Distribution

Archaeological records.--We did not conduct an exhaustive search for archeological records. Those we report were obtained while searching for early records and are based on skeletal remains retrieved from Native American middens. Though only four records were found, they indicate DCCOs were present in three of five breeding zones (Fig. 1) between 500-5000 years ago. Bones of young birds collected from middens on the Pacific and North Atlantic coasts indicate DCCOs were breeding on both coasts. Skeletal material retrieved from middens on Amchitka, Aleutian Islands, documents presence but does not verify breeding because no bones of young birds were found.

Early ornithological records (1500-1900).--Prior to 1900, cormorants were widely distributed across North America, and occurred in all five of the hypothesized breeding zones (Fig. 1). Approximately 80% of the locations shown are documented breeding sites (i.e., have records of nests, eggs, or young); we designated the remaining 20% as probable or possible breeding locations based on circumstantial or anecdotal information (e.g., time of year observed, author’s comment). Systematic surveys for birds were not formalized at most locations prior to the twentieth century; thus, the number of records within each breeding zone does not reflect abundance of breeding birds or extent of distribution. In some areas (e.g., Pacific and Atlantic coasts), multiple records prior to 1900 are available; in others (e.g. some parts of the Interior and Gulf Coast), little effort was made to survey avifauna prior to the twentieth century (Ridgway 1874).

Current records (1970-2000).--In general, distribution of the current population is similar to that of the historic population (Fig. 1). All records shown are documented nest sites, and, with the exception of states and provinces for which data were not available, represent the continental breeding distribution. Important areas that appear different from historic distribution (Fig. 1) include the Great Lakes and western portion of the Interior range; interior portion of southeastern U.S.; and southeastern Alaska. Comparison of historic and current distribution maps suggests the species currently has a wider distribution in these zones.

Abundance

Historic abundance.--While precise counts are not available for most colonies prior to the twentieth century, records located for each population zone suggest historic populations of DCCO were very large. To indicate relative abundance within each population zone we summarized colony observations and qualitative descriptions for each breeding region (see Zone Summaries below).

Current abundance.--Census efforts among states and provinces are rarely coordinated across breeding zones due to several factors (e.g., limited funding, political boundaries). Therefore, colony size typically is not estimated during the same year or using the same census technique across a population. Additionally, some important areas that provide breeding habitat for large numbers of cormorants (e.g., Mexico, Manitoba) are not regularly censused and little is known about cormorants in major portions of these areas. Thus, accurate estimates for regional (zone-wide) populations are not available. However, most areas within the five breeding zones are regularly censused and estimates of breeding pairs are available for many colony sites. To indicate relative abundance of DCCOs in each population zone we report (Table 2) population estimates for each breeding region based on the most recent and complete census efforts.

Alaska Summary

Birds in this zone comprise the subspecies P.a. cincinatus, the most restricted (Palmer 1962) and smallest of the five populations. Between 1970 and 2000, DCCOs were confirmed nesting at 126 colonies, mostly along the southern coast and on the Aleutian Islands (Fig. 1). Alaska has not conducted statewide censuses of all colonies in the same year and the number of breeding birds is not known. However, 106 sites were censused at least once between 1970 and 2000; most colonies (93%) were small (<100 pairs), and a total of 3,029 pairs was reported (Table 2). Because so few data are available over such a broad time period it is not possible to comment on how closely this estimate approximates actual numbers, but the abundance of small colonies suggests the population is small.

Determining historic distribution and abundance in this zone is complicated by the large-scale introduction of foxes (Alopex lagopus; Vulpes vulpes) to Alaska that occurred between 1750 and the early 1900s (Bailey 1993). Because introduced foxes had major impacts on nesting seabirds (Bailey 1993; Carter 1995), records collected during the historic period may reflect a recently modified distribution and abundance. Turner (1885) described DCCOs as abundant residents and breeders in the Near Islands in the western Aleutians in the late 1800s, and Clark (1911) reported they were still present on these islands as breeders in the early 1900s. However, the peak in fur farming did not occur until the early 1900s, with the most rapid growth occurring in the 1920s (Bailey 1993); by the mid-1930s, DCCOs were no longer breeding in the Near Islands (Murie 1959). Today the western-most breeding point is Chuginidak, Aleutian Islands (52°51’04”, 169°49’69”).

Gabrielson and Lincoln (1959) note that reports of DCCOs breeding throughout southeastern Alaska could not be confirmed with available information. They reviewed early records up to 1945, and defined the breeding distribution “from Kodiak westward at suitable places along the Alaska Peninsula far out into the Aleutian Islands.” All observed colonies were small (<100 pairs).

Based on available information, current DCCO distribution appears more restricted than in the past. In this portion of the range, DCCOs utilize level ground for nesting, which makes colonies vulnerable to terrestrial predators (Siegel-Causey et al. 1991). Introduction of foxes may have limited cormorant distribution, and if large colonies existed, they may have been substantially reduced before they were documented, particularly in the Aleutian Islands and Gulf of Alaska (Siegel-Causey et al. 1991; Carter et al. 1995). Murie (1959) noted a “drastic change” in the Alaska distribution since about 1906. Disappearance from portions of the range combined with increased predation pressure suggests DCCOs may have been more abundant historically, at least in parts of this region, than they are today. Additional causes that may have contributed to declines include other introduced predators and rabbits, and human disturbance (Siegel-Causey et al. 1991; Bailey 1993; Carter et al. 1995).

Pacific Coast Summary

Birds in this zone comprise the subspecies P.a. albociliatus. Between 1970 and 2000, nesting was documented at 248 sites (Table 2). The current breeding population occurs mostly along the coast from southern British Columbia to at least Bird Island, Sinaloa, Mexico (Fig. 2), and possibly further south (Carter et al. 1995). Significant colonies also occur inland. Summing available estimates for the late 1980s and 1990s, approximately 33,000 nesting pairs were documented in the region primarily during the mid-to-late 1990s. However, estimates from important portions of this region are dated, and large-scale movement among breeding birds occurs. In the Salton Sea, dramatic increases in breeding birds occurred during the mid-to-late 1990s, and cannot be attributed to productivity rates alone (D. Shuford, Point Reyes Bird Observatory (PRBO), pers. comm.); most birds were believed to be immigrants from other locations, possibly Mexico (D. Shuford, PRBO, pers. comm.). Because of large shifts and intercolony movement, frequent monitoring efforts over wide areas are required to estimate regional population size (Carter et al. 1995). Although data collection and census efforts have not been coordinated across the region, available information indicates this population is the third largest, but relatively small in comparison to the North Atlantic Coast and Interior populations.

With the exception of British Columbia, the DCCO was recorded as a breeding species throughout most of the Pacific Coast prior to 1900. The earliest breeding records we located were bones of young cormorants retrieved from a shell mound in Emeryville, CA. Samples from the mound indicate that aboriginal people utilized the area as long as 2,550 years ago and hunted a great diversity of waterbirds, including DCCOs (Howard 1929; Sher 1994; Broughton 2004). Although bones of DCCOs are abundant at archaeological sites throughout the Straits of Georgia, British Columbia, indicating the species occupied this area for the past 5,000 years, no skeletal remains of young birds were identified in excavations (Hobson and Driver 1989). Nesting was not documented in the province until 1927 (Munro 1928). However, based on archaeological evidence, Hobson and Driver (1989) suggest that the first documented nesting for this species may actually represent a re-colonization of the area.

Multiple records indicate breeding DCCOs were historically abundant across much of their Pacific Coast range (Squires 1917; Ray 1915; Chamberlin 1895; Willett 1910, 1933; Howell 1917; Grinnell 1908; Finley 1907, 1915; Linton 1907; Lamb and Howell 1913; Goldman 1908; Salvadori 1865; Lamb 1927; Bryant 1889; Bancroft 1927b; Bendire 1877; Dawson 1908, 1911). The largest cormorant colony ever recorded on the continent existed at San Martin Island, Baja California, MX; Wright (1913) estimated 348,840 nests at this site. The estimated size of this colony alone is larger than each of the five current regional populations and may be comparable to current continental population size. Although Wright’s estimate was later considered an overestimate and revised by J. R. Jehl to 213,500 pairs (Hatch 1995) based on the area of the island, the revised estimate still dwarfs the present size of the entire Pacific Coast and much of the North American population. Even if Jehl’s estimate is off by an order of magnitude, no colony approaches this size anywhere in North America.

By the late 1800s and early 1900s, the DCCO had experienced substantial decline and loss of breeding colonies along several portions of its Pacific Coast range. The species was heavily persecuted by humans, and breeding birds were shot at colonies and nests destroyed. In addition, habitat was lost due to agricultural and water developments (Carter et al. 1995). Review of early records suggests that the DCCO was far more abundant in this zone historically than it is today.

Interior Summary

Birds in this zone represent a substantial portion of P.a. auritus’ distribution. Between 1970 and 2000, nesting was reported at 704-754 sites (Table 2). This is the largest, most widespread breeding population; it spans the prairie provinces of Canada, the Canadian and U.S. Great Lakes, and southwestern Quebec. The distribution extends west of Minnesota to southwestern Idaho, and as far south as central Utah and central Colorado. Local breeding also occurs in central Kansas and possibly northern New Mexico. Breeding birds in Idaho were included in this zone based on Burleigh (1972), who reported that 11 specimens collected from different locations in Idaho were intermediate in their characters between P.a. auritus and P.a. albociliatus, but closer to auritus. Birds breeding in the Upper Rio Grande drainage and Middle Pecos River, New Mexico, were also included in this zone, based on plumage characters of P.a. auritus (S. Williams, NM Dept. Game and Fish, pers. comm.). Genetic analyses need to be undertaken to confirm these subspecies designations.

Recent estimates based on complete counts are available for about half the states and provinces where cormorants breed. With the exception of Manitoba, complete counts are available for most of the areas where cormorants nest in large numbers. The largest numbers occur in the Great Lakes and Prairie Provinces. Summing the most recent estimates available for each state and province gives a very rough estimate of 170,000 pairs for the region in the 1990s (Table 2). However, the breeding population is significantly larger than this. Recent estimates are not available for some areas, and several estimates for other areas were based on partial counts. With the exception of Manitoba, DCCOs are not known to breed in very large numbers (several thousands) in most portions of the Interior range where estimates were missing or incomplete (Hatch 1995, Wires et al. 2001). In Manitoba, however, DCCOs do breed in large numbers, particularly on Lake Winnipegosis. The last province-wide survey was conducted in 1979, at which time 22,642 active nests were documented; 9,053 nests were on Lake Winnipegosis, an extremely important area for waterbird nesting (Koonz and Rakowski 1985). The last census of Lake Winnipegosis in 1999 indicated pairs of nesting cormorants had quadrupled since the previous survey 20 years earlier (Table 2). Assuming cormorant numbers across Manitoba increased at the same rate, there may be as many as 90,000 pairs in Manitoba province-wide (W. Koonz, Manitoba Dept. Nat. Res., pers. comm.). Given these gaps, the Interior population in the late 1990s may have been closer to 270,000 pairs but the accuracy of this estimate is unknown.

In many parts of the Interior region, the early breeding history of the DCCO is well known. Pre-1900 records document nesting by the species across most of the region and suggest it had been a long-time and abundant breeder in several areas, particularly in the Prairie Provinces and the mid-western states, where many large colonies were documented. In Minnesota, Hatch (1892) reported DCCOs bred in nearly all parts of the state, and was “occasional to innumerable” depending on how close one was to breeding colonies. At the Minnesota-Iowa border, observers reported “the air is jist black with em’ an they’re nestin’ on the island so yer can’t see it for eggs” (Whitehead 1887). In Missouri, Widmann (1907) reported cormorants breeding in “considerable numbers.” In Ohio, “boatloads” were killed at St. Mary’s Reservoir (Langdon 1878). At Lake of the Woods, Ontario, “great numbers” of young were killed in the late 1700s (Tanner 1994). Lakes Winnipeg and Winnipegosis, Manitoba, have long been recognized for their very large colonies (Seton 1886; Bent 1922). Lake Isle a la Crosse, Saskatchewan, also had a “large” regular breeding colony (Seton 1908). For additional records see Barnes 1890; Cooke 1888; Roberts 1932; Agersborg 1885; Stansbury 1853.

The Great Lakes is the one area within the region where status (e.g., breeder or migrant) and early history of the cormorant are not clear. For example, investigators reported that DCCOs colonized and invaded the Great Lakes in the early 20^th century, moving in an easterly direction after the first nesting was reported on western Lake Superior in 1913 (Postupalsky 1978; Weseloh and Collier 1995). However, records and other evidence indicate DCCOs were present in the area prior to the 20^th century, suggesting this portion of the cormorant’s North American range was occupied earlier than 1913, and possibly breeding. These records and evidence, described below, fall into the following categories: anecdotal, geographic place names, and general breeding distribution.

In the first category, three anecdotal records are important in presumed colonization history. The earliest reported observation of cormorants breeding at a specific location on the Great Lakes proper is an anecdotal record; Postupalsky (1978) cites a personal communication from Frank Novy that cormorants were breeding in the Great Lakes on western Lake Superior in 1913. No additional information is available about this record. However, because no other records were known at the time of Postupalsky’s publication, this 1913 personal communication is regarded as the first nesting of cormorants in the Great Lakes. Additionally, the order in which the Great Lakes were hypothetically colonized is traced to it (Postupalsky 1978; Weseloh and Collier 1995). The first documented account of cormorants nesting on the Great Lakes proper is reported in Fargo and Van Tyne (1927). This record reports discovery of a small colony (10 nests) in 1926 at Agawa Rocks, Agawa Bay, on the eastern shore of Lake Superior. However, Fargo and Van Tyne noted that local residents told them “the cormorants had nested there for years.” No additional information is available about this colony, but this anecdotal information suggests cormorants had a history of breeding in the area prior to discovery of the colony. A third record (Baillie 1947) from Georgian Bay, Lake Huron, reports that cormorants were nesting in the general vicinity of the Mink Islands in 1919; though this record was never verified, based on other nesting records, Baillie (1947) believed the record was probably accurate. This record places breeding cormorants much farther east into the Great Lakes more than a decade earlier than was hypothesized by Postupalsky (1978) and Weseloh and Collier (1995).

In the second category, geographic place names suggest cormorants were present in the area and recognized by both Native Americans and settlers, but their status as breeders or migrants is not reported. The earliest of these place names is found on a General Surveyors Office Plat Map of Forsyth Township, Michigan. On 4 Aug, 1854, this map was signed and filed in the General Land Office in Detroit, and documents the existence of two lakes, Shag Lake and Little Shag Lake. Shag, a common vernacular name for cormorant, suggests cormorants were present in the area when the map was prepared. These lakes still exist and are located in the upper peninsula in Marquette Co, about 8-16 km from Lake Superior and within 64 km of Lake Michigan, Bay de Noc area, where cormorants are currently abundant. The map survey work was done during the third quarter (July-Sept) of 1844, and further subdivisions of the township were completed during the third and fourth (Oct-Dec) quarters of 1852. Thus cormorants occurred in the area during the summer and or fall months but their status (breeder, migrant) is unknown. The second geographic place name is from Lake Onaping, Ontario, about 64 km north of the North Channel, Georgian Bay area (Bell 1891). At this location a rock forming part of the cliffs on the west side of the lake was named by the Ojibwa “Kakakeshiwishtagwaning”, which means “the cormorant’s head”, and suggests that the Ojibwa were familiar with cormorants in this landscape.

In the third category, the first four editions of the Checklist of North American Birds (1886, 1895, 1910 and 1931) all report the Great Lakes as part of the breeding range of the DCCO but give no specific locations within the lakes. These records indicate cormorants were breeding in the Great Lakes region earlier than 1913, but because specific locales are not provided it is not possible to determine if cormorants nested on the Great Lakes proper, or on nearby inland lakes. An additional record reports capture of a cormorant in June, 1877, at Sandusky Bay, OH (Wheaton 1882). No other information is available about this record, but the presence of a cormorant at this time of year suggests the bird may have been breeding.

As occurred over much of the continental range, DCCOs in the interior underwent substantial declines in the late 1800s and early 1900s. Persistent human persecution and habitat degradation resulted in colony declines, abandonment and local extirpation from several states (e.g., South Dakota, Indiana, Ohio, Missouri, Iowa, Illinois, Arkansas) (Lewis 1929). Because so few estimates are available pre-1900 and data vary greatly among states and provinces, it is not possible to determine if cormorants were more abundant historically in this zone than they are today. For some areas (e.g., Indiana, Ohio, Iowa, Arkansas, Illinois, Minnesota, Missouri) available data suggest breeding cormorants were more abundant historically. Additionally, very large migrant flocks were observed that were much larger than those currently known to pass through the region. In Minnesota, a flock of migrating cormorants was described that was “four miles long and one-and-a-half miles wide” (Sennett 1891). Though numbers declined throughout much of the Interior by the turn of the century, large numbers were still occasionally observed; in 1926, a flock estimated at 100,000 to one million individuals was described migrating up the Mississippi River past LaCrosse, and was so large “that at times it was impossible to see the sunset sky through the mass” (Grassett 1926). During recent years (1970-2000), no flocks anywhere near this size have been reported.

In other areas of this zone (e.g., the Great Lakes), current numbers are much greater than they were in the late 1800s-early 1900s, and are the highest recorded in the history of the region. In the Prairie Provinces, cormorants were historically and are currently abundant, but it is not possible to determine if abundance has greatly changed.

Southeast U.S. and Caribbean

Birds breeding along the Gulf Coast, the south Atlantic Coast, and in the Caribbean comprise the subspecies P.a. floridanus, while birds breeding in the Bahamas and Cuba are classified as P.a. heuretus (Hatch and Weseloh 1999). This breeding zone extends from southern and central Texas to Florida, and along the Atlantic Coast through Florida, Georgia, and the Carolinas. Birds breeding in southern Texas and southwestern Arkansas were included in this zone based on subspecies distribution maps (Palmer 1962; Johnsgard 1993). Because North Carolina has been reported as the northern breeding limit for P.a. floridanus (Audubon 1840-1844; Bent 1922; Johnsgard 1993), and other authors have suggested that Carolina birds may comprise this subspecies (Palmer 1962; Clapp and Buckley 1984; Post and Post 1988), we included birds breeding in the Carolinas as part of the Southeastern U.S./Caribbean zone. Recent estimates based on complete counts are available for most of the known colonies, with the exception of southern Texas, the Yucatan and the Caribbean. The largest known numbers occur in Florida. Summing estimates available for the region in the 1990s gives an approximate estimate of 9,400 pairs at >108 sites (Table 2).

Prior to the 20^th century, the DCCO was “constantly resident in the Floridas and their Keys, and along the coast to Texas” (Audubon 1840-1844). Breeding was recorded in North Carolina as early as the 18^th century; at that time cormorants were reported to “lay their eggs in . . . the Islands, in the Sound and near the Sea Shoar in the Banks, and sometimes on high trees, as the Shags do” (Brickell 1737). Several records document breeding along both the Gulf and Atlantic coasts (Table 1).

Although no quantitative data were located to provide a comparison between current and historic abundance, descriptions suggest breeding cormorants were more abundant in this zone prior to the 20^th century than they are today. Presently, breeding cormorants are only abundant in Florida (Wires et al. 2001), while historically they were very numerous in multiple areas of the region. Audubon (1843) reported “many thousands” breeding in the Florida Keys in 1832; Brickell (1737) stated that in North Carolina, they were “as numerous all over these Parts of America as in any part of the World”; Audubon (1843) and Beyer et al. (1907) described cormorants as “abundant” in the Louisiana interior and on the coast; and Howell (1911) reported this species as “formerly abundant in the rivers and swamps of eastern Arkansas”. Additionally, based on Audubon’s (1840-1844) description, DCCOs may have been present as breeders along the Alabama, Mississippi and Texas coasts.

Northeast Atlantic Coast

Birds in this zone are also placed in the subspecies P.a. auritus, and breed along the Atlantic Coast from southern Newfoundland, the northern shore of the Gulf and estuary of the St. Lawrence River, Anticosti Island, Magdalen Island, south along the coast to New York City and Long Island (Hatch and Weseloh 1999). The range is expanding south with recent breeding and re-colonization in New Jersey, Delaware, Maryland and Virginia, as far south as Hopewell. Recent estimates based on complete counts are available for nearly all colonies known to be active in this region, with the exception of New Brunswick (Table 2). The largest numbers occur in Quebec in the St. Lawrence River, Estuary and Gulf, along the coast of Maine, and in the Maritime Provinces. Summing estimates available for the region in the 1990s gives an approximate estimate of 87,000 breeding pairs at 381-382 sites. This number should be interpreted cautiously because intercolony movement is common and the size of individual colonies may change dramatically from year to year in substantial portions of the region (J. F. Rail, Canadian Wildlife Service (CWS), pers. comm.). Additionally, a culling program in the St. Lawrence River in 1989 significantly reduced numbers in Quebec in the 1990s.

Archaeological records document cormorants as a breeding species in this zone from the late prehistoric period (about 1500 AD) (Luedtke 1980). Several historic records report breeding from the Straits of Belle Isle, northwest coast of the Island of Newfoundland, south to as far as Boston Harbor, Massachusettes (Table 1). The historic breeding distribution may have extended further south; Mendall (1936) notes that Williams’ (1643) observations of Native American use of cormorants for food referred either to Massachusetts or Rhode Island or both. Mackay (1894) noted that in the late 1800s, hundreds were observed every spring around Newport County, Rhode Island, at a rock outcrop near Sachuest Point called Cormorant Rocks, which was named on a map as early as 1776. Cormorants were also present historically in Virginia but seasonal presence and distribution is not clear (Lewis 1929).

As in the other breeding zones, sometime during European settlement DCCOs began declining in most parts of the Atlantic range due to heavy human persecution (Lewis 1929; Mendall 1936; Gross 1944; Krohn et al. 1995). By the early 19^th century, they were extirpated in New England. By the late 19^th and early 20^th centuries, breeding cormorants had disappeared from Nova Scotia and were greatly reduced in the other Maritime Provinces and Quebec (Lewis 1929; Erskine 1992). Although we found no early estimates or archaeological evidence to which historic and current abundance can quantitatively be compared, early comments and records suggest cormorants were very abundant in this breeding zone prior to European settlement (Williams 1643; Audubon 1843; Reeks 1869; Champlain 1922; Goss 1889; Townsend 1917). Mendall (1936) cites Morton (1637), who wrote of cormorants in New England, “There are greate store of Pilchers: at Michelmas, in many places, I have seene the Cormorants in length 3.0 Miles feedinge upon the Sent.” Lewis (1929) reported that several 17^th century references mention the abundance of cormorants in Virginia, and that in 1610, one anonymous author reported cormorants among the birds in Virginia rivers “in such abundance as are not in all the world to be equaled.”

The current breeding range of the DCCO is similar to pre-1900, although some changes have occurred. The range has retracted in Alaska and in southeastern U.S, but expanded along the Atlantic Coast and in the Interior. In the latter regions, some important changes may represent re-colonization events. DCCO historic abundance and widespread distribution has several implications for current conservation and management efforts, discussed below.

The Great Lakes Population

The Great Lakes region is one of the most important areas in the human-cormorant conflict, and concerns from this region have strongly influenced federal policy decisions for cormorants. Most of the conflict centers on perceived impacts to fisheries from cormorant predation. Currently many factors affect the health of the Great Lakes ecosystem; one of the most significant is establishment of >160 exotic invasive species (Mills et al. 1993; Ricciardi 2001). Of these, several (e.g., zebra mussel (Dreissena polymorpha) alewife (Alosa pseudoharengus), sea lamprey (Petromyzon marinus) directly and indirectly impact valuable fisheries (e.g., Mills et al. 1993; Belyea et al. 1999). Additionally, many other variables (e.g., water temperature variation, declines in benthos abundance) also impact fisheries (Hoyle et al. 1998, Pothoven et al. 2001). Nevertheless, in areas where declines in sport fish species and catch have occurred, many anglers and fisheries biologists attribute declines to recent increases in DCCO numbers (e.g., Fielder 2004). Additionally, some anglers believe cormorants are non-native to the region, and report that the species was intentionally introduced by the Japanese (F. Cuthbert, pers. obs.). The media has also published articles stating the species is an exotic (Sharp 2004). Finally, large cormorant colonies in this region have once again become the focus of vandalism and destruction as well as legal control activities. For example, in the late 1990s-2000, about 1,350 DCCO adults and chicks were illegally shot at Little Galloo Island, Lake Ontario, New York, and at Little Charity Island, Saginaw Bay, Michigan, by anglers who believed cormorants were responsible for poor catches (Kloor 1999; Lounsbury 2000). In summer 2004, all colonies in the Les Cheneaux Islands area, Michigan, were targeted by USDA/WS for control (e.g., harassment, oiling eggs, shooting adults) to aid the ailing yellow perch fishery (USDA/APHIS/WS 2004), despite an earlier rigorous study of cormorant predation in this area that concluded cormorants were not a substantial factor in yellow perch mortality (Belyea et al. 1999).

One critical component that shapes perception of a species’ role within an ecosystem is knowledge of species history. Previously, occurrence and movement of cormorants through the Great Lakes in the early 1900s was described as an “invasion” (Postupalsky 1978; Weseloh and Collier 1995). Perception of this species as an invader will strongly influence its management because invasive species, variously described as “exotics”, “nonindigenous” or “aliens,” typically spread and cause net harm to a system (Lodge and Shrader-Frechette 2003). Review of all available records to assess the history of the Double-crested Cormorant in the Great Lakes suggests characterization as an “exotic” or “invasive” is inappropriate, and history in the Great Lakes needs further exploration. Considering records documenting nesting prior to 1900 are available from the western and southern border states of the Great Lakes, and abundant nesting habitat would have been available, apparent species absence is puzzling.

Historic Populations and Current “Overabundance”

Most historic records cited in this work were obtained during the period of European settlement, particularly in the 19^th century (Table 1). We assume abundances that existed during this period represent populations under relatively natural conditions (i.e., abundance attained when relatively free from artificial/human modified elements that could lead to inflated or diminished numbers). However, it is possible that in some locations settlers were encountering spectacularly abundant concentrations of cormorants and other wildlife that were responding to release from Native American harvest. Broughton (2004) suggests that the abundant wildlife populations observed in pre-1900 California resulted after Native Americans experienced dramatic disease-based population declines through coastal contact with European explorers in the 16^th century. Thus, for some species, assuming that the pre-1900 time period represents the baseline or benchmark of past populations, may be a questionable assumption (Broughton 2004). Nevertheless, although the impact of Native American harvest on cormorants at a large scale (e.g., state, regional, continental) is not known, the breadth of cormorant records spanning hundreds of years and locations across the continent (Table 1) suggests the Double-crested Cormorant had been an abundant species for a substantial time. Additionally, the clearly documented record of human persecution throughout the European settlement period marks a significant era in the known history of the cormorant, to which massive population declines and local extirpations can be traced.

On a continental level, comparisons of historic and current records indicate cormorants were likely more abundant during initial European settlement than they are today. Therefore, describing the current continental population at an “all time” or “historic” high is misleading. Additionally, in several areas across the continent, portions of the public and some natural resource biologists view the DCCO as a recent addition to the regional avifauna despite historic records that demonstrate new nest records actually represent re-colonization of former breeding areas (Wires et al. 2001). Lack of knowledge about historical distribution and abundance has led to significant misconceptions about numbers of cormorants the environment may support. For example, some portions of the public in conflict with this species (e.g., anglers, aquaculturalists) consider mid-century (pre-1980) numbers to be the “norm” for DCCOs; however, numbers and distribution during this period were greatly diminished due to contaminants and persecution (e.g., Gress et al. 1973; Vermeer and Rankin 1984; Weseloh et al. 1995). Similarly, in the Great Lakes, where cormorants are currently at recorded highs, growth during much of the 20th century was suppressed through legal and illegal control activities and contaminants (Ludwig and Summer 1995; Weseloh et al. 1995). Thus population size prior to the 1970s (when national legislation protecting cormorants and reducing contaminants was introduced) was smaller due to human intervention and environmental degradation, rather than to natural regulation and biological carrying capacity. Misconceptions about historic numbers and what is “normal” in current ecosystems, rather than science, has fueled much of the debate about need to control DCCOs in North America.

The term “overabundant” is a judgment that needs qualification. Kushlan et al. (2002) imply one way to judge overabundance is to determine if a species has exceeded historical norms. This criterion is useful because comparison of current to historic numbers may help determine biologically if there are “too many” individuals of a species. Based on available records, the continental DCCO population has not exceeded historical norms. Additionally, no study has demonstrated DCCOs exceed biological carrying capacity by surpassing food supplies or habitat. Although concern about potential impacts to various resources has been reported in several areas (USDI/FWS 2003b), negative changes in natural resources directly linked to DCCO population growth via rigorous scientific study have been documented at only a few locations (Wires et al. 2001; Cuthbert et al. 2002).

In biological terms, we suggest characterization of the continental DCCO population as overabundant is inappropriate. This characterization is applicable, at best, only on a local and site-specific level where numbers can be demonstrated to outstrip resources. Applied to the eastern U.S. population, use of the term “overabundant” is meaningful only in terms of “wildlife acceptance capacity” (Decker and Purdy 1988). This term was created to highlight that many wildlife or social acceptance capacity levels may exist for a particular wildlife population at any point in time; this is unlike the concept of biological carrying capacity. Usually, wildlife acceptance capacity is far smaller and more variable than that determined by biological carrying capacity.

Conservation vs. Management Goals

Over the last decade and a half, large-scale conservation plans for birds in the Americas have developed that depart from the traditional narrow focus on threatened and endangered species, and now encompass broader and more representative goals, such as that defined by the vision of Partner’s in Flight’s (PIF) “keep common birds common” (Fitzpatrick 2002). Waterbird conservation plans at national and continental scales directly link species and population goals to historic distribution and abundance (Milko et al. 2003; Kushlan et al. 2002). Fitzpatrick (2002) takes current bird conservation initiatives one step further and offers the following as a robust mission statement: “Ensure persistence of all American bird populations in their natural numbers, natural habitats, and natural geographic ranges . . .” In these more recent initiatives and perspectives, historic distribution and abundance provide the basis for conservation focus. Thus, when designing management and conservation plans, and developing biologically meaningful population objectives, it is important to recognize where and in what numbers and habitats DCCOs formerly and naturally occurred. Because very large numbers have been typical for this species historically, population targets based on fishery objectives or other human imposed values may be readily surpassed and likely will require intensive management.

In some areas, return of DCCOs after a long period of absence or local/regional rarity has created conflicts over resources because humans are not accustomed to dense concentrations of DCCOs and now more intensively use areas that were formerly occupied by cormorants (Wires et al. 2001; Wires and Cuthbert 2003). These conflicts have led to policy decisions and management plans (Glahn et al. 2000; USDI/FWS 2003c; USDA/APHIS-WS 2004) that are incompatible with current conservation goals (Wires and Cuthbert 2001; Fitzpatrick 2002; Kushlan et al. 2002; Milko et al. 2003). For example, the management plan prepared for DCCOs to minimize damage to southern aquaculture (Glahn et al. 2000) proposes that southern breeding colonies be managed at levels compatible with aquaculture to forestall future depredation problems. Cormorants have recolonized portions of Arkansas and Mississippi after several decades of absence from these states; current colonies are small and few in number (Table 2). Despite no record of cormorant breeding in Arkansas since 1951, most (85%) of the birds at the only known nesting site were collected for research by USDA/APHIS-WS when they resumed breeding in 1999 (Hutchinson 1999). The Arkansas Wildlife Services state director commented that he hoped “we can keep this population down and maybe eliminate it . . . if we let it get out of control we could have cormorants throughout the year” (The Catfish Journal 1999; Hutchinson 1999). Though year-round residence in Arkansas by DCCOs may result in more conflicts with the aquaculture industry, year-round residence was documented historically (Howell 1911; James and Neal 1986).

In spring 2004, several islands in the Great Lakes were selected for large-scale cormorant control but these islands have also been identified as providing significant habitat for birds. For example, the islands in northern Lake Huron, MI, which include the Les Cheneaux Islands, have been proposed as an Important Bird Area (J. P. Cecil, Audubon, pers. comm.), and several of these islands with nesting cormorants were previously prioritized for conservation because of their regional importance to breeding colonial waterbirds (Wires and Cuthbert 2001). At Presqu’ile Provincial Park in Ontario, a formally designated Important Bird Area, over 6,000 breeding cormorants were killed in 2004, allegedly to protect an old-growth forest (Ontario Parks 2005).

Conservation of DCCOs and other abundant waterbird species cannot be effective without a unified conservation philosophy based on knowledge of historical populations and policy decisions founded on sound science. Public dissatisfaction with waterbirds is one of the greatest threats affecting long-term survival of waterbird populations (Kushlan et al. 2002). Much of this negative attitude is held because waterbirds eat fish and are perceived as competitors. In the case of the DCCO, its image is especially negative; the human-cormorant conflict is exacerbated because the species was not present or existed in very low numbers for all or most of current human memory in North America.

The fact that very large numbers of DCCOs existed prior to European settlement has fundamental implications for multiple natural resource management issues. Historically, vigorous fish populations sustained large numbers of fish-eating birds; in relatively natural systems, this is still possible (e.g., Davoren and Montevecchi 2003). In highly modified systems (e.g., Great Lakes), controlling native predators does not address the diverse problems that affect this ecosystem and its ailing fisheries. However, in many areas of the Great Lakes, DCCOs have been identified as the major factor limiting fish populations. This, in turn, has led to the development of population objectives for cormorants based largely on fishery objectives. For example, agencies managing cormorants in the U.S. eastern basin of Lake Ontario and on Long Island (Oneida Lake) (New York Department EC), and Canadian colonies on lakes Huron and Ontario (Ontario Ministry of Natural Resources) have established cormorant population sizes hypothesized to increase local fisheries or to achieve local fishery yields (Farquhar et al. 2001, 2004; Council of Lake Committees 2003; NYSDEC Newsletter 2004).

Although managing cormorants may benefit some fisheries and resolve human-cormorant conflicts, setting population objectives for cormorants based entirely on fishery or other objectives derived from human values significantly departs from the concept of conserving birds in natural numbers and natural habitats (Fitzpatrick 2002). Given the global state of ailing fisheries and increasing government sensitivity to wildlife acceptance capacity, self-regulating populations of cormorants may only be possible in cormorant “safe zones” where human interests are not allowed to influence cormorant numbers. Because few such “safe zones” are likely to exist, those committed to the conservation of fish-eating birds should vigorously oppose strategies that set population objectives based entirely on human interests rather than species biology and regional ecology. Finally, we urge the avian conservation community to support broad conservation strategies based on ecosystem health and process that recognize humans, fish and cormorants as three components of a complex system driven by many species and dynamic interactions.