of that total. However, new unpublished data (K. Vermeer) for Triangle Island and the northwest coast of Vancouver Island indicate that the population figures in Table 3 for this area are underestimates. Nevertheless, these two regions have nearly 80% of all the breeding seabirds in the Province. This results from the very large populations of the rhinoceros auklet and tufted puffin on Triangle Island and the fork-tailed storm-petrel, ancient murrelet, and Cassin's auklet on various islands on the east coast of the Queen Charlotte Islands (Table 3).

Continuing surveys of breeding seabirds are being conducted by personnel of the British Columbia Provincial Museum and the Canadian Wildlife Service.

General Environment

For this report, we have distinguished two major geographical areas in Washington where breeding seabirds are found—the western coast of the Olympic Peninsula and the San Juan Islands, including the Strait of Juan de Fuca.

On the Olympic Peninsula, seabirds breed on the offshore rocks, islands, and precipitous cliffs from Copalis Beach to Cape Flattery (Fig. 4). The offshore rocks and islands throughout this area (except Tatoosh Island) are now included in the Washington Islands National Wildlife Refuge. Most of the larger rocks and islands have dense stands of salmonberry, salal, and grasses, and a few support stands of stunted conifers (Fig. 5); most are inaccessible to man. The adjacent coast is dominated by the Olympic rain forest where the mean annual precipitation is about 337.1 cm (U.S. Weather Bureau 1956, 1965a, 1965b).

Because the San Juan Islands lie northeast of the Olympic Peninsula and east of Vancouver Island (Fig. 6) they are in a rain shadow; however, because of highly variable topography and aspect, most islands have a diverse assemblage of plant communities (Franklin and Dyrness 1973). Exposed south-facing slopes are occupied by grassland vegetation and frequently by scattered trees, usually Pseudotsuga menziesii and Arbutus menziesii. Most of the seabird colonies are located on rather small exposed islands with short, grassy, shrubby vegetation. In general, these islands are not suitable for burrowing species.

In the Strait of Juan de Fuca, the two most important sites are Smith and Protection islands. Both are composed of glacial deposits and heavy sod that has developed under dense grassy vegetation (Fig. 7). Consequently, these two islands support most of the burrowing seabirds in the region. Unfortunately, both islands have historically been subjected to much human disturbance (Richardson 1961; Manuwal 1974).

The existing information on seabird colonies in both the coastal and San Juan Island areas has been largely derived from aerial surveys by the U.S. Fish and Wildlife Service. These surveys are inherently biased toward surface-nesting species such as gulls and cormorants. Population estimates for guillemots, auklets, storm-petrels, and puffins are less accurate. Some additional information obtained by direct island visitation has been provided by Kenyon and Scheffer (1961), Richardson (1961), Thoresen and Galusha (1971), G. Eddy (unpublished data), and D. A. Manuwal (unpublished data). Although other accounts of Washington seabirds are available, the references listed above are specifically oriented toward population assessment.

Olympic Peninsula

Despite the large number of offshore rocks, islets, and islands along the Pacific coast of Washington, significant seabird colonies are present only on about 30 islands. Since Table 5 summarizes the population estimates for 12 species of seabirds breeding on 24 major sites, it represents only the majority and not the total number of breeding seabirds on the Pacific coast of Washington. About 74% of the entire Washington seabird population resides on the coastal rocks and islands.

Major colony sites with more than 2,500 breeding pairs are Grenville Arch, Willoughby Rock, Destruction Island, Cake Rock, Carroll Island, and Bodelteh Island. More intensive censusing, especially of nocturnal burrowing species will undoubtedly raise the population estimates for these and other islands off the coast. About 54% of the total coastal population is composed of the common murre and rhinoceros auklet.

San Juan Islands

There are about 86 actual or potential seabird colony sites in this area; 25 (30%) are now considered important. Eleven islands are under Federal protection as National Wildlife Refuges. Part of Protection Island is owned by the Washington State Game Department to protect the largest rhinoceros auklet colony in the State. Most colony sites are on small islands with poorly developed soil which prevents burrowing species from using them. Consequently, the dominant species are surface nesters (such as gulls and cormorants) and rock-crevice nesters (like the pigeon guillemot). In all, about 31,000 seabirds of 7 species breed in the San Juan Island area. Breeding seabird population estimates for 49 of the 86 nesting sites are given in Table 6. Even though this does not represent all the colonies, it covers the most important islands and those islands where there appears to be potential for seabird breeding.

The major colony sites with more than 250 breeding pairs are located at Protection and Smith islands, Bird Rocks, Colville Island, Hall Island, North and South Peapod rocks, Puffin Island, North Sisters, Viti Rocks, and Williamson Rocks (Fig. 6). Glaucous-winged gulls are the predominant species on all these islands except Protection and Smith islands, where there are large colonies of rhinoceros auklets. Rhinoceros auklets (65%) and glaucous-winged gulls (32%) make up 97% of the total San Juan Islands seabird population.

Food supply and availability of nest sites are two critically important factors influencing the distribution and abundance of seabirds. Whereas information on general diet composition is known for most seabird species, we know little about the availability of favored seabird prey. The dynamics of seabird food chains is reviewed elsewhere in these proceedings.

The nest-site preferences for seabirds of the northeast Pacific Ocean are given in Table 7, and Table 8 indicates the proportion of seabirds that belong to specific nest-site categories. These preferences, in conjunction with knowledge of the physical characteristics of seabird habitat, permit a partial explanation of the present distribution and abundance of seabirds. For example, if we compare the San Juan Island habitats with those of the Washington coast, it is apparent that there are more cliff-nesting species on the coast. This reflects the physical characteristics of the two habitats. There are few cliffs in the San Juan Islands, and those that exist are very unstable. Colony sites in the San Juan Islands are typically on low, flat islands. Glaucous-winged gulls are the most abundant nesting species there. Coastal islands, on the other hand, are either covered by dense vegetation or are large monolithic chunks of rock with few available flat areas. Population estimates for the Washington coast are heavily biased toward surface nesters, since most of the data have been gathered by aerial surveys. Consequently, the burrow and rock crevice categories are underestimated. The aerial survey is appropriate for only about 43% of the birds nesting on the Washington coast.

Northern and southern British Columbia provide another good example of habitat availability as revealed through seabird population estimates. The population data are more comprehensive and have largely been gathered by island visitations. The islands in the northern portion are heavily vegetated and many have well-developed soil into which storm-petrels, auklets, and murrelets can burrow. Indeed, 96% of the seabird population consists of nocturnal, burrow-nesting species. In southern British Columbia, however, there are more open-nest species, particularly glaucous-winged gulls and cormorants.

Overall, 68% of the breeding seabirds found along the northeastern Pacific coast are nocturnal and nest in burrows or rock crevices (Table 8). The most conspicuous nesting birds such as gulls, cormorants, and murres, comprise only 22% of the total population. Consequently, our current estimates of breeding seabirds still underestimate the more secretive, nocturnal, burrow-nesting species.

The available data are inadequate to detect changes in population distribution and density for most species (Table 9). In Washington, for instance, limited unsubstantiated information suggests an overall decline of the double-crested cormorant and tufted puffin in the San Juan Island area. Likewise, there seems to be an increase in glaucous-winged gulls there. In British Columbia, Drent and Guiguet (1961) were able to detect changes in some species. For example, they noted increases in the double-crested cormorant, pelagic cormorant, and glaucous-winged gull. No change was observed in the tufted puffin. Since then, the Brandt's cormorant has established a colony in Barkley Sound (Guiguet 1971). The data in southeastern Alaska are inadequate for all species except, perhaps, the Cassin's auklet which Gabrielson and Lincoln (1959) reported to be declining throughout Alaska. In short, no definitive statements can now be made concerning changes in seabird population numbers.

Species Accounts

Fork-tailed Storm-petrel (Oceanodroma furcata)

Storm-petrels are especially difficult to census because they are nocturnal, and the burrows and rock crevices where they breed are often difficult to locate, especially in mixed-species colonies. The census data are inadequate to determine whether there have been changes in population density and distribution. Indeed, the biology of this species is perhaps the least known of the North Pacific colonial seabirds. In southeastern Alaska, this species is outnumbered by at least 5 to 1 by the Leach's storm-petrel (Oceanodroma leucorhoa). The reasons for this are poorly understood. There is some evidence that the numbers of breeding fork-tailed storm-petrels on Forrester Island may fluctuate drastically from one year to the next (Gabrielson and Lincoln 1959).

Leach's Storm-petrel (Oceanodroma leucorhoa)

Of the two subspecies of this petrel (O. l. leucorhoa and O. l. beali), only O. l. beali is found in southeastern Alaska. The leucorhoa subspecies is more northerly in distribution. Where both fork-tailed and Leach's storm-petrels are sympatric, Leach's predominates; however, this relationship becomes more unpredictable in British Columbia and Washington. This species is undoubtedly widespread in the forested islands of the Alexander Archipelago.

Double-crested Cormorant (Phalacrocorax auritus)

The double-crested cormorant apparently does not breed in southeastern Alaska since Willett (1912), Gabrielson and Lincoln (1959), and S. Patten (personal communication) do not report breeding colonies for the area. The largest populations occur in southern British Columbia principally in the Gulf Islands, where 71% of all breeding double-crested cormorants are found (Table 10). According to Jewett et al. (1953), this species was less common in Puget Sound than was Brandt's cormorant, but is certainly not the case today (D. A. Manuwal, unpublished data). The only common cormorants in the San Juan Islands are the pelagic and double-crested species. The double-crested cormorant seems to have declined in numbers on both coastal and inland waters. On the basis of his observations, R. W. Campbell believes that this species is increasing in British Columbia.

Brandt's Cormorant (Phalacrocorax penicillatus)

Brandt's cormorant is the least abundant of the three cormorant species that nest in the study area. Washington is at the northernmost edge of the breeding distribution of this species. Only one more northerly colony exists, on Sartine Island off Vancouver Island (Vermeer et al. 1976). Brandt's cormorant comprises about 85% of the cormorant population in Oregon (U.S. Fish and Wildlife Service, unpublished data). However, in Washington it is only about 9% and in British Columbia 3% of the total cormorant population.

Comparing information in Jewett et al. (1953) with the current situation, it is apparent that there has been a drastic change in the distribution and probably in the numbers of this species in Washington. Today, there are no Brandt's cormorant colonies in the San Juan Islands or Strait of Juan de Fuca. Yet Jewett et al. (1953) reported colonies at Bellingham Bay and on Lopez and Matia islands. We have observed juvenile Brandt's cormorants in the San Juan Islands during the summer. This species may be particularly susceptible to human disturbance, since all three areas listed above are heavily used in the summer for recreation.

Pelagic Cormorant (Phalacrocorax pelagicus)

The distribution of breeding colonies of the pelagic cormorant is strongly determined by the availability of the steep cliffs on which it constructs its nest. This is the only common cormorant in southeastern Alaska. Throughout its extensive range, this species is generally found breeding in small numbers. Nothing is known about fluctuations in its numbers in Alaska.

This species is common in both British Columbia and Washington; nesting sites are of the same type as those in Alaska except in the San Juan Islands, where 200-300 birds nest on cliff faces composed of glacial deposits. Here, there is frequent nest loss due to slippage off the cliff face; this loss is especially severe on Smith and Protection islands. There do not appear to be any changes in the distribution of pelagic cormorants, but an accurate assessment of abundance is impossible from the data currently available.

Glaucous-winged Gull (Larus glaucescens)

The glaucous-winged gull is the characteristic gull of southeastern Alaska and British Columbia. In Washington, it is the dominant gull in the San Juan Island area but interbreeds with the western gull on the Washington outer coast from Tatoosh to Copalis Beach (Scott 1971). In Alaska, it is widely distributed and locally abundant on Forrester Island, St. Lazaria, and throughout Glacier Bay (S. Patten, personal communication). The biology of this species has been extensively studied in the southern part of its range, especially by Vermeer (1963) and James-Veitch and Booth (1954). The only study of the breeding biology of this species in southeastern Alaska is by Patten (1974) for Glacier Bay. Glaucous-winged gulls are apparently increasing in British Columbia (R. W. Campbell, unpublished data) and in Washington (T. R. Wahl, personal communication). This increase is undoubtedly a result of the proximity of breeding colonies to garbage dumps and commercial fishing fleets in both Canada and the United States. Little is known about changes in populations of gulls in southeastern Alaska.

Western Gull (Larus occidentalis)

The western gull is the common breeding gull of the Washington outer coast; however, there is increased interbreeding with glaucous-winged gulls northward from Destruction Island to Tatoosh Island. The percentage of glaucous-winged gulls steadily increases until Vancouver Island and the Strait of Juan de Fuca, where western gulls are rare. Population estimates of gulls on the outer coast of Washington are derived primarily from aerial flights. This makes identification of gulls difficult, and in view of the amount of interbreeding, it is probably impossible to classify many of the breeding gulls as to species. Western gulls appear to be increasing in the Grays Harbor area (G. D. Alcorn, personal communication).

Herring Gull (Larus argentatus)

The herring gull is typically found in inland Alaska but can be found uncommonly along the coast of southeastern Alaska, where it often forms mixed colonies with glaucous-winged gulls. These two species apparently hybridize where they are sympatric (Williamson and Peyton 1963; Patten and Weisbrod 1974; Patten 1974).

Black-legged Kittiwake (Rissa tridactyla)

The black-legged kittiwake is found only in the northern portions of southeastern Alaska. It apparently is a common breeding bird in Glacier Bay National Monument (S. M. Patten, Jr., personal communication). No population estimates are available for this species other than that it is locally abundant.

Common Murre (Uria aalge)

Common murres are common in southeastern Alaska and the coast of Washington but breed only in small numbers in British Columbia and are absent in the San Juan Islands. Since this species usually prefers cliffs or the tops of inaccessible rocks, they are probably limited by island topography in British Columbia, and are most certainly so limited in the San Juan and Gulf Island groups.

In Alaska, common murres breed in unknown numbers in Glacier Bay and in large numbers on St. Lazaria, Forrester, and the Hazy islands. No data on population changes are available for any of the three regions.

Pigeon Guillemot (Cepphus columba)

The pigeon guillemot is common throughout the region from Cape Fairweather to Washington. Even though it is not truly colonial, it may be locally abundant where there are suitable nest sites. Since these nest sites are usually difficult to find, population estimates are seldom accurate, usually being conservative. It is evident that guillemots appear to be small in number when compared with other seabirds nesting at major colony sites in the north Pacific region (Table 10). This disparity may be exaggerated by the difficulty of censusing guillemots.

Marbled Murrelet (Brachyramphus marmoratus)

Since the marbled murrelet has been found to nest in coniferous forests (Binford et al. 1975), traditional census techniques are unsuitable. This species is common in southeastern Alaska (Gabrielson and Lincoln 1959), in British Columbia (Drent and Guiguet 1961), and in Washington (Jewett et al. 1953).

Kittlitz's Murrelet (Brachyramphus brevirostris)

The difficulties in assessing breeding populations of Kittlitz's murrelet are the same as those for the marbled murrelet. This species nests on the ground at high elevation near the coast (Bailey 1973). The largest concentrations are in the vicinity of Glacier Bay National Monument (Gabrielson and Lincoln 1959). They are not found breeding in Washington or British Columbia.

Ancient Murrelet (Synthliboramphus antiquus)

Ancient murrelets appear to be locally common throughout southeastern Alaska. Their presence is probably strongly dependent upon a suitable soil in which to excavate burrows. The only available population estimates are those by Willett (1915) for Forrester Island (Table 1). Censusing this species is especially difficult because its burrows are easily confused with those of Cassin's auklet. There are no studies of this species in southeastern Alaska; however, it has been well studied in the Queen Charlotte Islands to the south by Sealy (1975).

Cassin's Auklet (Ptychoramphus aleuticus)

A synthesis of literature and unpublished observations led Gabrielson and Lincoln (1959) to conclude that Cassin's auklet has greatly decreased in numbers and is not abundant anywhere in Alaska. They also concluded that the colony on Forrester Island (Table 1) was the only well-documented colony in southeastern Alaska. Fishermen in the southeastern Alaska area occasionally see this species (M. E. Isleib, personal communication), but it is apparently still uncommon though more widespread than just Forrester Island. The nocturnal habits and burrowing in dense vegetation makes censusing this species very difficult. Nothing is known about the ecology of this species in Alaska.

Rhinoceros Auklet (Cerorhinca monocerata)

Rhinoceros auklets seem to be found breeding only on islands where there is a well-developed soil in which to excavate their extensive burrows. From the limited evidence available, it appears that the largest rhinoceros auklet populations probably are to be found in southeastern Alaska. Willett (1912) found a very large population on Forrester Island (Table 2), and the species has been found in the summer in the Barren Islands east of Kodiak Island (E. P. Bailey, personal communication). More intensive surveys of the Alexander Archipelago will probably reveal other populations of this species.

This species is less common in British Columbia than either Alaska or Washington. A possible reason for this is lack of suitable nesting areas. In Washington, the two largest colonies are at Protection Island in the Strait of Juan de Fuca and Destruction Island on the outer coast. Smaller numbers exist on other coastal islands and on Smith Island in the Strait of Juan de Fuca. The Smith Island colony is an interesting one since it appears that early human disturbance in the late 19th or early 20th century eliminated the species from the island. In their discussion of Smith Island, Jewett et al. (1953) made no mention of auklets, only of puffins and guillemots. Couch (1929) did not record the species in 1925. The colony now numbers about 600 pairs.

Horned Puffin (Fratercula corniculata)

Although the horned puffin is one of the most abundant seabirds in other parts of Alaska, it is much less abundant in the southeastern portion. In addition to the information discussed by Sealy (1973), it now appears that this species may breed as far south as Triangle Island, British Columbia (K. Vermeer, personal communication; D. A. Manuwal, personal observation). Here, as on Forrester Island, it is greatly outnumbered by the tufted puffin. No data are available on the breeding or status of this species in the study area.

Tufted Puffin (Lunda cirrhata)

The tufted puffin is found breeding on scattered islands throughout the region. The largest known colonies are on Forrester Island, Alaska, Triangle Island, British Columbia, and Carroll Island, Washington. It is notably absent from most of the gulf and San Juan Islands. Even though puffins have apparently never been numerous in the San Juan Islands, their population has noticeably declined during the past 35 years. For example, Jewett et al. (1953) reported a colony of 50 pairs on Bare Island in 1937, but in 1973 only 2 pairs were counted (D. A. Manuwal, unpublished data). Likewise, in 1915 there were more than 250 pairs on Smith Island, but by 1916 there were only 75 pairs (Jewett et al. 1953). The decline is attributed to rapid erosion of the glacial-deposit cliffs. There are no puffins on Smith Island today, and the largest colony in the Puget Sound area is the 35 pairs on Protection Island (D. A. Manuwal, unpublished data).

The total minimum estimate of the breeding seabird populations of British Columbia and Washington is 216,500 pairs (Table 10). No comprehensive estimates are available for breeding seabirds of southeastern Alaska. It is likely, however, that the number of breeding seabirds in the Alexander Archipelago may be equal to (or exceed) the populations of both British Columbia and Washington. Data are desperately needed from that area. Of the total seabird population in the study area (Table 10) 43% reside in northern British Columbia. The Washington State population represents 28% of the total. Fork-tailed storm-petrels comprise almost 25% of all the breeding seabirds in the area under consideration. The Cassin's auklet is the next most numerous species (18% of the total).

It is apparent that current data are, for the most part, inadequate for assessing anything but catastrophic changes in seabird breeding colonies. This inadequacy is due to inadequate censusing because of excessive reliance upon aerial surveys; in the past, this has often been a result of insufficient funding.

Of the several threats facing seabird populations, none may be as important as oil pollution. A general review of this subject is presented elsewhere by Vermeer and Vermeer (1975). It is apparent from this review that the most vulnerable species are those that dive beneath the sea surface, including all the alcids and cormorants breeding along the coast that are discussed in this paper. This group makes up almost 60% of all the breeding seabirds in this area. Unfortunately, our knowledge of several of these species is scanty and our current census techniques are unsuitable for most of these birds.

Studies of the changes in seabird numbers have been made in other oceans. For example, in Great Britain (Bourne 1972a, 1972b; Harris 1970), eastern Canada (Nettleship 1973), and the Atlantic coast of the United States (Kadlec and Drury 1968), two major trends seem apparent. First, there is an overall decline in alcid and tern numbers. The decline in auks may be due to their extreme vulnerability to oil pollution (Bourne 1972a, 1972b; Vermeer and Vermeer 1975). The Atlantic puffin, however, may be suffering the additional effects of gull cleptoparasitism (Nettleship 1972). Secondly, there seems to be an increase in gull populations on both sides of the Atlantic, particularly the herring gull and black-legged kittiwake.

Compared with the Atlantic coast of North America and northern Europe, the data base for seabird populations of the Pacific coast is poor. The fragmentary evidence now available indicates that there may be small population increases in the western and glaucous-winged gulls and range extensions of the Brandt's and double-crested cormorants and of the rhinoceros auklet (Scott et al. 1974). Whether these changes represent actual population increases or displacements remains unclear. The remote locations of most of the large Pacific seabird colonies may provide unofficial protection from human interference. Intensive surveys are needed to establish base-line inventories in these areas.

As a consequence of this first comprehensive review of the status of breeding marine birds of the northeast Pacific coast of North America, we recommend the following future research topics as necessary for the conservation of this great international resource.

---

by

David G. Ainley

Point Reyes Bird Observatory
Stinson Beach, California 94970

and

Gerald A. Sanger[15]

National Marine Fisheries Service
Marine Mammal Division
Seattle, Washington

The ecology, morphology, and much of the behavior of a seabird species are definable in terms of the food resources it exploits year-round and the spatial and temporal relations between food and breeding sites. This general point unifies such important reports as those by Kuroda (1954), Bédard (1969a), Ashmole and Ashmole (1967), Ashmole (1971), Spring (1971), and Sealy (1972). More concretely, information on trophic relations of seabirds is useful in several ways. In conjunction with biological oceanographic data, it can provide insight into geographic location, marine habitat, depth, time of day, and general method of food capture by seabirds. Collected over several years, it can provide a basis for understanding annual differences in seabird breeding phenology and success. Finally, supplemented with data on how much seabirds eat and excrete, it is necessary for an understanding of the energetic and ecological roles played by the birds in the functioning of marine ecosystems.

Several studies that describe trophic relations within seabird communities have helped to define the principals of community organization pertaining to the exploitation of available food resources and have given clues to food-chain pathways. Trophic relations have been described for breeding communities in the Barents Sea (Uspenski 1958; Belopol'skii 1961), in the tropical Pacific Ocean (Ashmole and Ashmole 1967; Ashmole 1968), in the North Sea (Pearson 1968), and in the Chukchi Sea (Swartz 1966). The last-named study pertained most directly to the geographic region discussed in this paper, but several other studies have provided sound information on segments of communities in the northeastern North Pacific and Bering Sea. These include the work on three species of auklets (Aethia, Cyclorrhynchus) in the Bering Sea (Bédard 1969a); investigations on cormorants and other fish predators in British Columbia by Munro (1941), Munro and Clemens (1931), and Robertson (1974); studies of murres in Bristol Bay by Ogi and Tsujita (1973); observations on several species near the Pribilof Islands by Preble and McAtee (1923); work on diving species off Oregon by Scott (1973); and studies of murrelets by Sealy (1975).

A review of available reports reveals three obvious gaps in the emphasis placed in seabird food studies. First, few studies have ever considered in detail the trophic relations of seabird communities during the winter or nonbreeding season. Partial exceptions are the works by Cottam (1939) and others on marine diving ducks, species that are seabirds only during the winter, and by several researchers (Munro and Clemens 1931; Munro 1941; Robertson 1974) on seabirds in British Columbia. Divoky (1976) studied diets of pack-ice gulls during the nonbreeding season, but those species are not included in the present analysis because they rarely are found south of the Bering Strait. Second, no study has considered the trophic relationships of an entire seabird community, i.e., not just breeding species but also nonbreeding species. In the rather broad communities considered here, 50-70% or more of the birds breed in another part of the world. To say that these nonbreeding species have no significant impact on resource exploitation or on organization and evolution among breeding members would be naive. Finally, few investigators have attempted to fit birds into an entire ecosystem, including lower trophic level origins as well as fish, marine mammals, and man.

The reasons for these gaps in study emphasis are readily apparent: the inconvenience of marine research during the winter when weather is stormy, the need for costly study platforms (boats), and the difficulties in organizing the specialized community of biologists required for such tasks. A less obvious but important reason is that oceanographers and fishery biologists have overlooked seabirds as important members of marine ecosystems.

Relatively good information exists for most pelecaniformes of the region. A notable exception is the brown pelican (Pelecanus occidentalis), an endangered species. This is unfortunate because dietary information is important for understanding the species' ecology. Observations in eastern North America (Palmer 1962) and Peru (Murphy 1936) indicated that their diet consisted of fish that occur at the surface. The larger cormorants are piscivorous, particularly on schooling fishes that occur at moderate to great depths (Table 1). The smaller cormorants feed more heavily on benthic fish and decapod crustaceans. Cormorants apparently feed only during daylight and then only for short periods because their wettable plumage loses its buoyancy. Thus they remain relatively close (50 km) to nesting and loafing areas.

Information on diets of marine ducks (Table 2) is more nearly complete than for most other seabirds. These birds fall into four groups with some overlap: species feeding on plants (Branta, Philacte, Anas-type, and Somateria fischeri); those feeding on benthic crustaceans (Clangula hyemalis, Histrionicus histrionicus, Polysticta stelleri, S. mollissima); those feeding on benthic molluscs (Somateria spp. and Melanitta spp.); and those feeding on fish (Mergus serrator, Clangula hyemalis, and Melanitta deglandi). A study by Perthon (1968), one of the few on a seabird's diet during most of a year, showed a seasonal change in diet for S. mollissima in Norway. In general, waterfowl seem to specialize in their diets much more than other seabirds and, for that reason, are perhaps more restricted in their distributions. Some marine ducks are known to dive to considerable depths (reviewed by Kooyman 1974), but usually they occur in shallow waters where plants and sessile invertebrates are readily available.

The summer diet of the pigeon guillemot (Cepphus columba) is the best known among seabirds in the region being considered here (Table 3). Only in the extreme southern part of its range (i.e., the California Channel Islands) is there no information available on its diet. The species feeds on organisms, mostly fish, from rocky habitat and apparently can dive to considerable depths (Follett and Ainley 1976). Because so much is known about guillemot diets during summer, a study of the winter diet would be valuable.

The diets of other alcids are known well enough to at least characterize them broadly. The larger species, murres, tufted and horned puffins (Lunda cirrhata, Fratercula corniculata), and the rhinoceros auklet (Cerorhinca monocerata), feed heavily on fish, mainly species that school in midwater (Table 4). To a great degree, these birds are opportunistic, feeding rather heavily at times on cephalopods and crustaceans, particularly nektonic forms. Morphological differences between the two murre species suggest that thick-billed murres (Uria lomvia) feed on benthic organisms much more than do common murres (U. aalge), and that the latter species is more piscivorous (Spring 1971); however, field data on diets are barely adequate to confirm this. Ogi and Tsujita (1973) analyzed the stomach contents of murres drowned in salmon gill nets but did not separate the two species. For the present paper we considered them to be mostly U. aalge, since this species predominates in the region of the food study (Bartonek and Gibson 1972). Adult murres sometimes eat different items than they feed to their chicks (Spring 1971; Scott 1973). The smaller alcids, ancient and marbled murrelets—Synthliboramphus antiquus and Brachyramphus marmoratus—(Table 5) and auklets (Table 6), feed on macrozooplankton: crustaceans, and fish and squid larvae. Little is known about the food or feeding ecology of Kittlitz's murrelet (B. brevirostris). Its diet is probably similar to that of the other murrelets, especially the marbled murrelet, its allopatric congener, but the diets of the other murrelets differ somewhat (Bédard 1969b; Sealy 1975). The Kittlitz's murrelet's shorter bill suggests that it feeds more on invertebrates. Alcids feed in deep or shallow water, depending on food distribution. Some alcid species can be found at great distances from land, particularly in winter (Hamilton 1958; Scott et al. 1971).

Information on the diets of other seabirds in the region is fragmentary and sometimes rather anecdotal. A little is known about the feeding habits of loons (Gavia spp.) and grebes (Podiceps spp. and Aechmophorus occidentalis), especially off British Columbia (Table 7). The larger of these birds feed mainly on inshore fish, but as species become progressively smaller, there is a tendency toward eating crustaceans. Work by Madsen (1957) in Denmark, indicated that loons and grebes tend to take prey near or on the bottom. Much more information is available on these birds' diets at their freshwater breeding sites but this provides only partial insight into what they might eat in marine habitats.

Information is especially poor for albatrosses and petrels (order Procellariiformes) (Table 8). Yet, based on sheer numbers alone, members of this diverse group are easily among the most ecologically dominant of the region (Sanger 1972; Ainley 1977). The Laysan albatross (Diomedea immutabilis) seems to be a squid specialist; the black-footed albatross (D. nigripes), northern fulmar (Fulmarus glacialis), scaled petrel (Pterodroma inexpectata), and the fork-tailed and Leach's storm-petrels (Oceanodroma furcata and O. leucorhoa) appear to be large, medium, small, and tiny versions, respectively, of surface-feeding generalists that eat whatever they can find, including live and dead fish, squid, coelenterates, crustaceans, and other organisms. The shearwaters (Puffinus spp.) feed to an unknown degree on schooling fish, squid, and crustaceans that occur near the surface. For these very abundant shearwaters, that, unfortunately, is close to the extent of our knowledge both for the North Pacific, where they winter, and the South Pacific, where they breed. Most petrels remain in oceanic habitats, but shearwaters, particularly the sooty shearwater (Puffinus griseus), and sometimes fulmars feed close to, if not within, the inshore neritic habitat. A much better understanding of the diets of this group is sorely needed.

Knowledge on the food of gulls, shorebirds, and related species is surprisingly scanty in view of all that is known about their breeding biology and social behavior. Little is known about the marine food of phalaropes, but by inference from their association with storm-petrels, plankton-feeding whales, and convergence lines (Martin and Myers 1969), their tiny size, and their method of feeding (picking at minuscule items on the water surface), one can guess that they feed on zooplankton and detritus. Skuas (Catharacta skua) and jaegers (Stercorarius spp.) apparently eat what they can find at the surface, as well as whatever they can steal from gulls and terns. Almost all the literature on their feeding (Bent 1946) dwells on accounts of their stealing from other birds. That spectacular behavior would seem to be so energetically costly, though, that it is probably less important than we have been led to believe. Rather surprisingly, the question of what foods the gulls and terns eat in the eastern North Pacific is difficult to answer from the literature (Tables 9 and 10). Some information exists for five of the larger larids at isolated places, but little is known about food elsewhere in their respective ranges, and the diets of the seven smaller gulls and the terns are practically unknown. Studies on gull diets in the Atlantic region (e.g., Spaans 1971; Harris 1965) provide information on what to expect from the same species in the Pacific, but that information must be considered only in general terms because, the birds being somewhat opportunistic, their diets differ greatly from one locality to another (Ingolfsson 1967). A few observations are available for arctic terns (Sterna paradisaea) in Alaska, but little information exists for other terns (Table 10). Bent (1921) noted that Aleutian terns (S. aleutica) sometimes associate with arctic terns during feeding.

Finally, we must include raptors, particularly the peregrine (Falco peregrinus) and bald eagle (Haliaeetus leucocephalus), because they are important predators on the smaller seabirds (White et al. 1971, 1973). Peregrines have, in fact, been observed feeding on storm-petrels far at sea (Craddock and Carlson 1970).

We have compared and summarized in general terms the food partitioning by species in five rather broad oceanographic regions and their subdivisions in the northeastern North Pacific and Bering Sea, based on the specific details on diets presented in Tables 1 through 10. The five broad regions, defined oceanographically by Dodimead et al. (1963) and Favorite et al. (1976) and modified by Sanger (1972), are shown in Fig. 1. The five oceanographic regions (domains) were divided further into inshore neritic, offshore neritic, and oceanic habitats (Sanger and King, this volume). We did not include estuarine habitats or sheltered bays in the analysis.

The oceanic habitat includes waters of the photic zone overlying the deep ocean and continental slopes beyond the continental or insular shelves. The Bering Sea and central subarctic domains are largely made up of oceanic habitat. The other three domains include both inshore and offshore neritic as well as some oceanic habitat. The boundary between the inshore and offshore neritic has yet to be defined in terms of bird life, but it lies at that line beyond which the bottom is too deep for a diving bird to exploit. A depth contour thus defines the boundary. In the antarctic South Pacific, emperor penguins (Aptenodytes fosteri) dive to depths of 275 m, but so far as is known, no comparable bird exists in the North Pacific. Some marine ducks and loons reportedly dive to 50-60 m (Kooyman 1974). The inshore-offshore neritic boundary for seabirds may lie near the 70-m depth contour.

Food resource partitioning by seabirds in the five oceanographic domains are shown in Tables 11-15. Within each domain, the common and usual members of the seabird community are listed, and the major and minor categories in each of their diets are shown (on the basis of available literature, Tables 1-10). The categories are grouped further, and rather tenuously, according to the trophic level at which a bird is presumably feeding: I = herbivore, II = secondary carnivore, III = tertiary carnivore, IV = final carnivore, and Sc = scavengers (carnivorous) feeding at many levels. Birds at level I feed on large algae and seed plants and are not directly part of the same food webs involving other species. These food webs originate with phytoplankton (Fig. 2). So far as is known, no bird feeds on phytoplankton and few, if any, feed on microzooplankton; hence birds do not generally feed as primary carnivores. An exception at times might be the least auklet (Aethia pusilla) when it feeds on small copepods (see Bédard 1969b).

The above groupings are "tenuous" because prey in each category may represent more than one trophic level, and a single prey species could occur at one level one day or place and at another level the next day or place, depending upon what it happened to be eating. This is shown in Fig. 2, where the parakeet auklet (Cyclorrhynchus psittaculus) can occur in the food web at different levels, depending both on the prey it is eating and on what its prey is eating. Even without this complication, many seabirds feed at more than one level in the food web. For instance, murres eating euphausiids would be feeding at a different level than murres feeding on larger fish. It might be "safer" to regard prey organisms in level II as macrozooplankton, prey organisms in level III as micronekton, and prey organisms (seabirds themselves) in level IV as macronekton (after Sverdrup et al. 1942).

Information contained in Tables 11-15 can be summarized to show characteristics of seabird trophic relations. One such characteristic is the range of diet breadth or diet complexity (Table 16). Few species (about 6%) feed on only one type of prey and might, therefore, be referred to as "specialists." Included are eared grebe (Podiceps caspicus), Laysan albatross, brown pelican, emperor goose (Philacte canagica), black brant (Bernicia bernicla), peregrine falcon, and whiskered auklet (Aethia pygmaea). Consideration of these species as specialists may require revision when more data become available. Except for the albatross and auklet, these species are members of the inshore neritic cohort. Food specialization does not seem to be characteristic of oceanic birds in particular or of most seabirds in general.

Most species (roughly 53% in any community) include two or three prey categories in their diets—usually midwater schooling fish, squid, and crustaceans. These birds include the most numerous in the communities—the shearwaters and some alcids—which feed largely on three prey types, and also include some of the less abundant birds, the marine ducks, which feed mostly on two prey categories.

The remaining seabirds are more general in their feeding. Many have large populations, but are not as abundant as shearwaters or most alcids. The true "generalists" are the species that feed on as many as eight or more types of prey, and relatively few (12%) such species exist in each avian community. These birds, the scavengers, include black-footed albatross, fulmar, storm-petrels, and large gulls. The petrels are the scavengers of the oceanic habitat and the gulls are their counterparts in the neritic habitat (but see Sanger 1973).

Another comparison is shown in Table 17, where the species in each community are categorized according to the number feeding at each trophic level. If a species feeds at more than one level, it is tallied once in each level. Most seabirds (66-77%) feed at the second and third levels as secondary and tertiary carnivores. Few feed as terminal carnivores, and relatively few are scavengers. Actually, most scavenging occurs at levels II and III, so about 90% of the seabirds in each community feed at levels II and III. Communities including an inshore neritic feeding element are the only ones that include herbivores, and even then, few of these species exist in significant numbers in the marine environment (discounting estuaries and sheltered bays).

It is readily apparent from the foregoing comparisons that much overlap exists in the prey eaten by seabirds within each community. The question whether real competition ever exists is academic. Competition perhaps exists only rarely because seabirds partition resources through use of different feeding methods, selection of different-sized prey, and habitat zonation. Table 18 lists feeding methods (after Ashmole 1971 and Ainley 1977) and the body size and bill length of each species considered in this review. Bill length is usually related directly to body size (Ashmole 1968; Bédard 1969b), but note, for instance, that the longer species of the two kittiwakes has the shorter bill. Body weight would be a better measure of relative size than body size, but few reliable weight data are available for seabirds.

The use of different feeding methods by species in each community grossly assigns birds to feeding at different depths. Thus, whereas shearwaters, puffins, and small gulls (Xema sp., Rissa spp.) overlap almost entirely in prey categories and even prey species, the gulls can capture these organisms only at the surface; the shearwaters capture them at shallow depths; and the puffins capture them at much deeper depths. Direct field observations of this phenomenon are few but Gould (1971) and Sealy (1973a) compared the diets of birds feeding in mixed-species flocks. An example of how even finer divergence in feeding methods helps to partition food resources has been provided by Spring (1971) in his comparison of the two murres. Both species feed by diving to great depths, but the thick-billed murre is able to hover over the bottom and thereby is better able to capture benthic organisms.

The scavengers (generalists) offer a good example of how a range of bird and bill sizes is usually represented among species having similar diets and feeding methods. The progression of oceanic scavenger sizes is graded rather evenly from the black-footed albatross down to the northern fulmar, to the scaled petrel, to the storm-petrel. All these species capture prey that occur only at or near the water surface. Recently Sanger (1973) reported appreciable numbers of glaucous-winged gulls (Larus glaucescens) and herring gulls (L. argentatus), noted neritic scavengers, out in the oceanic realm of the petrel. He presented limited data that suggested an overlap between the diet of these gulls and that of black-footed albatrosses, as noted by Miller (1940). It would not be surprising if these gulls were as much generalists in the oceanic habitat as they are in the neritic. Interestingly, their bill and body sizes fall between those of the albatross and the fulmar, thus in theory enabling them to invade the oceanic habitat without great competition. It is likely that their invasion occurred during historical times and is related to their habit of following fishing boats from shore out to sea (Sanger 1973). If so, the gulls might be assuming from other species part of a previously uncontested resource.

Another interesting group of species that shows close similarities in diet consists of the piscivorous loons, grebes, and mergansers. All these birds, including seven or eight species, apparently feed on fish occurring on or near the bottom in the inshore neritic habitat. Again, however, an even progression in size exists: yellow-billed loon (Gavia adamsii), common loon (G. immer), arctic loon (G. arctica), red-throated loon (G. stellata), western grebe (Aechmophorus occidentalis), red-necked grebe (Podiceps grisegena), and common merganser (Mergus merganser). Most likely then, they select different-sized fish. Another example of this phenomenon is provided by the eight neritic gulls, which are largely scavengers and show a remarkably even progression in bill and body size. Finally, as shown clearly by Bédard (1969a, 1969b) and Harris (1970), alcids of different sizes select different-sized prey, often of the same species.

A final important way in which seabirds partition available resources is by inhabiting different zones. Zonation is especially evident during the breeding season when species common to the same breeding site sort themselves out according to the distances they range for food. This phenomenon was discussed by Murphy (1936), Shuntov (1974), Sealy (1972), Cody (1973), and Scott (1973).

The species that appear to have specialized food habits (if further research confirms that indeed they do) are probably very sensitive to vagaries in food availability or are, at least, much more sensitive than other species. Some specialists which also have very restricted distributions would, therefore, be susceptible to localized catastrophes occurring where specialists are concentrated around the food resource. This is proved in the case of the scoters, which are both specialized and rather restricted to nearshore beds of molluscs and have fallen victim to local oil slicks (Smail et al. 1972). An example of another potentially critical situation is that of the black brant, which at certain times of the year concentrate their entire population around eelgrass beds in Bristol Bay, Alaska, where much offshore oil drilling may soon occur.

Birds adapted to feed by diving, with the exception of cormorants, spend most of their time in the water. These species are therefore most susceptible to oiling (Smail et al. 1972), but pursuit plungers (the shearwaters) are also highly susceptible (Point Reyes Bird Observatory, unpublished data). A characteristic of polar and subpolar seabird communities is the high percentage of birds that feed by diving and pursuit plunging. These birds are mostly absent from tropical and subtropical communities because feeding by these methods is not adaptive there (Ainley 1977). Hence, oil pollution has all the potential of rendering maladaptive the principal feeding methods of many polar seabirds.

Another way in which seabird feeding relates to conservation problems concerns competition between birds and man for commercially valuable fishes. A related problem is the mass mortality of seabirds due to man's fishing gear. An acute situation is the drowning of seabirds caught in salmon gill nets (Bartonek et al. 1974; Pacific Seabird Group 1975; Ripley 1975; King et al., this volume). Immediate action is definitely required.

Further, competition between birds and man for the same resource has the potential for disastrous effects on bird populations if humans out-compete the birds and overfish the resource. A classic example, reviewed by Idyll (1973), is the possible collapse of the Peruvian anchovy (Engraulis ringens) fishery; if overfishing and an El Niño should coincide, the Peruvian seabird populations could collapse as well. The California fisheries and apparently the double-crested cormorants that nest on the Farallon Islands have both suffered from the demise of the Pacific sardine (Sardinops caerulea) in the California current (Ainley and Lewis 1974). In regulating fish harvests, fishery organizations should include in their calculations the harvest by creatures other than man (Schaefer 1970), rather than evading the issue by referring to a vague "natural mortality."

---

Finally, fishing by humans can benefit seabirds by removing fish (or whales) that compete with birds for food (Laws 1977). A potential example is that of northern California, where salmon and seabirds both feed heavily on juvenile rockfishes (Fitch and Lavenberg 1971; Point Reyes Bird Observatory, unpublished data). Harvest of salmon should theoretically leave more rockfish available for birds to eat. This sort of situation has not yet been fully documented and definitely warrants further study, especially in such areas as the Bering Sea, where some fish stocks have become depressed due to overfishing (Gulland 1970).

Many people realize intuitively that seabirds are important members of marine ecosystems. Although the supporting evidence is not now available, it will be needed if seabirds are to be protected. Emotion alone will not justify the protection of seabirds in an age when the human race moves steadily toward global famine. The job at hand is, in part, to sell seabirds, not just to the public, government officials, executives of oil companies, or fish-packing concerns, but also to marine biologists and oceanographers, for the scientists have the best means to study organisms at sea. We must move away from the concept that seabirds are merely yo-yos of various sizes, shapes, and colors on strings of various lengths that venture forth to sea from the land, grab a quick lunch, and then return to the safety of terra firma. Seabirds are marine organisms and deserve at least as much research attention as that currently given marine mammals.

The information now available on seabird diets is largely presented in terms of the number and volume of various prey species taken. Whereas these data provide the relative importance of prey, fishery data on prey stocks are usually measured in terms of biomass. Thus, it is difficult to relate seabird data to the immense wealth of information on biological oceanography. If we are to recognize the importance of seabirds in the nutrient and energy cycling of marine ecosystems, rather than considering them merely as "yo-yo predators," we must relate them to the total marine community.

The goal of marine ornithologists should be to refine and broaden considerably in detail such studies as those by Sanger (1972), Shuntov (1974), and Laws (1977), who attempted to assess the relations between seabird populations and stocks of other marine organisms for the northern North Pacific, the world oceans, and the Antarctic, respectively. The trophic roles played by seabirds must be studied in detail at the community level year-round before those analyses can be properly refined. Another exemplary work is that done by Brownell (1974), who studied trophic relations of higher vertebrates off Uruguay, including dolphins, pinnipeds, seabirds, and some large fish. In a review study, Sanger (1974) considered the food-chain relations of similar vertebrates in the Bering Sea. These sorts of studies will serve to bring the role of seabirds into perspective with other upper trophic level feeders.

We much appreciate the opportunity to participate in the symposium at which this paper was presented. The encouragement and help given by J. C. Bartonek was indispensable. D. G. Ainley's participation in the symposium was supported by the Point Reyes Bird Observatory. This is contribution No. 124 of the Point Reyes Bird Observatory.