The diets of common and Briinnich's guillemots Uria aalge and U. Zomvia in the Barents Sea region ROBERT T. BARRETT, VIDAR BAKKEN and JURI V . KRASNOV Barrett, R. T., Bakken, V. & Krasnov, J . V. 1997: The diets of common and Briinnich's guillemots Uria aalge and U. lomvia in the Barents Sea region. Polar Research 16(2). 73-84. Guillemots Uria spp. account for ca. 70% of the total harvest of prey taken by seabirds breeding in the Barents Sea region. This paper presents guillemot chick diet data collected recently at four localities (Finnmark, Murman, Bjplmplya and Spitsbergen) and collates all the data found by the authors in the literature and in the archives of Tromspl Museum, the Norwegian Polar Institute and Kandalaksha State Nature Reserve. Guillemots consume a wide variety of prey and, in comparison to the harvest by predatory fish and marine mammals, their impact on the Barents Sea ecosystem is considered to be minimal. We point out the need for more systematic collection of data from different regions and at different times of the year before a final impact assessment can be made. Robert T. Barrett, Zoology Department, Troms# University Museum, N-9037 Trams#, N o w a y ; Vidar Bakken, Nowegian Polar Institute, P . O . Box 5072 Majorstua, N-0301 Oslo, Notway; Juri V. Krasnov, Kandalaksha State Nature Reserve, Kandalaksha, Munnanskaja obl.. Russia 184040. Introduction The Barents Sea is one of the most productive marine ecosystems in the world with vast stocks of commercially important fish, seabirds and sea mammals (Sakshaug et al. 1994). Of the ca. 3.7 million pairs of seabirds (with a biomass of ca. 4200 tonnes) that breed in the Barents Sea region (as defined by Mehlum & Gabrielsen 1995), Briinnich's guillemots Uria lomvia predominate both in numbers (ca. 1.6 million pairs = 43%) and biomass (ca. 2600 tonnes = 61%). They are in addition the major consumer among the Barents Sea seabirds, taking in the order of 60% of an estimated 2213 tonnes of food consumed daily by breeding seabirds during the summer (Mehlum & Gabrielsen 1995). When including the ca. 266,000 pairs of common guillemots Uria aalge which also breed in the Barents Sea, the total consump- tion by adults and chicks of both guillemot species has been estimated to be ca. 1500 tonnedday or nearly 70% of the total seabird consumption by breeding seabirds (Mehlum & Gabrielsen 1995). However, guillemots are known to consume a wide variety of prey (Erikstad & Vader 1989; Vader et al. 1990; Barrett & Krasnov 1996), and this paper addresses the impact guillemots may have on the Barents Sea ecosystem. It first documents considerable diversity in prey use at four breeding colonies in the southern and western Barents Sea regions from 198CL-1995 (Fig. 1) and evaluates data found in the literature and in the authors' respective institution's archives. Regio- nal diet patterns are then presented, impact analyses are made and suggestions are given as to where effort should be made in future feeding studies. Little emphasis is, however, placed on temporal or inter-species statistical comparisons as these are dealt with in more detail by Barrett & Krasnov (1996) and Barrett et al. (1997), respectively. Methods and material Most of the recent data concerning the diet of both species during the breeding season were collected by the authors by direct observation of food items brought to the chick by the adults. Observations were made from 5-20 m using binoculars. In most cases the food items could be identified to species level. Identification controls were made by catching fish-carrying birds with a noose-pole and identifying the food item in the hand. Such studies have been carried out on B j ~ m g y a (Bear Island) (74"21'N, 19"06'E) nearly every year since 1988, on Hornoya, East Finnmark (70"22'N, 31"lO'E) since 1980 and on Kharlov, off the Kola Peninsula (68"49'N, 37"20'E) since 1985 (Fig. 1). 74 R. T. Barrett et al. 80’N 70% Fig. 1. Map of the Barents Sea showing the four colonies where guillemot diet data were collected by the authors (solid circles) and other localities mentioned in the text. Open circles indicate approximate positions of sampling sites listed in Tables 1 and L Briinnich’s guillemot chick diet was also studied on Kovalskifjellet, Spitsbergen (77”03’N, 17’17’E) in 1989 (Mehlum & Gabrielsen 1993) and 1992. In most years, observations were spread over two or three weeks during the main chick- feeding period (July) thereby reducing potential errors in shorter-term sampling due to seasonal changes in diet. More details are given by Barrett et al. (1997). Because Briinnich’s guillemots generally land directly on their breeding site, conceal the prey by bowing their heads, and feed their chicks almost immediately on arrival, sample sizes for this species on Hornoya and Kharlov were often much smaller than for conunon guillemots which commonly hold the fish aloft before feeding the chicks. Some stomach samples from adults were also collected from both species on Kharlov in 1992 and from Briinnich’s guillemots on Novaja Zemlja (1992) and Franz Josef Land (1993). Comparisons of diets between colonies were made using x2 tests, and p-values <0.05 were considered to be significant. When numbers of food items from individual food categories were low in a sample, several categories were com- bined before testing. Supplementary data gleaned from the literature and the Norwegian Polar Institute, Troms@ Museum and Kandalaksha State Nature Reserve archives differed so much in degree of quantifi- cation that they were impossible to compare directly. Furthermore, some of the data were based on observations of chick diet while other data were based on adult stomach samples. Because the relationship between chick and adult diet is unknown, chick and adult data are presented separately. Results Chick diet Capelin Mallotus villosus, sand eels Ammodytes sp., and I-group herring Clupea harengus (Barrett & Krasnov 1996) constituted most of the diet of common guillemot chicks on Kharlov, Bjornoya and Hornoya (Fig. 2). There were, however, clear differences among the colonies in the relative composition of the diets. On Kharlov, sand eels were the most important prey items fed to common guillemot chicks. In some years, capelin was also important, and, in 1986, dominated the diet (Fig. 2). Herring Fig. 2 . Composition (% by number) of diet fed to chicks of common guillemots on Kharlov, Bjgmgya and Horngya. Figures above columns = no. of fish counted. 21” indicates that the fish were collected on the neighbouring colony, Syltefjord. (See Fig. 3 text for English names of prey.) The diets of common and Briinnich’s guillemots 75 occurred in some years but generally constituted 110% of the items recorded. In 1995, hemng made up 25% of the fish recorded. On Bjarnaya, capelin constituted 90% or more (by number) of the diet in five of the six seasons for which data exists. In the sixth year (1995), capelin made up >60% and the little squid Gonatus fabricii accounted for most of the rest (Fig. 2). On Horngiya, capelin was again the most common prey item caught by common guillemots, while sand eels were regularly eaten but in lower proportions (Fig. 2). After 1990, hemng was a periodically important constituent of the chick diet. In nearly all the years for which there is comparable data from Hornaya, Kharlov and Bj~rnaya, there were significant differences in the diet composition among the colonies (x’ > 13, df = 1-3, p < 0.001). The only insignificant differences found were between Hornaya and Kharlov in 1985 and 1991, but these may have been due as much to the small sample sizes as to the actual geographical differences. While Briinnich’s guillemots tended to feed their chicks a much more varied diet than did common guillemots on Bj@rn@ya, capelin, sand eels and I-group hemng again made up the samples collected on Hornaya and Kharlov (Fig. 3). On Bjamaya, the Briinnich’s guillemot samples contained six different prey items (capelin, squid, sculpins (Cottidae), polar cod Boreogadus saida, blennies Lumpenus sp. and eelpouts (Zoarcidae)) which occurred in varying proportions. Capelin were, however, the most common prey with squid constituting ca. 30% of the food items in three of the five years. There were significant differences between the prey compositions on Hornoya and Bjornoya in the three years data were collected on both colonies (1989, 1991, 1993, x2 > 30, df = 3, On Kovalskifjellet, Briinnich’s guillemots fed their chicks almost entirely on polar cod in 1992. Two of the 148 food items observed were crustaceans, the only two specimens of this taxa observed fed to chicks in any of the colonies studied. On Kharlov, sand eels constituted 100% (in 1992) and 80% (in 1994) of the fish seen. The remaining fish recorded in 1994 were capelin (14%) and hemng (5%). p < 0.001). Adult stomach samples Of eight adult Briinnich’s guillemot stomachs sampled on Kharlov in 1992, three contained capelin, two contained herring, two contained gadoids and one contained sand eels. Data from the literature Chick diet Common guillemots on Hjelmsoy, West Finn- mark, fed their chicks almost exclusively capelin in 1983, but in 1984 saithe Pollachius virens dominated their diet (Vader et al. 1990; Tromsgi Museum unpubl.). On Rost, Lofoten, saithe, sand eels and occasionally butterfish Pholis gunnellus Fig. 3. Composition (% by number) of diet fed to chicks of Briinnich’s guillemots on Homoya, Kharlov, Bjemeya and Kovalskifjellet, Spitsbergen. Numbers above columns = no. of fish counted. C. harengus = hemng, M . villosus = capelin, Ammodytes sp. = sand eel, G. fubricii = squid, Lumpenus sp. = blenny, B. suida = polar cod. -L T ab le 1 . F oo d ite m s in th e di et o f co m m on g ui lle m ot s in th e B ar en ts S ea re co rd ed in th e lit er at ur e. N = n um be rs o f ad ul t s to m ac hs (S ) o r c hi ck f oo d ite m s (F ) s am pl ed . N um be r af te r a s to m ac h fo od i te m = n o. o f st om ac hs c on ta in in g th at f oo d ite m . ? = s am pl e si ze n ot g iv en . $ % D ur in g br ee di ng s ea so n in o r ne ar c ol on ie s N ov ja Z em lj a D at e N F oo d it em s 19 47 1 3 s Po la r co d 13 , A tla nt ic c od 3 . 19 48 -5 0 20 s Po la r co d 8, A tla nt ic c od 2 , sa nd e el s 3, c ap el in 2 , s cu lp in 2 . B j@ m @ ya 19 48 19 96 ?F ?S M ai nl y sm al l g ad oi d fi sh M ai nl y eu ph au si id s K ol a P en in su la ( K hn rl ov ) 19 35 46 F A tla nt ic c od , sa nd e el a nd h er ri ng , Fi g. 4 . 19 38 19 8F M ai nl y sa nd e el , Fi g. 4 . 19 37 -3 9 85 F M ai nl y ca pe lin a nd s an d ee l, Fi g. 4 . 19 47 -4 9 19 60 21 s 19 92 9 s 2 4 ,4 7 & 4 1s A tla nt ic c od , sa nd e el a nd h er ri ng , Fi g. 4 . G ad oi ds 1 3, c ap el in 5 , c ru st ac ea 3 a nd N er ei s 1. Sa nd e el s 5, h er ri ng 2 , A tla nt ic c od I , ca pe lin 1 . N o w a y 19 70 s 19 83 19 84 19 87 ?F 79 F 24 F 1 8 s V ed oy , R gs t. M ai nl y sa ith e an d sa nd e el s, o cc . P ho lis g un ne llu s H je lm sg y. 9 9% c ap el in . H je lm so y. S ai th e 16 , s an d ee l 3, A na rh ic ha s lu pu s 3, s qu id 1 . B le ik sg y. M os tly s qu id b ea ks . O ut si de b re ed in g se as on in c oa st al w at er s A pr 1 98 6 1 7 s E as t Fi nn m ar k. E xc lu si ve ly c ap el in . A pr 1 98 5 ? T ro m sa . M os tly c ap el in . So ur ce K an da la ks ha R es er ve a rc hi ve s U sp en sk i 19 56 Se rg ea nt 1 95 1 M eh lu m e t al . i n pr es s B el op ol ’s ki i 19 57 K af ta no vs ki 1 93 8 K an da la ks ha R es er ve a rc hi ve s B el op ol ’s ki i 19 57 G er as im ov a (i n K ra sn ov e t al . 19 95 ) K ra sn ov u np ub l. T sc ha nz & B ar th 1 97 8 V ad er e t al . 19 90 T ro m sg M us eu m u np ub l B ar re tt un pu bl . E ri ks ta d & V ad er 1 98 9 V ad er e t al . 19 90 The diets of common and Briinnich’s guillemots 77 were fed to chicks in the 1970s (Tschanz & Barth 1978). In 1938, the only single year for which there is published data for the Briinnich’s guillemots on Kharlov, sand eels constituted >80% of the diet of both species (Kaftanovski 1938). Of 20 fish observed in 1994 being fed to Briinnich’s guillemot chicks on Novaja Zemlja by Strom et al. (1994), seven were polar cod and four were capelin. On Bjernoya, chicks of both guillemot species were fed mainly fish (Roi 191 1; Duffey & Sergeant 1950; Sergeant 1951). These were identified as “mainly gadoids” and may well have been polar cod. On Kovalskifjellet in 1989, Briinnich’s guillemot chick diet consisted almost entirely of polar cod (Mehlum & Gabrielsen 1993). Adult stomach samples Most of the data found in the literature were based on stomach samples from adult birds shot or otherwise collected on or near breeding colonies. Four other studies were of birds collected in ice- covered waters during the summer; another was from the central Barents Sea and five were from coastal waters during the autumn and spring (Tables 1 and 2). The methods of quantification varied considerably among the studies, from near anecdotal notes (e.g. the 19th century references) to detailed estimates of frequencies of occurrence of prey items (e.g. Lydersen et al. 1989; Erikstad 1990; Mehlum et al. 1996; in press) or percentage contribution by mass or energy (Weslawski et al. 1994). This prohibited any quantitative compari- Fig. 4. Composition (%) of the diet of common guillemots at Seven Islands and of Briinnich’s guillemots at Novaja Zemlja in the 1930s and 1940s. Data from Krasovskii (1937). Kaftanovski (1938), Uspenski (1956), Belopol’skii (1957) and Kandalaksha State Nature Reserve archive. Sample sizes are given in Table 1. (See Fig. 3 text for English names of prey.) sons between sites. Certain patterns do, however, arise from the data. There is little data on the common guillemot from Novaja Zemlja (in Bezymyannaya Bay), and the main food items documented in the late 1940s were polar cod and Atlantic cod Gadus morhua (Table 1). Further west, on Kharlov off the Kola Peninsula, common guillemot stomachs sampled in the late 1930s and 1940s again contained mainly fish (capelin, Atlantic cod, herring and sand eels, Table 1, Krasovskii 1937; Kaftanovski 1938; Belopol’skii 1957). These data are also summarised in Fig. 4. Of nine adult stomachs sampled on Kharlov in 1992, five contained sand eels, two contained herring, one contained Atlan- tic cod and one contained capelin (Table 1). Polar cod and Atlantic cod constituted most of the Briinnich’s guillemot diet on Novaja Zemlja in the late 1940s, but other fish such as sand eels, capelin, herring and sculpins were also eaten (Table 2, Fig. 4) (Krasowski 1937; Uspenski 1956; Belopol’skii 1957). In 1992, 10 of 31 Briinnich’s guillemot stomachs sampled on Novaja Zemlja contained food (Table 2). Gadoids were found in four, snail fish Liparis sp. in two, and capelin, sand eels, sculpins, unidentified fish and crustaceans in one each. In 1993, 12 of 14 Briinnich’s guillemot stomachs analysed in Franz Josef Land contained polar cod, four contained crustaceans, one contained polychaetes, and one contained unidentified fish. Belopol’skii (1971) otherwise summarises the contents of 11 1 Briinnich’s guillemot stomachs sampled on Kharlov in 1935, 1941 and 1 9 4 7 4 9 . Herring was found in 30, sand eels in 18, Atlantic T ab le 2 . F oo d ite m s in th e di et o f B ri in ni ch ’s G ui lle m ot in th e B ar en ts S ea r ec or de d in th e lit er at ur e. R ef er en ce s m ar ke d w ith a n as te ri sk ( *) a re q uo te d fr om L gv en sk io ld ( 19 64 ). N = n um be rs of a du lt st om ac hs ( S) o r ch ic k fo od i te m s (F ) sa m pl ed . N um be r af te r a st om ac h fo od i te m = n o. o f st om ac hs c on ta in in g th at f oo d ite m . ? = s am pl e si ze n ot g iv en . D ur in g br ee di ng s ea so n in o r ne ar c ol on ie s N ov ja Z em lja D at e N 19 25 ? ?F 19 34 62 F 19 42 , 4 7 46 , 29 28 19 48 -5 0 44 , 52 1, & 3 12 s 19 92 31 s 19 94 20 F F ra ns J os ef L an d 19 31 16 s 19 91 -9 3 11 s 19 93 14 s Sv al ba rd 18 89 ? 18 96 ?F 18 98 ? 18 98 ?S 18 98 , 1 90 0 ? ? ? 19 10 ?S 19 24 ? 19 33 4 s 19 67 49 1 s 19 89 16 1F 19 92 23 s F oo d ite m s Po la r co d, g ad oi ds , c ap el in . M ai nl y A tl an ti c co d, F ig . 4. M ai nl y A tl an ti c an d po la r co d, F ig . 4 . V ar ie d: A tl an ti c co d, p ol ar c od , sa nd e el s, F ig . 4. G ad oi ds 4 , sc ul pi ns 2 , A tla nt ic c od 1 , c ap el in 1 , s. e el 1 , L ip ar is 1 , c ru st . 1 Po la r co d 7, c ap el in 4 , sa nd e el s 2, b le nn ie s 2. C ru st ac ea ns 8 , u ni d. f is h 7, p ol yc ha et es 2 . 93 % (b y m as s) p ol ar c od , 6 sp p. o f cr us ta ce a, m ai nl y P ar ar he m is to li be llu la . Po la r co d 12 , c ru st ac ea ns 4 , p ol yc ha et es 1 a nd u ni d. f is h 1. E dg eg ya . A m ph ip od s, m ai nl y G am m ar us lo cu st a. Is fj or de n, P ol ar c od . M id te rh uk en , va n M ije nf jo rd en . Po la r co d. A m ph ip od s an d ot he r cr us ta ce a, fi sh . Sp its be rg en . C ru st ac ea ns , p ol ar c od , L um pe nu s sp . Fi sh . M id te rh uk en , B el ls un d. M ai nl y po la r co d. Y ou ng f ed o n fi sh . A lk ef je lle t, H in lo pe n. A du lts a nd y ou ng f ed e xc lu si ve ly o n fi sh . A ll co nt ai ne d T hy sa no es sa in er m is . E dg eg ya . O nl y ga m m ar id s. 84 % p ol ar c od , 10 % L um pe ni da e, 2 % C ot ti da e, 4 % u ni d. f is h. St or fj or de n. P ar at he m is ro s pp ., T. in er m is , G . w ilk its ki i, po la r co d. So ur ce G or bu no v 19 25 in U sp en sk ii 19 56 K ra so vs ki i 19 37 B el op ol ’s ki i 19 57 U sp en sk i 19 56 K ra sn ov 1 99 5 S tr gm e t al . 19 94 D em m e 19 34 W es la w sk i et a l. 19 94 K ra sn ov 1 99 5 W al te r 18 90 * T re vo r- B at ty e 18 97 * N at ho rs t 19 00 R om er & S ch au di n 19 00 * K ol th of f 19 03 B ir ul ja 1 91 0* M un st er hj el m 1 91 1 M on ta gu e 19 26 * H ar tle y & F is he r 19 36 de K or te 1 97 2 M eh lu m & G ab ri el se n 19 93 M eh lu m e t a l. 19 96 B j@ m #y a 18 99 ? 19 08 ? 19 48 ? 19 95 , 1 99 6 12 , ? K ol a P en in su la ( K ha rl ov ) 19 38 36 F 19 35 , 1 94 0s 11 1s 19 60 11 s N or w ay 19 83 49 F A du lts a te c ru st ac ea , po ly ch ae te s an d fi sh . Y ou ng f ed o n fi sh . Y ou ng f ed o n fi sh . C ru st ac ea ns in a du lt st om ac hs . Y ou ng f ed c hi ef ly s m al l (c a. 5 0 m m ) ga do id s. O ne s to m ac h co nt ai ne d re m ai ns o f ga do id s & p ol yc ha et es . M ai nl y eu ph au si id s T. in er m is . 86 % s an d ee l, 8% h er ri ng , 3 % g ad oi d, 3 % “ ot he r” . H er ri ng 3 0, s an d ee ls 1 8, g ad oi d 16 , c ap el in 1 4, c ru st ac ea ns 5 C ap el in 5 , g ad oi d 5, S eb as te s 1. H je lm sg y. 3 3% c ap el in , 2 4% s qu id , 2 2% s an d ee l. D ur in g br ee di ng s ea so n in o pe n w at ed ic e- co ve re d w at er s Sv al ba rd 19 85 /8 6 18 s 19 86 16 s 19 82 -8 7 13 s Pe la gi c, i ce -c ov er ed w at er s. P ar at he m is to l ib el lu la , p ol ar c od . 19 84 -8 5 12 s K on gs fj or de n. B en th ic a m ph ip od s, p ol ar c od , u ni d. f is h. Fi rs t ye ar i ce . M ai nl y P an da lu s bo re al is , a m ph ip od s an d po la r co d. M ul ti- ye ar i ce . 4 8% f is h, m os tly p ol ar c od . 33 % a m ph ip od s, m os t G . w ilk its ki i. O ut si de b re ed in g se as on in c oa st al a n d o pe n w at er s S ep 1 98 4 21 s S pr in g 19 85 -8 7 7 6 s S pr in g 19 85 14 s H om su nd . T . in er nt is , ga m m ar id s, p ol ar c od . F eb 1 98 7 3 0 s H op en . P . lib el lu la a nd p ol ar c od . M ar 1 98 7 24 s Se nt ra lb an ke n. A tla nt ic c od , p ol ar c od a nd c ru st ac ea ns . A pr 1 98 6 28 s E as t Fi nn m ar k. E xc lu si ve ly c ap el in . A pr 1 98 5 ?S T ro m sg . M os tly c ap el in . H om su nd . P ol ar c od d om in at ed , f ol lo w ed b y P . lib el lu la a nd P ol la ch iu s vi re ns M ar gi na l ic e zo ne . P . lib el lu la , p ol ar c od , T. in en ni s, P an da lu s bo re al is . S w en an de r 19 00 R oi 1 91 1 D uf fe y & S er ge an t 19 50 M eh lu m e t al . in p re ss K af ta no vs ki 1 93 8 B el op ol ’s ki i 19 71 G er as im ov a in K ra sn ov e t al . 19 95 V ad er e t al . 19 90 L gn ne & G ab ri el se n 19 92 Lp cn ne & G ab ri el se n 19 92 M eh lu m & G je rt z 19 84 , G je rt z et a l. 19 85 , M eh lu m & G ab ri el se n 19 93 M eh lu m & G ab ri el se n 19 93 L yd er se n et a l. 19 89 M eh lu m & G ab ri el se n 19 93 M eh lu m & G ab ri el se n 19 93 B ak ke n 19 90 E ri ks ta d 19 90 E ri ks ta d & V ad er 1 98 9 V ad er e t al . 19 90 Y z- a 3 R d 4 W 80 R. T. Barrett et al. cod in 16, capelin in 14 and crustaceans in five (Table 2 ) . Very little data have been published from any of the huge Briinnich’s guillemot colonies on Svalbard. Adult stomach samples and a few observations of food fed to chicks indicate that polar cod and crustaceans (mainly amphipods) are the most frequent items taken during the breeding season (Table 2 ) . However, prior to and after breeding, the diet of Briinnich’s guillemots shot in or near the ice edge consisted nearly solely of crustaceans dominated by amphipods Parathe- misto libellula and Gammarus spp. and, in one study, the prawn Pandalus borealis (Table 2 ) . Three studies of adult Briinnich’s guillemots at and around Bjgrngya (Swenander 1900; Roi 191 1; Mehlum et al. in press) recorded crustaceans, polychaetes and fish among stomach contents. There are also little data from the coast of Norway, but it seems that fish again constituted the total diet of both species with capelin comprising 100% of the prey of both species when they gathered off the coast of Tromsg and Finnmark just before breeding in the mid-1980s (Tables 1 & 2). Squid beaks dominated the food remains found in 18 stomachs of adults sampled off Bleiksgya, Vesterailen, in June and July 1987 (Barrett unpubl.). The total number of prey taxa recorded in the Briinnich’s guillemot diet in the Barents Sea region is at least 37, including 17 crustaceans, 16 fish, molluscs, polychaetes and squid. Thirteen taxa, mostly fish (capelin, herring, sand eel, Atlantic cod, polar cod, wolf fish Anarhichas sp., butterfish, sculpins (Cottidae), saithe and red fish Sebastes sp.) but also squid, polychaetes and molluscs have been documented in the common guillemot diet. Discussion General comparisons of Briinnich’s and common guillemot diets in Canada are discussed by Tuck (1960) and Bradstreet & Brown (1985). This paper presents the first such comparison from the Barents Sea region. As in the Canadian studies, Briinnich’s guillemots in the Barents Sea have a diet of fish and crustaceans, many of which are benthic or associated with ice. Their diet is also much more varied than that of common guille- mots which catch mainly small pelagic, schooling fish. More details concerning the food differences between the two species at Horngya and Bjgrngya are discussed by Barrett et al. (1997), while annual differences in chick diet at Horngya and Kharlov are discussed in relation to prey avail- ability by Barrett & Krasnov (1996). It is, however, likely that adult birds consume food different from what they feed their chicks. This was found, for example, in gulls (Laridae) by Nogales et al. (1995) and partly shown for common guillemots on Kharlov in 1935 and 1992 by Belopol’skii (in Kandalaksha State Reserve archives) and JVK (Table l ) , respec- tively. On Kharlov, chicks were fed mainly high energy, pelagic and schooling fish (capelin, herring and sand eel) while many adult stomachs also contained more benthic fish such as Atlantic cod and other gadoids. Similarly, Briinnich’s guillemot chicks on Kovalskifjellet were mainly fed polar cod in 1992 (Fig. 3), while adults shot offshore, near the colony in the same season, had large proportions of amphipods and euphausiids in their stomachs (Mehlum et al. 1996). The same pattern was found at Bjernaya in 1993 when stomachs of adults shot offshore contained pre- dominantly euphausiids Thysanoessa inermis (Mehlum et al. in press), while chicks were fed mainly capelin and polar cod (Fig. 3). A study of both guillemot species in the same area in 1996 showed that euphausiids were also the dominant prey of common guillemots (Mehlum et al. in press). The potential for such differences between adult and chick diets, plus the paucity of data of either kind from all parts of the Barents Sea, seriously restricts the identification of definitive, broad-scale geographic trends. Despite this, the data collected along the mainland coasts of Norway and Russia do suggest that capelin is especially important as guillemot prey in the early spring (March-April). This is the time when huge flocks of both guillemot species gather to feed on the mature, energy-rich capelin as it approaches the Finnmark coast to spawn (Barrett 1979; Erikstad & Vader 1989; Strann et al. 1991; Krasnov 1995; Nikolaeva et al. 1996). These flocks consist not only of birds which breed in the Barents Sea colonies but also of young common guillemots from, for example, British colonies (Strann et al. 1991). Furthermore, the distribution of guillemots out at sea early in the year (January-March) and in August-September also seems to coincide with the distribution of capelin (Fauchald & Erikstad 1995; Krasnov The diets of common and Briinnich’s guillemots 81 unpubl.). This suggests that capelin is an im- portant prey for both species in the central and southern Barents Sea for at least 4-6 months of the year. Although the numbers of capelin along the coast drop after the spawning season, there are always some late spawners available to guillemots throughout the summer (Barrett & Furness 1990; Barrett & Krasnov 1996). At least half the numbers of capelin fed to common guillemot chicks on Hornpya during seven breeding seasons in the 1980s and 1990s were gravid females (Barrett unpubl.). That fewer capelin are taken at Kharlov reflects the fact that while capelin almost always occur and spawn in East Finnmark, their movements along the Kola Peninsula are limited to the years when large influxes of warm Atlantic water spread further east than normal (Ozhigin & Luka 1984; Gjpszter et al. 1994). Similarly, when herring stocks are at their Iowest (as was the case in the early 1980s), the youngest year classes of herring do not spread as far east along the coast as they do when the stocks are large (Dragesund et al. 1980). Such variations in stock levels explain the near absence of herring in the Hornpya samples before 1985 and the subsequent increase in the 1990s (Fig. 2) and the differences in the amount of herring caught by guillemots on Kharlov and Hornpya (Barrett & Krasnov 1996). Unfortunately little is known about the distri- bution of sand eels in the southern Barents Sea, but there are spawning grounds around Kharlov which are large enough to support a periodic local fishery (Krasnov & Barrett 1995) and hence contribute to the predominance of sand eels in the local guillemot chick diet (Figs. 2 and 3). The data from the literature also document the importance of crustaceans in the diet of adult Briinnich’s guillemots around Spitsbergen and Franz Josef Land. The dominance of crustaceans in the diet is especially true of birds sampled at sea around Spitsbergen during the late winter, spring and summer. Food samples in five of ten studies were dominated by the large hyperiid amphipod Parathemisto libellula, three by other crustaceans (amphipods or euphausiids) and only two by fish (polar cod). Among the 14 studies of birds collected in or near colonies around Spitsbergen and Franz Josef Land, seven recorded crustaceans as the main prey type (Table 2). Among the crustaceans listed are the amphi- pods P . libellula and Gammarus wilkitskii which are often associated with waters near or covered by ice (Sakshaug et al. 1994) and which are much more abundant than, for example, polar cod. Although of poorer quality energetically (Gabrielsen et al. 1994; Gabrielsen pers. comm.), these amhipods are more widely spread, probably more accessible, and thus taken in preference to solitary fish hiding in the ice or on the sea bottom. However, because guillemots chicks are fed food items brought to them singly, the adults probably seek out the larger fish to make the transport into the colony energetically worthwhile. This would explain the high proportion of polar cod in the diet of chicks at Kovalskifjellet compared to the crustacean-rich diet of the adults collected off- shore in 1992. Despite the differences in chick and adult diets, geographical variation in the species of fish caught is apparent from our data and those found in the literature. Polar cod is an important prey in Svalbard, Franz Josef Land and Novaja Zemlja but not along the coasts of Norway and the Kola Peninsula where capelin, sand eel and herring dominate. The restriction of polar cod to the colder northern and eastern localities stems from the relatively limited distribution of the species’ spawning grounds. The main spawning ground in the southeast comer of the Barents Sea, between Kolguyev and Novaja Zemlja, and a smaller one is located east of Spitsbergen (Gjpszter et al. 1994). Adult polar cod are spread over much of the Barents Sea during the summer and are thus taken in very small proportions by guillemots on Bjprnpya where there are otherwise more acces- sible stocks of capelin. The main bulk of the polar cod population avoids the southwestern areas which is influenced by the inflow of warm Atlantic water (Gjpszter et al. 1994). Other fish recorded in the literature and in our studies in the high arctic waters included benthic species of blennies (Lumpenidae or Stichaeidae), sculpins and eelpouts (Zoarcidae). The absence of these species in the more southern samples probably reflects the more northerly distribution of some of the species, e.g. Lumpenus sp., as much as the relative shallowness of the water (enabling the guillemots to reach the benthic species) in the northern areas where guillemots were sampled. The occurrence of the squid Gonatus fabricii in the diets of guillemots on Bjprnoya and Bleiks@y reflects the proximity of these colonies to the main flow of Atlantic water in the Norwegian and Barents seas with which this species is often associated (Wiborg 1979). 82 R. T. Barrett et al. Ecosystem effects Of the ca. 1.6 million pairs of Briinnich’s guillemots which bred in the Barents Sea in 1986, ca. 405,000 pairs bred on the southeastern coasts of Svalbard (excluding Bjomoya) and 1 million pairs on Novaja Zemlja. Furthermore, 245,000 of the total of 266,000 pairs of common guillemots bred on Bjornoya together with ca. 100,000 pairs of Briinnich’s guillemot (Mehlum & Gabrielsen 1995). In other words, ca. 25% of the guillemot food consumption during the breeding season in the Barents Sea in the mid 1980s took place around Svalbard, ca. 50% near Novaja Zemlja, ca. 20% around Bjornaya and ca. 5% along the coast of North Norway and the Kola Peninsula. The annual food requirements of both guillemot species in the Barents Sea is approximately 70% of t h e t o t a l s e a b i r d r e q u i r e m e n t s : i.e., 70% x 3500 TJ (1 TJ = 10l2 Joules) = 2450 TJ (Sakshaug et al. 1994; Mehlum & Gabrielsen 1995). If all the guillemots preyed on one or two fish species only, for example, capelin or polar cod, their annual harvest would have had a significant impact on these single prey popula- tions, especially in years when the fish stocks were at a minimum. For example, when the capelin stocks plummeted to <0.5 million tonnes in 1986/87, their annual production was in the order of 1-2000TJ only (Sakshaug et al. 1994). The same would have applied to the polar cod when stocks were at a minimum in e.g. the early 1970s and early 1980s (Gjosater 1995). However, this paper has shown that the guillemot diet constitutes many species and varies considerably both spatially (Svalbard; mainly crustaceans, Novaja Zemlja; polar and Atlantic cod, BjornGya; euphausiids, capelin and periodically squid, Nor- way and the Kola Peninsula; capelin, sand eels, herring and Atlantic cod) and temporally. Guille- mots also take prey items from the lower trophic levels where annual production is an order of magnitude higher (Sakshaug et al. 1994). As a result, the overall impact of the guillemot population on the Barents Sea ecosystem is very slight compared to the annual production at the different trophic levels and of the different fish populations (Sakshaug et al. 1994). Furthermore, estimates of the annual food requirements of the top predators in the Barents Sea (Atlantic cod, seals, whales and seabirds) by Sakshaug et al. (1994) show that the total seabird harvest is equivalent to only ca. 7% of the total food requirements of higher predators in the Barents Sea, or to ca. 25% of the average annual capelin fishery outtake in the early 1980s (Sakshaug et al. 1994). On the other hand, the impact of the ecosystem on the guillemots can be great. This was illustrated in 1986/87 when the collapse in the capelin stocks resulted in huge declines in the breeding populations of common guillemots on Bjomoya, in Finnmark, and on the Kola Peninsula (Vader et al. 1990; Krasnov & Barrett 1995). More dietary data are needed before a full assessment can be made and a prey-consumption model for these species in the Barents Sea can be developed. This applies especially to the region around Novaja Zemlja where huge numbers of guillemots breed but where their diet is poorly documented. Further comparisons of chick and adult diets should also be made in the various regions to test if the documentation of chick diet (which involves a non-destructive method) can also be used to monitor adult food consumption during the breeding season. Finally, most data have, up to now, been collected during the summer and very little is known about the diet of guillemots at other times of the year. There is a need for systematic collection of birds from different parts of the Barents Sea and at different times of the year, especially from areas where large numbers are known to spend at least part of the autumn and winter (Fauchald & Erikstad 1995; Krasnov unpubl.). Acknowledgements. - Parts of this study were financed by the Norwegian Polar Institute, The Norwegian Research Council, Tromsp, Museum. the University of Tromsp, and the Kandalak- sha State Nature Reserve. We are grateful to numerous field assistants and to W. Vader, G. W. Gabrielsen, F. Mehlum and two referees for their constructive comments and criticisms of earlier drafts of this paper. References Bakken, V. 1990: The distribution and die1 movements of Btiinnich’s Guillemots Uria lomvia in ice covered waters in the Barents sea, Februarymarch 1987. Polar R e x 8, 55-59. Bamett, R . T. 1979: Small oil spill kills 10-20,000 seabirds in North Norway. Mar. Poll. Bull. 10. 253-255. Barrett, R. T., Asheim, M. & Bakken, V. 1997: Ecological relationships between two sympatric congeneric species, Common and Thick-billed Murres, Uria aalge and (1. loniviu, breeding in the Barents Sea. Can. J . Zoo/. 75, 61 8-63 1. The diets of common and Bdnnich’s guillemots 83 Barrett, R. T. & Furness, R. W. 1990: The prey and diving depths of seabirds on Homdy, North Norway after a decrease i n the Barents Sea capelin stocks. Ornis Scand. 21, 179-186. Barrett, R. T. & Krasnov, Y. V. 1996: Recent responses to changes in stocks of prey species by seabirds breeding in the southern Barents Sea. ICES J. Mar. Sci. 53, 713-722. Belopol’skii. L. 0. 1957: Ecology of Sea Colony Birds of the Barents Sea. Israel Progr. For Scientific Translations, Jerusalem. 346 pp. (Translated from Russian, 1961). Belopol’skii, L. 0. 1971: Sostav kormov morskikh ptits Barentseva moqa (The diet of Barents Sea seabirds). Uchenye zapiski Kaliningradskogo gos. universiteta, Kali- ningrad 6, 41-67. Birulja, A. 1910: Zoologiceskie reul’taty russkich 2kspedicij na Spicbergen: Biologiceskija nabludenija nad pticanii Spicbergena. Annu. Mus. Zool. Acad. 15. St. Petersbourg. Bradstreet, M. S. W. &Brown, R. G. B. 1985: Feeding ecology of the Atlantic alcidae. Pp. 263-318 in Nettleship, D. N. & Birkhead, T. R. (eds.): The Atlantic Alcidae. Academic Press, London. de Korte, J. 1972: Birds, observed and collected by the “De Nederlandse Spitsbergen Expeditie” in West and East Spitsbergen, 1967 and 196849; third and last part. Beaufortia 20, 23-58. D e m e , N. P. 1934: Ptichii bazar na skale Rubini (Ostrov Gukera, Zemlya Frantsa Iosifa) (Bird nesting colony on the cliff ledge of Rubini Island (Hooker Island, Franz Josef Land). Trudy Arkticheskogo institutu 11, 55-86. Dragesund, 0.. Hamre, .I. & Ulltang, 0. 1980: Biology and population dynamics of the Norwegian spring-spawning hemng. Rapp. P.-v. Rlun. Cons. Int. Explor. Mer. 177, 43-71, Duffey, E. & Sergeant, D. E. 1950 Field notes on the birds of Bear Island. Ibis 92, 5 5 4 4 6 3 . Erikstad, K. E. 1990 Winter diets of four seabird species in the Barents Sea after a crash in the capelin stock. Polar B i d . 10, 6 19-627. Erikstad, K. E. & Vader, W. 1989: Capelin selection by common and Briinnich’s guillemots during the prelaying season. Ornis Scand. 20, 151-155. Fauchald, P. & Erikstad, K. E. 1995: The predictability of the spatial distribution of guillemots (Uria spp.) in the Barents Sea. Pp. 105-122 in Isaksen, K. & Bakken, V. (eds.): Seabird populations in the northern Barents Sea. Norsk Polarinst. Medd. 15. Oslo. Gabrielsen, G. W., Ryg, M., Mehlum, F., Markussen, N. M. & ehitsland, N. A. 1994: Sjbfugl og sjbpattedyr. Pp. 203-221 in Sakshaug, E., Bjbrge, A,, Gulliksen, B., Loeng, H. & Mehlum, F. (eds.): 0kosystem Barentshavet. Norges Forsk- ningsrld, Univ.forlaget, Oslo. Gjertz, I., Mehlum, F. & Gabrielsen, G. W. 1985: Food sample analysis of seabirds collected during the “Lance”-cruise i n ice-filled waters in Eastern Svalbard 1984. Norsk Polar Insf. Rapp. Ser. No. 23, 1-17. Gjbsaeter, H. 1995: Pelagic fish and the ecological impact of the modern fishing industry in the Barents Sea. Arctic 48, 267-278. Gjdsaeter, H., Godd, 0. R. & Ulltang, 0. 1994: De viktigste fiskeslagene i Barentshavet. Pp. 181-202 in Sakshaug, E., Bjdrge, A., Gulliksen, B., Loeng, H. & Mehlum, F. (eds.): 0kosystem Barentshavet. Norges Forskningsrld, Univ.forla- get, Oslo. Gorbunov, G. P. 1925: Ptich’i bazary Novoi Zemli (Cliff-ledge nesting colonies of birds on Novaja Zemlja). Trudy nauchno- issledovatel’skogo instituta p o izucheniyu Severa 26, 1 4 7 . Hartley, C. H. & Fisher, J. 1936: The marine foods of birds in an inland fjord region in west Spitsbergen. J . Anim. Ecol. 5 , 37G389. Kaftanovski, Y. M. 1938: Kolonial’noe gnezdov’e kair i faktory vyzyvayushchie gibel’ yaits i ptentsov (Colony nesting of the murre and factors causing the death of the eggs and nestlings). Zoologicheskii zhurnalI7, 695-705. Kolthoff, G. 1903: Bidrag ti1 k h e d o m om norra polamakter- nas daggdjur och foglar. Kgl. Sv. Vet. Akad. Handl. 36. Stockholm. 104 pp. Krasnov. Y. V. 1995: Ekologiya i povedenie morskikh ptits nu sovremennom etape expluatatsii resursov Barentseva niorj (The ecology and behaviour of seabirds during the recent stage of exploitation of Barents Sea resources). Dissertatsia na soiskanie uchenoi stepeni doktora biologicheskikh nauk. 369 pp. Krasnov, Y. V. & Barrett, R. T. 1995: Large-scale interactions among seabirds, their prey and humans in the southern Barents Sea. Pp. 443-456 in Skjoldal, H. R., Hopkins, C., Erikstad, K. E. & Leinaas, H. P. (eds.): Ecology of Fjords and Coastal Waters. Elsevier Science B.V., Amsterdam. Krasnov, Y. V., Matishov, G. G., Galaktionov, K. V. & Savinova, T. N. 1995: Morskie kolonial’nye ptitsy Murmana ( M m a n ’ s colonial seabirds). St. Petersburg “Nauka” publ. 224 pp. (In Russian). Krasovskii, S. K. 1937: Etyudy PO biologii tolstoklyuvoi kairy (Studies of the biology of the thick-billed murre). Trudy Arkticheskogo instituta, Biologiya 7 7 , 32-92. Lydersen, C., Gjertz, I. & Weslawski, J. M. 1989: Stomach contents of autumn-feeding vertebrates from Hornsund, Svalbard. Polar Rec. 25, 107-1 14. Ldnne, 0. J . & Gabrielsen, G. W. 1992: Summer diet of seabirds feeding i n sea-ice-covered waters near Svalbard. Polar Biol. 1 2 . 685-692. Lbvenskiold, H. L. 1964: Avifauna Svalbardensis. Norsk Polarinst. Skr. No. 129, 1 4 5 5 . Mehlum, F. & Gabrielsen, G. W. 1993: The diet of high-arctic seabirds i n coastal and ice-covered, pelagic areas near the Svalbard archipelago. Polar Res. 12, 1-20. Mehlum, F. & Gabrielsen, G. W. 1995: Energy expenditure and food consumption by seabird populations in the Barents Sea region. Pp. 4 5 7 4 7 0 in Skjoldal, H. R., Hopkins, C., Erikstad, K. E. & Leinaas, H. P. (eds.): Ecology of Fjords and Coastal Waters. Elsevier Science B.V., Amsterdam. Mehlum, F. & Gjertz, I. 1984: Feeding ecology of seabirds in the Svalbard area - a preliminary report. N o w . Polar Inst. Rapp. Ser. No. 16. 1 4 1 . Mehlum, F., Hunt. G. L., Klusek, Z., Decker, M. B. & Nordlund. N. 1996: The importance of aggregations to the distribution of Briinnich’s guillemots in Storfjorden, Svalbard. Polar Biol. 16, 537-547. Mehlum, F., Nordlund, N. & Isaksen, K. In press: The importance of the Polar Front as a foraging habitat for guillemots Uria spp. breeding at Bjbrnbya, Barents Sea. J . Mar. Systems. Montague, F. A. 1926: Notes from Spitsbergen. Ibis 68, 136151. Munsterhjelm, L. 191 1: Beobachtungen w&end einer ornitho- logischen Studiereise nach dem Nordpolmeer und Spitz- bergen im Sommer 1910. Ofversikf af Finska Vet.-Soc. F6rh. Ed. 53, Afd. A, No. 20, 1 4 2 . 84 R. T. Barrett et al. Nathorst, A. G. 1900: TvaS somrar i Ishafet. Vul. 1 . Stockholm. 352 pp. Nikolaeva, N. G., Krasnov, Y. V. & Barrett, R. T. 1996. Movements of Common Uria aalge and Briinnich’s Guillemots U. lomvia breeding in the southern Barents Sea. Fauna now. Ser. C, Cinclus 19, 9-20. Nogales, M., Zonfrillo, B. & Monaghan, P. 1995: Diets of adult and chick Herring Gulls Larus argentatus on Ailsa Craig, south-west Scotland. Seabird 17, 56-63. Ozhigin, V. K. & Luka, G. I. 1984: Some peculiarities of capelin migrations depending on thermal conditions in the Barents Sea. Pp. 135-147 in GjWeter, H. (ed.): Pruc. Sov.- NOW. Symp. on Barents Sea Capelin, Bergen. Roi, 0. le 1911: Spezieller Teil. In Koenig, K. (ed.): Avifauna Spitzbergensis. Furschungreisen nach der Baren-Insel und dem Spitzbergen-Archipel, mit ihrem faunistischen und floristischen Ergebnisse. Bonn. 294 pp. Romer, F. & Schaudin, F. 1900 Fauna Arctica. Bd. 1. Jena. Sakshaug, E., Bjorge, A., Gulliksen, B., Loeng, H. & Mehlum, F. 1994 Structure, biomass distribution, and energetics of the pelagic ecosystem in the Barents Sea: A synopsis. Polar Biol. 14, 405411. Sergeant, D. E. 1951: Ecological relationships of the guillemots Uria aalge and Uria lomvia. Pruc. X Int. Om. Congr. (Uppsala, 1950). 578-587. Strann, K.-B., Vader, W. & Barrett, R. T. 1991: Auk mortality in fishing nets in north Norway. Seabird 13, 22-29. Strcim, H., (Zliene, I. J., Opheim, J., Kuznetsov, E. A. & Khakhin, G. V. 1994: Seabird censuses on Novaya Zemlya 1994. NOF Rapportserie, Rupp. nr. 2-1994, 1-38. Swenander, G. 1900 Beitrage zur Fauna der B a e n Insel. Bihang till Kgl. Sv. Vet. Akad. Bihang 26. Afd. IV, 1-50. Trevor-Battye, A. 1897: The birds of Spitsbergen, as at present determined. Ibis 39, 574-600. Tschanz, B. & Barth, E. K. 1978: Svingninger i lomvibestanden p i Vedoy p i Rost. Fauna 31, 205-219. Tuck, L. M. 1960: The Murres: their distribution, populations and biology - a study of the genus Uria. Can. Wildlife Munogr. Ser. No. 1, Ottawa. 260 pp. Uspenski, S. M. 1956 The Bird Bazaars of Novuya Zemlya. CWS translation of Russian game Reports, Vol. 4, Ottawa. 159 pp. (Translated from Russian, 1958). Vader, W., Barrett, R. T., Erikstad, K. E. & Strann, K.-B. 1990 Differential responses of common and thick-billed mums to a crash in the capelin stock in the southern Barents Sea. Stud. Avian Biol. 14, 175-180. Walter, A. 1890 Omithologische Ergebnisse der von der Bremer Geographischer Gesellschaft im Jahre 1889 veran- staltete Reise nach Ostspitzbergen. J. Om. 38. Weslawski, J. M., Stempniewicz, L. & Galaktionov, K. 1994: Summer diet of seabirds from the Frans Josef Land archipelago, Russian Arctic. Polar Res. 13, 173-181. Wiborg, K. F. 1979: Gunatusfabricii (Lichtenstein), en mulig fiskeriressurs i Norskehavet. Fisken Hav. 1979 ( 1 ) . 3 3 4 6 .