Feeding by grey seals in the Gulf of St . Lawrence and around Newfoundland

Diet composition of grey seals in the Gulf of St. Lawrence (Gulf) and around the coast of Newfoundland, Canada, was examined using identification of otoliths recovered from digestive tracts. Prey were recovered from 632 animals. Twenty-nine different prey taxa were identified. Grey seals sampled in the northern Gulf of St. Lawrence fed mainly on capelin, mackerel, wolffish and lumpfish during the spring, but consumed more cod, sandlance and winter flounder during late summer. Overall, the southern Gulf diet was more diverse, with sandlance, Atlantic cod, cunner, white hake and Atlantic herring dominating the diet. Capelin and winter flounder were the dominant prey in grey seals sampled from the east coast of Newfoundland, while Atlantic cod, flatfish and capelin were the most important prey from the south coast. Animals consumed prey with an average length of 20.4 cm (Range 4.2-99.2 cm). Capelin were the shortest prey (Mean = 13.9 cm, SE = 0.08, N = 1126), while wolffish were the longest with the largest fish having an estimated length of 99.2 cm (Mean = 59.4, SE = 2.8, N = 63). In the early 1990s most cod fisheries in Atlantic Canada were closed because of the collapse of the stocks. Since then they have shown limited sign of recovery. Diet samples from the west coast of Newfoundland indicate a decline in the contribution of cod to the diet from the pre-collapse to the postcollapse period, while samples from the southern Gulf indicate little change in the contribution of cod. Hammill, M.O., Stenson, G.B., Proust, F., Carter, P. and McKinnon, D. 2007. Feeding by grey seals in the Gulf of St. Lawrence and around Newfoundland. NAMMCO Sci. Publ. 6:135-152.


INTRODUCTION
Marine mammals are often considered as important consumers because of their large size and abundance, which may lead to their having an important influence on the structure and function of marine ecosystems (Bowen 1997, Savenkoff et al. 2004a, Morissette et al. 2006).Furthermore, the consumption of commercially valuable species, by marine mammals, has the potential to reduce their commercial yield (Bogstad et al. 1997) and in areas where commercial stocks are quite low, predation by seals may slow their recovery (Bundy 2001, Chouinard et al. 2005).Evaluating this impact is complex because information is needed on both predator and prey populations as well as the functional relationships between them (Hammill andStenson 2000, Yodzis 1994).
Feeding is a major link between organisms and their environment and an understanding of feeding ecology of the various species is needed to understand ecosystem dynamics (Härkönen and Heide-Jørgensen 1991, Savenkoff et al. 2004a, Yodzis 1994).However, obtaining quantitative diet information is not a trivial problem.Several studies have shown strong temporal and spatial patterns in diet composition and the uncertainties associated with diet composition are among the most sensitive variables affecting estimates of consumption (Benoit and Bowen 1990a,b;Bowen et al. 1993, Beck et al. 1993, Bowen and Harrison 1994, Lawson et al. 1995, Shelton et al. 1997).
Studies of diet of marine mammals have been conducted over several decades using a variety of techniques (Andersen et al. 2004), which have included the use of serological methods, stable isotopes, fatty acid profiles and identification of hard parts from gastro-intestinal contents or scats (Olesiuk et al. 1990;Pierce et al. 1991Pierce et al. , 1993;;Lesage et al. 2001;Iverson et al. 2004;Hammill et al. 2005).The most frequently used techniques have involved the recovery, identification and measurement of hard parts from scats at haul-out sites or from digestive tract contents from harvested animals (Murie and Lavigne 1985, 1986, 1992;Jobling 1987;Jobling and Breiby 1986;Hammond et al. 1994;Olesiuk et al. 1990;Bowen et al. 1993;Bowen and Harrison 1994).All methods have certain limitations and associated biases with the result that no single method appears to be ideal (Jobling and Breiby 1986, Jobling 1987, Lesage et al. 2001, Bowen 2000, Hammond and Rothery 1996, Grahl-Nielsen et al. 2004, Thiemann et al. 2004, Hammill et al. 2005).
The Northwest Atlantic grey seal (Halichoerus grypus) occurs along the Atlantic seaboard from the northeastern United States to the northern tip of Labrador, and throughout the Gulf of St. Lawrence (Mansfield andBeck 1977, Lesage andHammill 2001).Although rare early in this century (Lavigueur and Hammill 1993), their numbers have increased from around 30,300 animals in 1970 to approximately 246,500 animals in 2000 (Hammill et al. 2007).Early studies into grey seal diet composition were largely qualitative (Fisher and McKenzie 1955;Mansfield and Beck 1977;Benoît and Bowen 1990a), but in recent years much effort has been directed towards obtaining quantitative information in order to obtain a better understanding of variability in diet composition (Benoît and Bowen 1990b;Murie and Lavigne 1992;Bowen et al. 1993;Bowen and Harrison 1994).Some quantitative information data are available for the northern Gulf (Benoit and Bowen 1990b; Murie and Lavigne 1992), but little information was available for the southern Gulf of St. Lawrence and waters surrounding Newfoundland.
Here we examine diet composition of grey seals collected from Anticosti Island area in the northern Gulf of St. Lawrence, the coasts around Newfoundland and the southern Gulf coastal areas of New Brunswick, Prince Edward Island and Nova Scotia between 1985 and 2004.

MATERIALS AND METHODS
Stomach and intestinal contents were obtained by Department of Fisheries and Oceans employees or from contract hunters as part of normal programs to monitor pinniped diets.Animals were sampled in the Gulf of St. Lawrence (Gulf), the south coast of Newfoundland and the east coasts of Newfoundland and Labrador (Fig. 1).Stomachs and digestive tracts were removed in the field, and frozen at -20°C, until analysis.Stomachs were opened along the external curve and were classified as "containing food" if food remnants, including otoliths and other hard parts, were present.Contents were weighed sorted using a sieve with 0.45 mm mesh.Invertebrates were identified to order; cephalopods were identified using beak identification guides (Clarke 1986).Fish were identified using otoliths and less frequently from whole fish found in the stomach.
Scientific names for all identified species are listed in Appendix 1. Otoliths and other hard parts were sorted manually and conserved dry for later identification.Fish were identified to species when possible, using reference collections (Fisheries & Oceans Canada, Mont-Joli, Québec) and an identification guide (Härkönen 1986).The number of fish in each stomach was determined by pairing left and right otoliths and counting the total number of paired and unique specimens.
Otoliths were sorted, visually, into 3 different classes depending on their degradation state: class D1, including perfectly conserved otoliths (generally found in intact skulls or whole fish in seal stomach); class D2, otoliths with very few degradation marks, but margins showing some signs of erosion; class D3, very eroded otoliths, with dorsal and ventral margins and internal and external areas showing advanced digestion marks.Only D1 and D2 otoliths were used to determine total fish length.If a large number of otoliths of a single species were present in a stomach, a random subsample of 30 otoliths was measured.
Otolith-fish metric and energy density relationships were developed from samples collected during Department of Fisheries and Oceans research missions in the Gulf of St. Lawrence and off the east coast of Newfoundland, or using values from the literature (e.g.Härkönen 1986, Lawson et al. 1995, Proust 1996, D. Chabot, Dept. of Fisheries and Oceans, Mont-Joli, QC, unpublished data)(Appendix A).Otoliths not measured were identified to species and it was assumed that their weight and energy density were equivalent to the mean size and energy density of the measured otoliths for that species in the sample.Otoliths that could not be identified to species were assumed to have size and energy density equivalent to the mean of all measured otoliths.In the case of invertebrates, total mass and energy contribution were determined by multiplying the number of identified individuals by the mean weight and species energy density.In some cases, only eyes or telson were present.The contribution of this material to the diet was determined by multiplying the number of individuals determined from the number of eyes and telson times a mean mass and a mean energy density using all identified invertebrates.Diets were reconstructed for each seal, using the seal as the sampling unit.
Diet composition is expressed as follows: Frequency of occurrence i (FO i ) =(S i /S t ) •100, where S i is the number of stomachs containing species i and S t is the total number of stomachs; Numerical abundance i (NA i ) =(N i /N t ) •100 , where N i is the number of individuals of species i and N t is the total number of individuals of all prey.% wet weight= (w i /w t ) •100, where w i is the reconstructed weight of species i in a digestive tract, and w t is reconstructed weight of all prey found in an individual digestive tract.% gross energy= (e i /e t ) •100, where e i is the reconstructed energy content (kj/g) of species i in a digestive tract, and e t is reconstructed energy content (kj/g) of all prey found in an individual digestive tract.
Diet diversity was examined using species richness and calculating a Shannon index (H').Species richness is the number of different species in the sample collection.The Shannon index is a measure of species diversity, taking into account the number of individuals examined, and was calculated using: H' = -Σ{ pi*log(pi)}, where pi is the propor-tion of species x in the sample (Legendre and Legendre 1998).
Owing to the small sample sizes and individual variation, standard deviations around the means were expected to be quite large.To reduce this variability, simulated data sets of total energy and total mass consumed were created using a bootstrapping technique (Hammill et al. 2005;Resampling Stats, Arlington VA, USA 1999).
Each digestive tract was treated as a unit for resampling purposes.This process was repeated 1000 times to generate estimates of total mass and total energy, from which proportions contributed by each prey group were calculated.Differences in energy density between regions were examined by ANOVA, followed by Tukey's least squared difference test using SAS (SAS Institute Inc., Cary, NC).

RESULTS
A total of 1,118 animals were collected between 1985 and 2004.The largest sample was obtained from Anticosti Island in 1988 and 1992 (N = 506), followed by the southern Gulf of St. Lawrence, between 1994 and2003 (N = 414) and the coast around Newfoundland between 1985 and2004 (N = 198).Prey were recovered from 632 animals.Only stomach contents were examined in samples from Anticosti Island, and Newfoundland.Complete digestive tracts were examined in samples from the southern Gulf.
In samples obtained from both Anticosti Island and Newfoundland, a greater number of stomachs obtained during May-July contained food than stomachs obtained during August to October (Table 1) (χ² Anticosti = 121, df = 1; χ² Newfoundland = 17.2, df = 2 (χ², P<0.05).Among samples collected between May and July, 114 of 256 (45%) stomachs contained prey remains.Overall, a larger proportion of samples obtained from the southern Gulf, contained prey remains because the complete digestive tract, not just the stomach, was examined for identifiable prey.With the exception of animals collected in late November and December, and in February, 68-100% of the digestive tracts contained some identifiable prey (Table 1).
No difference was observed between samples in reconstructed prey mass recovered from stomachs or complete digestive tracts.Significant differences in energy density (kJ/g) were observed between the Anticosti Island, the southern Gulf and Newfoundland samples (ANOVA: F 2,623 = 39.6,P<0.0001).The energy density of diets from Anticosti Island samples (Mean = 6.33 kJ/g, SD = 3.10, N = 183) was significantly higher than the energy density of diets from the southern Gulf (Mean = 5.22 kJ/g, SD = 1.18,N = 323) and from Newfoundland (Mean = 4.42 kJ/g, SD = 1.49,N = 120), while the energy density of the southern Gulf diet was significantly  1).
Twenty-nine different prey or taxa, including 8 invertebrate taxa were identified in the food containing stomachs/digestive tracts.A wider range of prey species was consumed by grey seals from the southern Gulf and the south coast of Newfoundland (species richness = 9-35, Shannon index = 1.46-2.43),compared to grey seals collected from Anticosti Island and east coast of Newfoundland (Species richness = 8-18, Shannon index = 0.28-1.71)(Table2).At Anticosti Island, Atlantic cod, capelin, mackerel, herring, wolffish, and lumpfish dominated the diet in terms of weight and energy, while sand lance was also important in terms of frequency of occurrence or relative abundance (Tables 3, 4).Noticeable differences were observed between the May-July 1988 sample, and the August-September 1992 sample.In the former, lumpfish, capelin, wolffish and mackerel were the 4 most important prey, account-ing for almost 90% of the diet (Table 3).In the August-September 1992 Anticosti Island sample, Atlantic cod, wolffish, mackerel, winter flounder, and herring were the 5 most important prey species accounting for over 90% of the diet by weight and energy.Capelin and sand lance were not important in terms of mass or energy, but were important in frequency of occurrence and numbers of individuals (Table 4).
In the southern Gulf, cunner, white hake, sand lance, Atlantic cod, and herring, were the 5 most important prey accounting for 68% of the diet by weight and 74% of energy (Table 5).In the southern Gulf, cunner was the most important prey in February accounting for 88.7% of the diet by weight.During May-July, a more diverse diet was consumed, with sand lance, pleuronectids (flatfish), cunner, Atlantic cod, and Atlantic herring, accounting for 81.4% of the diet by weight (Table 6).Little change in diet composition was observed between May-July and August-October.The importance of Atlantic cod to the diet increased slightly from 12.8% to 17.3% and Atlantic herring increased from 10.9% to 13.3%.Little change was observed in the contribution of cunner, while the contribution by sand lance declined slightly (Table 6).Considerable interannual variation occurred in sampling effort.To compare between years, only samples collected in September and October were examined (Table 7).The contribution of different prey varied among the 7 years.Using only prey that made a 5% or greater contribution to the diet, winter flounder was an important prey in all years, followed by sandlance in 5/7 years, herring 4/7 years, cod for 3/7 years, cunner for 2/7 years and hake in 1 year only (Table 7).
Samples from Newfoundland were separated by coast, but were not analysed on a seasonal basis because of small sample sizes.A total of 25 foodcontaining stomachs were obtained from the east coast of Newfoundland and southern Labrador.
Most of these samples (76%) were obtained between August and October.Capelin, gadoids and winter flounder were the most important prey by weight and energy in this area (Table 8).Along the south coast of Newfoundland, 88% of the May and July.Atlantic cod, capelin, Gadus spp., pleuronectids and herring were the most important prey by weight and energy, while sand lance and shrimp were also important in terms of relative abundance (Table 9).Seventy-eight samples from the west coast of Newfoundland, were collected primarily between April and July (81%).Atlantic cod, Gadus spp., winter flounder, sand lance, lumpfish and mackerel were the most im- portant prey, accounting for 80% of the diet by weight and 84% by energy.(Table 10).Capelin, Atlantic herring, mackerel, shrimp and smelt were also important using relative abundance.
The smallest fish consumed were capelin with a mean length of 13.9 cm (SE = 0.08; N = 1126, Range 4.2-20.2).The largest fish consumed were wolffish (Anarhichus spp.), with a mean length of 59.4 cm (SE = 2.8, N = 63; Range 4.8-99.2cm).Some regional differences in prey size were observed.The mean length of Atlantic cod consumed in the northern Gulf, which included Anticosti Island and western Newfoundland, was significantly greater (37.9 cm, SE = 0.

DISCUSSION
Major limitations to the hard part / reconstruction approach to quantify diet composition include the failure to find hard parts in the sample and under-estimating hard part size due to erosion while in the stomach (Jobling and Breiby 1986, Tollit et al. 2003, Christiansen et al. 2005).The degree to which these problems occur is affected by foraging behaviour, species composition of the diet, activity levels of the animal and meal size (Murie and Lavigne 1985, Jobling and Breiby 1986, Jobling 1987, Lawson et al. 1995, Tollit et al. 1997, Marcus et al. 1998).The impact of variability in otolith erosion rates, including complete otolith digestion on diet reconstructions, has been examined in captive studies and some solutions have been proposed (Tollit et al. 1997(Tollit et al. , 2003)).We did not measure eroded otoliths because suggested correction factors to adjust otolith lengths to account for partial digestion are quite variable (reviewed by Bowen 2000) and when this variability is taken into account, considerable uncertainty to estimates of diet composition is added (Hammond and Rothery 1996).
In the Northwest Atlantic grey seal population reproduction occurs from late December until mid-February (Mansfield and Beck 1977).
Little feeding occurs at that time.After moulting, animals forage intensively from the early spring (April-June) until July, after which a decrease in foraging bouts are observed until early autumn (October), when foraging activity again intensifies (Beck et al. 2003).The greater number of food-containing stomachs in samples obtained during the spring compared to late summer samples reflects this seasonal change.
Little difference was observed between regions and seasons in estimated mean meal size, but samples obtained from Anticosti Island had a higher energy density than samples obtained from other regions, reflecting in part the importance of high energy species such as capelin and mackerel in the diet of grey seals from this area.Grey seals are primarily piscivorous, with invertebrates accounting for only a very small fraction of their diet (Benoît and Bowen 1990a,b;Murie and Lavigne 1992;Bowen et al. 1993).
Although a wide range of species were consumed, only about 3 to 5 species accounted for over 80% of the diet of grey seals in the northern Gulf of St. Lawrence and waters around Newfoundland.Grey seals from the southern Gulf of St. Lawrence had a more diverse diet, with up to 8 species accounting for about 80% of the diet composition.Major prey items included cunner, white hake, winter flounder, herring, cod, capelin, lumpfish and sand lance, which have also been reported as important prey elsewhere (Benoit and Bowen 1990a,b;Bowen et al. 1993;Bowen and Harrison 1994).We also identified wolffish as an important prey species in samples from the northern Gulf.Diet composition of samples from the southern Gulf of St. Lawrence had a higher species richness and Shannon Index than diets from the northern Gulf of St. Lawrence and the coasts of Newfoundland, which likely reflect ecosystem differences between the 2 regions.The northern Gulf is characterized by 2 deep channels (Laurentian and Esquiman), with an average depth of 420 m.Zooplankton biomass is high, while species diversity is low (De Lafontaine et al. 1991).Capelin, redfish, cod and sand lance are important species in the fish community (Savenkoff et al. 2004a).In contrast, the southern Gulf is characterized by a large relatively shallow area with an average depth of 50 m, called the Magdalen Shallows (De La-fontaine et al. 1991).Compared to the northern Gulf, zooplankton biomass is lower, but diversity is higher (De Lafontaine et al. 1991).Major species in the fish community include cod, white hake, mackerel, herring, shanny, cunner and flatfish such as flounders (Savenkoff et al. 2004b).
Substantial seasonal and inter-annual variation in the contribution of different prey was also observed.In the northern Gulf, lumpfish, capelin, mackerel and wolffish were the dominant prey in the early summer diet.The importance of capelin and lumpfish was also reported by Benoit and Bowen (1990b) and probably reflects the concentration and inshore movement of these species to spawn (Jangaard 1974, Scott andScott 1988).Later in the season, cod, herring, mackerel and wolffish become dominant prey, which with the exception of wolffish, is similar to what was observed by Benoit and Bowen (1990b).This seasonal change also points to a change in diet from energy rich species consumed in early summer when foraging activity is more intensive to a less energy rich prey during the fall (Benoit andBowen 1990b, Beck et al. 2003).However, similar changes in composition were not observed in the southern Gulf.In that area, sandlance, Atlantic herring, Atlantic cod, and white hake were the 4 most important prey species in both early summer and late summer-fall periods.
Grey seals consumed prey with a mean length of 20.4 cm.Capelin were the smallest prey consumed (mean = 13.9 cm), while wolffish were the largest (mean = 59.4 cm).Overall, mean prey size was similar to what has been reported elsewhere in Atlantic Canada and the Northeast Atlantic (Bowen et al. 1993, Hammond et al. 1994, Mikkelsen et al. 2002).The largest fish consumed, a wolffish with an estimated length of 99 cm, was longer than any fish previously reported consumed by grey seals in Atlantic Canada.Although the majority of fish were less than 35 cm long, predation on larger fish by grey seals has been documented elsewhere, particularly wolffish (catfish in NE Atlantic, Anarhichas lupus; Mikkelsen et al. 2002).
A 13.6 kg wolffish was found beside a ringed seal hauled out on the ice in Arctic Canada, and saithe (Pollachius virens L.) and ling (Molva molva L.) with estimated lengths of around 90 cm have been reported for grey seal diets in the United Kingdom (Hammond et al. 1994, Smith 1977).At the same time, detection of otoliths from such large wolffish in the stomachs of grey seals does not guarantee that the entire fish was consumed.There are some indications that seals may only consume soft parts of fish, leaving the heads behind (Lunneryd 2001).Perhaps in instances where seals attempt to take potentially aggressive prey such as the wolffish, seals might only have consumed the head and little else.
In the early 1990s, a moratorium on fishing for Atlantic cod was declared after several eastern Canadian cod fisheries had collapsed.Almost a decade later, evidence of marked changes in ecosystem structure are still evident, with almost all of these stocks showing no or very limited signs of recovery (Rice and Rivard 2003).In samples examined during 1986and 1987, by Benoit and Bowen (1990b), just prior to the collapse, Atlantic cod made up 41.5% of the diet by weight which is similar to the 46.4% we observed in samples obtained at about the time of the collapse.Unfortunately no data are available from the Anticosti Island region since the collapse of the northern Gulf cod stock.Although samples are small, data from the west coast of Newfoundland suggest that the contribution of cod to the diet has declined from 32.5% (SD = 14.5, N = 11) between 1988 and 1992, just prior to the collapse of the cod, to 14.6% (SD = 4.7, N = 51) in samples obtained between 1995 and 2004, well after the stock had collapsed.In the southern Gulf, Benoit and Bowen (1990a) examined diet samples obtained during the mid-1980s, prior to the collapse.Expressed as frequency of occurrence, cod accounted for 13.5% of the diet, which is very similar to our 16.7%, suggesting that, in spite of major changes in cod biomass, there has been little change in the contribution of cod to the diet of grey seals in the southern Gulf of St. Lawrence in the areas we sampled.The contrasting inferences between the pre and post collapse diets from western Newfoundland, and the southern Gulf of St. Lawrence point to the dilemma in understanding foraging patterns in marine mammals.On the one hand, larger spatial scale commercial fish surveys point to the decline in cod biomass, while samples from individual grey seals reflect local prey choice which will be influenced by a much smaller spatial scale of local prey abundance, energy value and energy cost to obtain that prey.

Fig. 1 .
Fig. 1.Map showing region where grey seals were collected.The dots represent locations where animals were collected.

Fig
Fig. 2. Frequency distribution (%) of estimated lengths (cm) of prey consumed by grey seals in the Gulf ofSt.Lawrence and  around Newfoundland  between 1985 and  2004.

Table 1 .
Number of animals collected (N), number of animals with food in digestive tracts (Ns), average reconstructed mass (g) in tract and average energy density (kj/g), with standard deviations in parentheses.Samples from Newfoundland, the northern Gulf and southern Gulf(Feb  2000, Nov-Dec 2003)examined stomachs only.The remaining samples from the southern Gulf examined complete digestive tracts.

Table 2 .
Species richness and Shannon indices for grey seal diet samples from Newfoundland, Anticosti Island and the southern Gulf.

Table 4 .
Diet composition of 69 seals collected at Anticosti Island August-September 1992.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parentheses, and percent (%) mass and energy contribution to the diet with standard deviation in parentheses.

Table 3 .
Diet composition of 114 seals collected at Anticosti Island during May-July 1988.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parentheses, percent (%) mass and energy contribution to the diet with standard deviation in parentheses.

Table 5 .
Diet composition of 322 seals collected from the southern Gulf of St. Lawrence between 1994 and 2003.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parentheses, and percent (%) mass and energy contribution to the diet with standard deviation in parentheses.

Table 6 .
Percent mass contribution of different prey by season to grey seal diets from animals collected in the Southern Gulf of St.Lawrence between 1994 and 2003.Samples include all years combined.Average with standard deviation in parentheses.

Table 7 .
Annual changes in diet composition (percent mass) of grey seals collected in the southern Gulf of St. Lawrence between 1995 and 2003.Only samples collected between September and October included in table.Average with standard deviation in parentheses.Latin binomial names are as in Table6.

Table 8 .
Diet composition of 25 seals collected from the east coast of Newfoundland and southeastern coast of Labrador between 1985 and 2004.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parentheses, and percent (%) mass and energy contribution to the diet, average with standard deviation in parentheses.

Table 9 .
Diet composition of 24 seals collected from the south coast of Newfoundland between 1985 and 2004.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parenthesis, and percent (%) mass and energy contribution to the diet with average and standard deviation in parentheses.

Table 10 .
Diet composition of 78 seals collected from the west coast of Newfoundland between 1985 and 2004.Frequency of occurrence and % frequency of occurrence in parentheses, numerical abundance with relative percent in parenthesis, and percent (%) mass and energy contribution to the diet with average and standard deviation in parentheses.

Table 11 .
Number of measured otoliths (N), average prey length (cm), with standard deviation in parentheses and range of prey lengths (cm) consumed by grey seals in samples collected from the Gulf of St. Lawrence and around Newfoundland between 1985 and 2004.Latin binomial names as in Table6.