Diet of the ringed seal ( Phoca hispida ) in Greenland

Analysis of 454 stomachs of ringed seals (Phoca hispida) collected from the Inuit hunt in six municipalities in West Greenland and 30 alimentary tracts collected by scientists in East Greenland, showed seasonal and regional differences in the diet. In Northwest and East Greenland polar cod (Boreogadus saida) and Arctic cod (Arctogadus glacialis) were the most dominant prey items. In contrast, seals in central West Greenland mainly preyed upon amphipods (Parathemisto spp.), capelin (Mallotus villosus), redfish (Sebastes sp.) and squid (Gonatus sp.), while capelin was the most important prey item in Southwest Greenland. The material from Uummannaq demonstrates seasonal variations, as euphausiids were common in spring, and snailfish (Liparis spp.) dominated from October through January, where after snailfish disappeared with the formation of fast ice. No age related differences in the diet were found in stomach samples from Avanersuaq, but in Upemavik crustaceans were more abundant in stomach samples from younger immature seals. Greenland halibut (Reinhardtius hippoglossoides) was only present in samples from seals older than 4 years. In Upemavik Arctic cod was the dominant prey item for seals caught in May at the ice edge, whereas polar cod dominated the samples from seals caught in open water. INTRODUCTION T he ringed seal (Phoca hispida) is a diverse feeder with a diet ranging from pelagic amphipods to Arctic gadoids (i.e. polar cod (Boreogadus saida) and Arctic cod (Arctogadus glacialis» and more temperate species like capelin (Mallotus villosus) and redfish (Sebastes spp.). The ringed seal takes planktonic, nektonic or benthic food and generally turns from inshore feeding on the epibenthic mysids, shrimps, other crustaceans or polar cod, to strictly macroplanktonic feeding in deeper, offshore waters (Frost and Lowry 1981a). Ringed seals sampled in drifting ice had stomach contents consisting of pelagic amphipods, while polar cod constituted an important prey species for seals in coastal areas (Johansen 1910). Pedersen (1931) found that mature ringed seals mostly appeared in coastal areas on the East Coast of Greenland, whereas younger seals mostly were found in open water. NAMMCO Scientific Publications, Volume I The most important prey item of mature seals was crustaceans, while the major prey item of younger seals was polar cod. An analysis of ringed seal stomach contents showed that the ringed seal in Avanersuaq municipality basically predated on polar cod (Vibe 1950). Ringed seals collected at Baffin Island in the summer periods in 1938 and 1940 were mainly feeding on Parathemisto libellula and Dunbar (1941) concluded that the ringed seals were planktonic feeders. McLaren (1958) studied ringed seal diet in southwestern Baffin Island, and P libellula , Mysis oculata and polar cod were found to be the most important prey items. He concluded that ringed seals were opportunistic feeders and he found no food preferences related to age. Several studies have documented seasonal cycles of feeding habits (e.g. McLaren 1958, Johnson et a1.1966, Smith, 1987, Weslawski et


INTRODUCTION
T he ringed seal (Phoca hispida) is a diverse feeder with a diet ranging from pelagic amphipods to Arctic gadoids (i.e.polar cod (Boreogadus saida) and Arctic cod (Arctogadus glacialis» and more temperate species like capelin (Mallotus villosus) and redfish (Sebastes spp.).The ringed seal takes planktonic, nektonic or benthic food and generally turns from inshore feeding on the epibenthic mysids, shrimps, other crustaceans or polar cod, to strictly macroplanktonic feeding in deeper, offshore waters (Frost and Lowry 1981a).Ringed seals sampled in drifting ice had stomach contents consisting of pelagic amphipods, while polar cod constituted an important prey species for seals in coastal areas (Johansen 1910).Pedersen (1931) found that mature ringed seals mostly appeared in coastal areas on the East Coast of Greenland, whereas younger seals mostly were found in open water.

NAMMCO Scientific Publications, Volume I
The most important prey item of mature seals was crustaceans, while the major prey item of younger seals was polar cod.An analysis of ringed seal stomach contents showed that the ringed seal in Avanersuaq municipality basically predated on polar cod (Vibe 1950).Ringed seals collected at Baffin Island in the summer periods in 1938 and 1940 were mainly feeding on Parathemisto libellula and Dunbar (1941) concluded that the ringed seals were planktonic feeders.McLaren (1958) studied ringed seal diet in southwestern Baffin Island, and P libellula , Mysis oculata and polar cod were found to be the most important prey items.He concluded that ringed seals were opportunistic feeders and he found no food preferences related to age.Several studies have documented seasonal cycles of feeding habits (e.g.McLaren 1958, Johnson et a1.1966, Smith, 1987, Weslawski et al. 1994).
In the northern part of Greenland the sea is covered by continuous fast ice in late winter.Ringed seals maintain breathing holes or seek cracks or openings in the fast ice.Hunting methods for ringed seals vary according to season (Kapel 1975, Teilmann and Kapel this volume) .During late winter to early spring seals are caught in nets under the ice.In spring they are shot in open leads at the ice edge, or when they haul out on the ice (a hauled-out seal is called an uuttoq in Greenlandic, uuttut in plural).The break-up of the sea ice in late spring is the beginning of the open water hunting season, which lasts until late autumn.During this peri- Earlier studies of ringed seal diet in Greenlandic waters are based on fragmentary or limited material (Johansen 1910, Pedersen 1931, Vibe 1950).The intentions with the present study are to elucidate the variation in ringed seal diet in seven widely separated areas in Greenland and to examine the seasonal variation in one area.The stomach mass and contents are examined in relation to hunting method, age and season.
Ringed Seals in the North Atlantic  , 1967) was measured to the nearest cm.From each animal the lower jaw was removed and the of 254 seals was estimated by counting annular growth layers of the dentine of one canine tooth (Dietz et al. 1991).

MATERIALS AND METHODS
The remaining 200 animals were given an approximate age calculated from the correlation between the observed standard length and age (age = exp(-19.978)* stdlength 449 , r = 0.4653 , N = 254).Additional information on sex, date of capture, locality and hunting method was recorded.The ringed seal stomachs and intestines were kept frozen until analysed in the laboratory.After thawing, the content of the stomachs and intestines was removed, weighed and placed in a water bath.Undigested fish and invertebrate remains were removed from the sample and the remaining mixture of water and prey items were vibrated until otoliths and stones were deposited at the bottom of the water bath.After extracting the deposit, the rest of the contents were removed through a fine-mesh sieve.
The length of all otoliths, telsons and cephalopod beaks were measured to the nearest mm under a microscope.The ringed seal diet was determined by reconstructing the pre-ingestion size of the prey items using published otolith length to fish length and fish mass equations (Frost and Lowry 1981b, Finley and Gibb 1982, Bradstreet and Finley 1983, Harkonen 1986), NAMMCO Scientific Publications, Volume I telson plus uropod length to invertebrate mass relationship (Finley andGibb 1982, Bradstreet andFinley 1983) and squid beak length to squid mass relationship (Clarke 1962).There is no published otolith length (OL) to fish length and fish mass equations for snailfish, Liparidae.Based on information on otolith length and fish mass (FW) from 22 fresh Liparis fabricii collected in West Greenland, the otoliths from Liparis spp.were grouped in three size categories (0.4mm<OL>0.8mm,0.8mm<OL> 1.2 mm, 1.2mm<OL>2.0mm)and given an approximate mass (1g, 5g, 109).
All data were grouped by municipality, year and hunting method.The calculated mass of all invertebrate and fish remains in a stomach sample was summed to a total mass.Diet composition was expressed as frequency of occurrence and relative contribution by fresh mass.

RESULTS
The samples were collected discontinuously at different municipalities, where samples from Uummannaq were taken during nine months, while the sampling extended over two to three months at the remaining municipalities (Table 1).Of the 484 examined stomachs 149 were empty and 60 additional stomachs only contained small amounts of pebbles/stones.A total of 71 % of the stomachs from the uuttoq (fast ice) hunting in West Greenland were empty, compared with 24% in net-caught and 14 % in open-water hunted seals.A similar pattern is shown in spring (April to June) catches from 1984-1987 in Uummannaq, where a total of 71 uuttoq seals and 70 netted seals were taken.In the uuttoq seals 78 % of stomachs were empty, compared with 17% among seals caught in nets (Table 1).
Further analyses on the influence of hunting methods, species composition by region and season are based on 275 seals with recognisable prey in their stomachs.

Influence of hunting methods
Seals taken in open water generally have larger and less fluctuating masses of stomach contents compared with seals taken in other habitats (Table 1).The weight of stomach contents of seals taken in open water is by average four times larger than in seals taken as uuttut.In Uummannaq some increase in mean stomach mass is noted in late autumn followed by a decrease in January, the time when the fast ice appears (Table 1).The same could be noted in Kangaatsiaq, where mean stomach mass dropped when the ice appeared in March.

Species diversity in diet
Pooling all data from included regions indicates that the food of ringed seals in Greenland mainly comprises gadoids (polar cod and Arctic cod), capelin, redfish, pelagic crustaceans (Parathemisto spp.), benthic crustaceans (mysids) and cephalopods (Table 2, Figs 2-3).At least 21 different species were taken by ringed seals but the number is probably considerably higher because a number of species could only be determined to family, although it was obvious that more than one species was present (e.g.Zoarcidae, Cottidae and Liparidae).Exact species determination based on otoliths alone is impossible for several of these families.Furthermore, sampling over the full annual cycle in the different regions would certainly increase the number of prey species.

NAMMCO Scientific Publications, Volume I
Diet in regions Samples in Avanersuaq were taken in August and September.The primary prey items of ringed seals were hyperiid amphipods and polar cod, which made up 92.3% of all prey both by numbers and by relative contribution of mass.Most ingested specimens of polar cod were young and immature with a mean length of n.lmm (SD 40.8), which suggests that the seals foraged on shoaling polar cod from the same age class.Also in the Upemavik municipality (most samples taken in May) polar cod dominated numerically in the diet but due to larger size, Arctic cod dominated the composition of the diet by weight.Amphipods and mysids were present in large numbers, but they contributed insignificantly to the diet on a mass basis.In the samples from Upemavik both the polar and especially the Arctic cod, were larger than those found in stomachs from Avanersuaq.Also, five otoliths from Greenland halibut (Reinhardtius hippoglossoides) were found in stomachs from Upemavik.Although these halibuts were among the largest fish eaten by the ringed seals they were only present in low numbers and thus did not contribute more than 4.3 % to the mass of the diet.

Cottidae
In Uummannaq, data from 1983-1984, showed that >90% of the prey items by numbers or mass percentage were hyperiid amphjpods, euphausiids and snailfish (Liparis spp.) in 83 stomachs with contents.Although present, apparently the cod species were less important in Uummannaq in 1983-84, whereas halibut on a mass basis is the second most important fish species after Liparis spp.The importance of the cod species is re-established in the 5 stomachs from Uummannaq in 1985, whereas the 8 stomachs from 1987 on a mass basis had a preponderance of cephalopods (Gonatus fabricii).Except for Avanersuaq this cephalopod is present in stomachs from all regions.
In Qeqertarsuaq more boreal species like capelin and redfish are new but important components in the diet that also has substantial contri- bution from euphausiids.In the Kangaatsiaq municipality Gonatus fabricll again makes up the bulk of the mass of the stomach contents, but also gadoids, redfish and capelin are important prey items.The only herring (Clupea harengus) was found in a stomach from Kangaatsiaq.Since otolith width of this herring (2.5mm) was outside the range of the regressions given in Harkonen (1986) the size estimation must be considered uncertain.
Capelin followed by Gonatus fabricii was the primary prey items in the Nanortalik municipality, corresponding to >95 % by both numbers and mass of stomach content.

DISCUSSION
McLaren (1958( ), Fedoseev (1965)), Lowry et al. (1980) and Bradstreet and Finley (1983) all observed that ringed seals have a reduced food intake during the moulting period from April-May whereas feeding activity remain high throughout the winter.We noted similar seasonal and spatial variations in mass of stomach contents.In Uummannaq and Kangaatsiaq a reduced food intake is evident after the appearance of ice in January and March, respectively.However, at this time of the year hunting methods change from open water hunting to netting, which may affect the quantity of food contained in the stomachs of the seals.It is evident that stomachs collected from the open water hunting provide the largest and most stable sample of prey items.This is mainly due to the fast digestion of prey items in seal stomachs.Feeding experiments with captive seals indicate that no otoliths can be recovered in the stomach 12.9 hours after feeding (Murie and Lavigne 1986).Seals that have been hauling out on the ice for more than 12 hours can be expected to have only few remains from previous meals in their stomachs.
A large number of invertebrate (>60) and fish species (> 10) have been reported as prey items of ringed seals from the Arctic (cf.(Weslawski et al. 1994).In Greenland ringed seals apparently select 2-4 key prey items that constitute more than 90% of the diet in each area.The most common prey size is 5-lOcm for fish with a maximum size of about 20cm although prey as large as 35cm may occasionally be taken depending on age/size of the seals (see also Dunbar 1941, McLaren 1958, Smith, 1987).Ringed Seals in the North Atlantic thic prey like Cottidae, Liparidae, halibuts and perhaps some of the unidentified cod species, seem generally less important than pelagic prey that occur in schools.
Apparently ringed seals in the northern seas of the Russia, Alaska, Canada and Northwest Greenland feed heavily on polar and Arctic cod most of the year (Chapski 1940, Vibe 1950, McLaren 1958, Lowry et al. 1980, Bradstreet and Finley 1983).Polar cod occur in the water column from 0-50 metres depth, however, they are also known to occur at greater depths and annual migrations of juvenile polar cod in late summer are well documented (Craig et al. 1981, Bradstreet et al. 1986).Polar cods are reported to be the major prey item for ringed seals in northern Alaska in the autumn and winter seasons (Lowry et al. 1980).According to Lowry et al. (1980) the amount of food eaten by ringed seals decreases in late spring and summer.In the same period the volume of polar cod in the diet decreases and benthic crustaceans become the most important food item.Bradstreet and Finley (1983) found polar cod in ringed seal stomachs collected throughout the year in the Canadian high Arctic.In our study, polar cod and Arctic cod were key components in the ringed seal diet for most of the year in Northwest Greenland and in East Greenland, but their importance in the diet was replaced by more boreal species in central and Southwest Greenland.
The ringed seals from Kangaatsiaq foraged on juvenile cephalopods.Kristensen (1984)  and marine birds all forage on polar cod at the ice edge, although both polar cod and Arctic cod remains were found in the stomachs of four ringed seals shot at the ice edge (Bradstreet and Cross 1982).
The samples from Uummannaq covered all seasons except summer.Parathemisto spp., Liparis spp.and euphausiids were the most important prey items, but there was some evidence of seasonal changes in the diet.The importance of Parathemisto spp.seemed to decrease during early spring as the importance of euphausiids increased during the same period.Many species of zooplankton, including euphausiids and hyperiid amphipods, have patchy distribution.Parathemisto spp. is known to form dense swalms during breeding in early spring and they could be easy food for ringed seals (Dunbar 1941).In this study the increasing presence of euphausiids during spring suggests that they at that time are more available food for ringed seals than Parathemisto spp.
The biology of Liparis spp. is almost unknown but apparently they are present in the ringed seal diet in large numbers in late autumn and almost disappear when the ice is formed in January.Similarly not much is known about Lycodes spp.But they are obviously taken by ringed seals in East Greenland.Both Liparis spp.and Lycodes spp.are reported from ringed seals from Svalbard (Weslawski et al. 1994).Lowry et al. (1978) found no clear time-, sex-, or age-related differences in the ringed seal diet.In this study samples from Avanersuaq and from the ice edge in Upernavik showed no agerelated differences in diet.Samples from the open water catch in Upernavik, however, showed that younger seals took relatively more crustaceans than adult seals.The decline in importance of crustaceans with age is in accordance with Bradstreet and Finley's (1983) observations from Canada but the pattern was not as clear as in ringed seals collected in Alaskan waters (Lowry et al. 1980).
It appears from our study that ringed seals in West Greenland concentrate their predation on shoaling prey, which is probably also the most abundant and easily available prey species.This is also in accordance with several other studies (e.g.McLaren 1958, Fedoseev 1965).Based on the knowledge that ringed seals feed on a variety of organisms representing several trophic levels, McLaren (1958) concluded that "food is not likely a limiting factor in the local distribution and numbers of the seal."Lowry et al. (1980), however, emphasised the role of food

REFERENCES
availability in influencing the movement of ringed seals.Certainly if ringed seals preferentially seek blooms of shoaling prey and they utilise a low number of species this will determine their movements since they will have to search for concentrations of prey.

Fig. 1
Fig. 1 Map of Greenland with municipalities and areas mentioned in the text.230 Mean mass (g) of ringed seal stomach contents distributed by municipality or area, month and hunting methods.Note that in Uurnrnannaq and Kangaatsiaq the fjord ice forms in January and March, respectively.N indicates the number of examined stomachs.Municipality or

Fig. 2
Fig. 2The amphipod Parathemisto libellufa taken by pelagic trawl at the east coast of Svalbard.Photo: T Haug)

Table 2 .
Food items of ringed seals in different municipalities from West Greenland and from Kong Oscars Fjord in East Greenland.The frequency of occurrence is given as well as the relative contribution by mass and the estimated size (length) of each food item.N indicates the number of examined stomachs.
Ringed Seals in the North Atlantic

Table 3 .
Relative contribution by mass of the dominating prey items in the ringed seal diets in May and June in the Upemavik municipality distributed by three hunting methods.N indicates the number of stomachs examined.
-Ringed Seals in the North Atlantic

Table 5
In Upemavik the seals were caught at the ice edge and in open water.In samples taken from the open water the proportion of crustaceans decreased with age and larger fish species (e.g.Arctic cod and Greenland halibut) were only present in seals older than 5 years.
).In Avanersuaq seals were caught in the open water season in August and September.Arctic cod was only present in seals older than 5 years, whereas polar cod and crus-NAMMCO Scientific Publications, Volume I

Table 4 .
Monthly frequency of occurrence of the four most important prey items for the ringed seal stomach samples collected in the Uummannaq municipality in 1983 and 1984.Quantification of the relative contribution by mass for Liparis spp. is not possible; thus instead the frequency of occurrence is shown.

Table S .
Dominating prey items in terms of estimated mass in ringed seal diet in relation to age.The stomach samples were collected in the Avanersuaq municipality and from ice edge and open water catches in the Upernavik municipality.N indicates the number of examined stomachs.