Main Foods Taken
At most sites almost entirely fish, generally slow-moving or schooling species; size range 3–40 cm, but commonly <15 cm. Less frequently other aquatic animals, including insects, crustaceans, eels, and amphibians (Palmer 1962a), with single reports of voles (Microtus pennsylvanicus; C. Adams unpubl.) and a snake (Bent 1922). Foraging is opportunistic and flexible. More than 250 species of fish from >60 families reported as prey.
Microhabitat For Foraging
Generally feeds in shallow, open water (< 10 m deep) and close to shore (< 5 km away) (Stapanian et al. 2002, Colman et al. 2005). On Lake Erie cormorant flocks were found more frequently than expected on water < 10 m deep (Stapanian et al. 2002). Cormorants on Oneida Lake showed a clear preference for depths of < 7.5 m (Coleman et al. 2005). May feed over sandy bottoms or among rocks and in beds of sea grass or kelp. On Oneida Lake, NY, cormorants were found foraging in regions with cobble and rubble lake bottom disproportionately (22%) to the area available (12%; see Colman et al. 2005). Often feeds at the bottom, but also in midwater. On the Pacific Coast, more likely than other cormorants to forage in open water and in estuaries (Ainley et al. 1981a). Responds rapidly to concentrations of prey (see Conservation and Management: conflicts, below). Although captive Great Cormorants feed at night in Japan, Double-crested Cormorants are almost exclusively diurnal.
Food Capture And Consumption
Dives from surface and pursues prey underwater. Grasping of fish is aided by hooklike nail at tip of maxilla (upper half of bill) and by muscles attached to occipital style (xiphoid). This pointed bone, articulating at posterior part of skull, is present only in cormorants and anhingas (Siegel-Causey 1988b). Wide jaw-opening is facilitated by nasal-frontal hinge at junction with cranium. Small prey may be swallowed underwater; those noticed at the surface are likely to be large or otherwise difficult to handle, such as eels, flounders, or spiny fish. Handling of larger prey includes shaking and hammering on the water. Legs may be shaken from crustaceans. Then prey may be thrown in the air, to be caught and swallowed headfirst. To catch skulking prey such as mudpuppy (Necturus) and burbot (Lota lota) in freshwater, or gunnels (Pholis) in the ocean, cormorants probably root around at the bottom.
When feeding on schooling prey, may form loosely coordinated foraging flocks, sometimes in lines or shallow crescents. The line rolls forward as individuals at rear fly short distances to “leapfrog” the diving birds (Bartholomew 1942). In Georgian Bay, Ontario, in late summer and early fall, coordinated flocks of 50–175 birds drive fish to heads of narrow bays (Glanville 1992). In Massachusetts, foraging in flocks is most noticeable by transient fall migrants (JJH). In Green Bay, WI, flock foraging during breeding season was associated with clear water, shifting to solitary foraging where water was muddied by rainfall (Custer and Bunck 1992); associated with turbid water according to van Eerden and Voslamber (Van Eerden and Voslamber 1995). Feeding territoriality not reported.
Flying cormorants sometimes detect the presence of fish below and may rapidly descend to feed; aborted responses of that kind may explain mid-flight “hesitations” shown by flocks. Swimming cormorants occasionally hold their heads underwater before diving, apparently looking for prey. Several features of the eye indicate effective vision underwater as well as in the air; cornea is notably thick and flattened, and intraocular muscles are well developed, apparently to facilitate underwater accommodation (Sivak et al. 1977). However, Double-crested Cormorants commonly catch bottom-dwelling fish in turbid water where visibility is very low, so tactile sensitivity may be important for both locating and capturing prey.
Information on foods eaten has been widely gathered, generally in the context of conflicts with anglers, and not in all seasons. Foods of cormorants wintering in marine areas have not been not described. Representative studies and comparative papers are summarized below, and only major prey species are listed (see Appendix 1 for listing of scientific names of species by family); for additional lists of prey, see Palmer 1962a. Seasonal variation is evident in all appropriately designed studies, but ecological factors contributing to cormorant predation and its impacts are not revealed by mere lists of prey or simple percentages. Occurrence of individual specializations has not been examined. See also Conservation and Management: conflicts, below.
Most data are from regurgitations collected in nesting colonies, or from pellets (especially fish otoliths); contents of gizzards and esophagi of collected birds are less frequently reported. It is not known if chick regurgitations reflect the adult diet. Collecting adults is not only expensive but also likely to undersample known spatial and temporal variation in prey captured each day. Pellets may integrate meals over an entire day, but with unknown completeness. Difficulties with otoliths and other data from pellets include differential loss and digestion by size and/or species (Duffy and Laurenson 1983, Craven and Lev 1987, Johnstone et al. 1990), as well as secondary consumption (evidence for prey's prey; Blackwell and Sinclair 1995, Johnson et al. 1997b). Thus, otoliths are difficult to use for estimates of size or number of fish ingested. Reliance on the abundant dietary remains in pellets may lead to overestimates of invertebrate prey and underestimates of small or soft-bodied fish (Seefelt and Gillingham 2006a). Comparing food items from pellets to those from collected birds in Wyoming, Derby and Lovvorn (Derby and Lovvorn 1997) concluded that some of the differences revealed by the methods had resulted from the cormorants' daily schedules of hunting for different prey. For development of standardized methods, for diet and bioenergetics see Carss 1997 and Ridgway 2010a.
From the nw. Atlantic, Cairns (Cairns 1998) summarized 40 sets of collections (1914–1996) from the Gulf of Maine (16), Atlantic Nova Scotia (7), Prince Edward I. (9), and Quebec (8). Total > 5,000 samples, mostly stomach contents and regurgitations; about 60 fish taxa reported. Salmon were important prey during smolt run in stocked rivers, but no data were available from rivers where runs are exclusively wild. Estimates include percentage composition from biomass and from number (within groups, percentages sum to 100).
During Smolt Runs. (671 samples). Atlantic salmon 26%; other salmonids 4.6%; nonsalmonid freshwater and estuarine fish 53% (including yellow perch, 3.1; white sucker, 3.6; gaspereau [alewife and blueback herring], 7.0; rainbow smelt, 24.5; American eel, 3.8; sticklebacks, 3.3). Marine prey totaled 15.4%, including sand lance, 2.2; sculpins, 3.4; and crustaceans, 3.7.
Outside Smolt Runs. (> 4,554 samples). Freshwater fish, 4.9% (including negligible salmon); brook trout, 1.9 (chiefly reflecting numbers from freshwater ponds on Prince Edward I.); nonsalmonid fish, 2.9. Estuarine/diadromous fish, 23% (including rainbow smelt, 6.4; gaspereau, 5.3; sticklebacks, 3.4; American eel, 2.6; cyprinodonts, 2.1). Pelagic fish, 14% (including sand lance, 5.5; herring, 4.2; capelin, 4.0). Marine bottom fish, 54% (including sculpins 15.7; cunner, 15.6; blennies, 9.4; Atlantic cod 2.9; pollock, 2.2; winter flounder, 2.1; and hake, 1.4. No equivalent studies found for cormorant diets near salmon rivers in the Pacific Northwest, where consumption of salmon smolts angers fishermen (Bayer 1989c).
In Penobscot Bay and River, ME, adult diets showed seasonal changes. In spring, birds feeding on the river consumed 12 fish species and concentrated at dams where they caught many salmon smolts in May. Adults feeding below the head of tide consumed 28 species, including freshwater and seasonally available marine and estuarine forms (Blackwell and Krohn 1997, Blackwell et al. 1997). Nestling diets (n = 742 regurgitated samples collected in Jul at colonies on islands in the bay): >40 taxa; 5 benthic taxa were consistently important: cunner, rock gunnel, sculpins, wrymouth, and sand shrimp (Crangon septemspinosa). Anadromous and catadromous species occurred inconsistently (Blackwell et al. 1995). In contrast to Mendall's (Mendall 1936a) report from the same area in 1935, economically important groundfishes (Atlantic cod, winter flounder) were negligible components of nestling diet in 1993. From the St. Lawrence, a shift from benthic to schooling species over a similar period was reported by Rail and Chapdelaine (Rail and Chapdelaine 1998).
From the Pacific Ocean, samples of pellets and regurgitations from 7 sites from Mandarte I. (British Columbia) to the Gulf of California in the breeding season included 95 taxa, 2,722 individuals; predominantly schooling fish from well above flat bottoms (Ainley et al. 1981a), or inshore benthic species (Robertson 1974). Contributions of >5% (by number) at 1 or more sites were derived from 21 species (in 15 families).
Information from breeders in the se. U.S. is limited, and no studies of coastal wintering birds are reported. In Biscayne Bay, FL, fish samples collected below nests included 507 individuals of 32 species: Gulf toadfish, 35.9% by number (40.8% by weight); blue-striped grunt, 9.1 (6.6); white grunt, 7.3 (5.2); bucktooth parrot fish, 5.7 (5.3); and pinfish, 5.1 (5.6). Parrot fish (Scaridae) together made up 20.5% by number (21.9% by weight); grunts (Hemulidae), 14.4 (12.9); filefish (Balistidae), 6.5 (1.9); barbfish, 3.2 (3.9); sandperch, 2.4 (4.9); and gray snapper, 1.4 (2.7). Mean weight of individual fish was 39 g (Cummings 1987).
From freshwater sites in Florida, important fish in 67 cormorant stomachs were shad, brown bullhead, sunfish, and black crappie at both natural wetlands and phosphate-mine settling ponds, but in the latter habitat mosquito fish was the most frequent and insects were present in unusual numbers (but <0.4% aggregate volume). These included dragonfly nymphs (Aeschnidae, Libellulidae), midge larvae (Chironomus), and water boatmen (Corixidae; O'Meara et al. 1982). It is unclear if some of these insects resulted from secondary consumption.
For the Great Lakes, many diet studies of cormorants, 1986–2013, have been published. In these ecosystems, populations of prey fish have fluctuated widely in response to almost complete loss of native large piscivorous fish in the 1940s and the stocking of several salmonids (Christie et al. 1987). Introduced alewife, abundant in Lake Ontario by the 1880s and in remaining Great Lakes (except Lake Superior) by the mid-1950s, has been important in most studies in Lakes Huron, Michigan, and Superior during Apr–Sep (Belyea et al. 1999, Seefelt and Gillingham 2006a, Seefelt and Gillingham 2008, Dorr et al. 2010a). Ludwig et al. (Ludwig et al. 1989) collected 8,512 regurgitated food items from adults and young birds handled. Alewife and nine-spine sticklebacks each accounted for 41% by number (1 species of crayfish, all the rest were fish, in 18 taxa). By biomass, the most important species were alewife (57%), yellow perch (13%), rainbow smelt (8%), and white sucker (7%); weights of sticklebacks averaged 1.9 g, of alewives 19 g. Diet varied both seasonally and geographically; by Aug, diet at all 4 study areas was 100% alewife in all colonies sampled. Overall, forage fish accounted for 88% of food items (66% of biomass). Fish species of local commercial importance (such as panfish) constituted 12% by number (34% by weight), including yellow perch (3%, 13%), smelt (7, 8), and sucker (<1, 7). No lake trout or common whitefish, the two important commercial fish in Lake Superior, were found. Cormorant diet in Lake Superior in 1983 was examined by Craven and Lev (Craven and Lev 1987).
On Lakes Huron, Erie, and Ontario, 2,133 pellets and boli, collected Apr/May–Jul 1992 and 1994, revealed significant spatial and temporal variation among 8 colonies in the frequency of occurrence of fish species (Neuman et al. 1997): smallmouth bass, 5.3–87.5%; alewife, 0–49.6; smelt, 2.0–65.3; slimy sculpin, 2.5–41.3; white perch, 0–21.4; yellow perch, 4.3–77.8. Nonfish items included crayfish (0–6.4%) and snails (0–1.0%). Variation was attributed to movements of fish into shallow spawning areas and to spatial heterogeneity of fish habitat. Average length of prey consumed varied by species: 7.6–30.5 cm (26 species identified).
A review of 9 years of diet data from 12,024 pellets of cormorants on Little Galloo Island (LGI) in the eastern basin of Lake Ontario indicated that Alewife was the major prey (31.5 %) of LGI cormorants, ranging from 20.8% (post-chick feeding) to 55.5% (chick feeding) numerically over the season (Apr-Oct; see Johnson et al. 2002). Yellow perch was the second most abundant fish in the diet (24.7%) and composed 12.3% (chick feeding) to 34.2% (post-chick feeding) of the diet. Three centrarchid species, pumpkinseed (6.4%), rock bass (5.6%) and smallmouth bass (3.7%) contributed 15.7% of the diet, and cyprinids made up 13.4%. All prey species with the exception of rock bass and slimy sculpin varied over time in their contribution to the diet (see Johnson et al. 2002). On average, LGI cormorants consumed about 32.8 million fish annually (1.4 million kg) comprising at least 28 taxa (Ross and Johnson 1995, Ross and Johnson 1999).
Biomass of smallmouth bass and yellow perch consumed annually by cormorants exceeded what is taken by sport (bass and yellow perch) and commercial perch fishermen (Johnson et al. 2002). The total length of perch consumed by cormorants ranged from 59 to 236 mm, the majority of which were age-1 (48%), age-2 (20%), and age-3 (20%; see Burnett et al. 2002). At a high yellow perch standing stock estimate of 65 kg/ha, cormorants were capable of consuming 29% of the age-3 perch stock (Burnett et al. 2002).
From the Les Cheneaux Is. area of n. Lake Huron in 1995, contents of 373 cormorant stomachs were analyzed to evaluate cormorant impact on yellow perch fishery (Belyea et al. 1997). Yellow perch (75–150 mm) varied from 47% of diet (by weight) in early spring to 2% in late spring and early summer (overall 10%). Sticklebacks and alewives varied from 0 to 45% and from <30 to 85% of diet, respectively, in various months (together, 62% overall). Diet and feeding locations varied according to spawning of the various prey species.
Lake Erie differs from other Great Lakes in that Gizzard Shad (Dorosoma cepedianum) tend to be more prominent in the diet, largely replacing alewife as a forage fish. In the western basin of Lake Erie, Bur et al. (1999) collected 302 cormorants (Apr 15 to Oct 15, 1997) representing 15 species of fish and 2 invertebrates. Cormorant diet composition by weight was Gizzard Shad (48.3%), Freshwater Drum (33.4%), and emerald shiner (8.5%), with remaining prey being < 4% of diet. Hebert and Morrison (2003) reported remarkably similar proportions for these species in the eastern and central basin of Lake Erie, other than rainbow smelt (10%), which was not present in cormorant diets in the western basin.
In some areas of the Great Lakes there has been shift in diet to a recent and abundant invasive species; the Round Goby (formerly, Neogobius melanostomus now Apollonia melanostomus: see Somers et al. 2003, Johnson et al. 2010, Coleman et al. 2012). In w. Lake Ontario in 2002, round goby made up 1.8% to 11% of the number of fish identified from chick regurgitates despite relatively recent establishment of goby populations (Somers et al. 2003). Of 600 cormorant stomachs examined from 2004 to 2007 on the Upper Niagara River, the contribution by weight of Round Goby to the annual cormorant diet ranged from 38% to 85% (Coleman et al. 2012). On average 62% of the biomass in cormorant diets consisted of Round Goby, followed by Gizzard Shad (8%) and Emerald Shiner (7%; see Coleman et al. 2012). Johnson et al. (2010) found a shift in diet of cormorants from 2 colonies in the eastern basin of Lake Ontario from 1999-2007: Pre-Goby years, Yellow perch (19.9% - 39.4%), Alewife (10.9% - 22.6%), and 3 spine stickleback (15.0% – 28.9%); Post-Goby years, Round Goby (74.7% - 79.4%), Yellow perch (7.3% - 11.1%), and Alewife (2.9% - 8.3%).
An examination of 15 yr of diet data collected from Oneida Lake, NY, indicated that cormorant diet changed as the fish community changed (DeBruyne et al. 2013). Diet samples were variable based on season and year with Emerald Shiner (Notropis atherinoides), Gizzard Shad (Lepomis spp.), Logperch (Percina caprodes), Walleye (Sander vitreus), and Yellow Perch (Perca flavescens) having the highest overall relative importance (DeBruyne et al. 2013). In years when age-0 Gizzard Shad were abundant they dominated cormorant diets in the fall after the shad reached a length of 45 mm. Negative effects of cormorants had been documented for Walleye and Yellow Perch (Rudstam et al. 2004), but diet shifts indicate that continued monitoring is needed to fully evaluate potential impacts as fish communities change over time.
On Lake Champlain VT and NY, diet studies pre-and post-alewife establishment showed a shift in diet due to invasion of alewife (DeBruyne et al. 2012). The diets varied considerably by colony location. During 2001– 2002 and pre-alewife invasion, the most common prey items identified for cormorants at Young Island were yellow perch (62-95% by weight). Young Island cormorants consumed less but still large numbers of yellow perch post alewife invasion (overall pre-alewife 46% and post-alewife 38% of identified prey; see DeBruyne et al. 2012). At Four Brothers Islands alewife became a major component of cormorant diets (48% of prey items consumed; see DeBruyne et al. 2012). Yellow perch, remained between 14% and 18% of the overall identified prey items at Four Brothers during the post-alewife period. Overall rainbow smelt frequency decreased from 74% during pre-alewife to 24% during the post-alewife period (DeBruyne et al. 2012).
Dorr et al. (2010b) examined food habits of spring (Apr 9 to May 13, 2005-2007) migrating cormorants congregating at fish spawning sites at two locations; one in the Potagannissing Bay (Drummond Island) area of n. Lake Huron, and Brevoort Lake, MI, just inshore of N. Lake Michigan. The cormorant diet at Drummond Island consisted of 13 species and biomass was primarily yellow perch (43.6%), and minnows/carp (41.6%). Biomass of perch consumed at Potagannissing Bay exceeded that of the bays fishery. At Brevoort Lake the cormorant diet consisted of 20 fish species and 1 crustacean. Biomass of the diet was primarily yellow perch (57.7%) and sunfishes (15.1%), followed by round goby (5.9%) and walleye (5.6%). The observed level of cormorant predation could consume 55% of the 17-year record age-1 walleye year-class of 2006.
On Lake Winnipegosis, Manitoba, in 1987, 6,169 prey samples from regurgitates represented 16 fish species and 1 crayfish (Orconectes sp.; Hobson et al. 1989). Three species constituted >90% frequency (biomass): yellow perch, 63.8 (27.6); white sucker, 14.2 (46.4); and tullibee, 14.0 (20.4). Commercially valuable species, walleye and sauger, accounted for maximum 0.1 and 0.2% of prey biomass, respectively.
Amphibians are rarely reported as prey, but salamander Ambystoma occurred in up to 25% of stomach samples in Wyoming (Derby and Lovvorn 1997) and mudpuppy Necturus in New York (C. Adams unpubl.). Crustaceans are frequently recorded as prey, but usually in small numbers (e.g., 0.2% frequency; Hobson et al. 1989), and some may be derived from gut contents of fish prey. They are less frequent in stomach samples than in pellets (Johnson et al. 1997b). In Penobscot Bay, ME, however, sand shrimp was among the highest-ranking prey taxa in nestling regurgitations (average percentage by volume 0.9–3.3; Blackwell et al. 1995) and also for spring adults feeding in the estuary (to 45.8%; Blackwell et al. 1997). Eight species of crustaceans were identified in nestling regurgitates (Blackwell et al. 1995). At aquaculture sites in Louisiana, some wintering Double-crested Cormorants feed on crawfish (Procambarus spp.) and shrimp (Macrobrachium spp.; J. Huner pers. comm.). Other invertebrates reported in small numbers include gastropods and pelecypods, which may often be secondarily consumed (Johnson et al. 1997b).
Winter diets infrequently reported, except near catfish farms. From 420 wintering cormorants shot on 8 reservoirs in Texas, however, stomach contents included only fish, of 29 species. Largest was a catfish 415 mm long, but 90% were ≤125 mm, ≤200 g. Most were shad (79.2% by number, 26.1% by weight) or sunfish (7.8, 15.0). Sport fish (3 catfishes, 2 bass, crappies) together accounted for 3% by number (32% by weight). Other prey included common carp (0.2, 2.2) and blue tilapia (5.2, 18.2). From Nov to Mar, consumption of shad decreased while that of sunfish increased (Campo et al. 1993).
From 142 cormorants collected in winter 1996-97 at Lake Beulah, Mississippi and 51 cormorants collected at Lake Eufaula, on the Alabama/Georgia border, the diet at both sites consisted primarily of shad and sunfishes but also included some catfish (Glahn et al. 1998). Intact fish in the diet averaged 111 mm but varied with fish species. Based on fish availability data from Lake Beulah, cormorants appeared to have a preference for sunfishes, particularly bluegill (Lepomis macrochirus). However, based on bioenergetic projections cormorants consumed only a small percentage of the bluegill available at Lake Beulah, except possibly harvestable size bluegill (Glahn et al. 1998).
Near catfish farms in Mississippi, wintering cormorants that were collected at night roosts in 1989–1991 had eaten >10 species of fish, but channel catfish and gizzard shad constituted >90% of diet by weight. Proportions of these 2 species were approximately equal overall, but differed by month and location, and between sexes of cormorants sampled. Proportion of catfish ranged from 0–27% in Oct–Nov to 90–96% in Feb–Apr. Catfish was most often consumed by males during spring (Feb–Apr) and in areas with more farmed catfish (Glahn et al. 1995, Dorr et al. 2012a, Dorr et al. 2012b, in press). Cormorants foraging in catfish aquaculture-producing areas had greater omental fat reserves (Glahn et al. 1999) and on Lake Erie breeding grounds were in better condition than birds that foraged mainly in marine environments (Hebert et al. 2008)
Some of the recent changes in diet (and increases in cormorant numbers) reported in both marine and freshwater areas are associated with significant changes in underlying fish populations including; declines in in major predatory fish that may be attributable to overfishing—e.g., cod and other groundfish in Penobscot Bay, ME (Blackwell et al. 1995), and walleye in Lake Winnipegosis, Manitoba (Hobson et al. 1989), declines in forage species such as alewife, (Fielder 2008, Fielder 2010) and changes in lake ecosystems due to many factors including water quality and invasive species such as dreissenid mussels, alewife, and round goby (Williams et al. 1998, Lantry et al. 2002, Somers et al. 2003, Fielder et al. 2008, Fielder et al. 2010, Johnson et al. 2010, Coleman et al. 2012, Debruyne et al. 2012, DeBruyne et al. 2013).