2009 NORTHEASTERN NATURALIST 16(4):501–518 Composition of Eiders Harvested in Newfoundland Scott G. Gilliland1,* and Gregory J. Robertson2 Abstract - Somateria mollissima (Common Eider) is an important game species throughout its circumpolar range, including eastern Canada and northeastern United States. In eastern Canada, the largest harvest of Common Eiders occurs in Newfoundland; however, the age, sex, and subspecific composition (S. m. borealis and dresseri are both present) of this hunted population is not well quantified. The species, subspecies, age, and sex composition of the harvest was determined by examining heads collected from 1672 eiders (including Somateria spectabilis [King Eider]), taken mostly by hunters, from 1980–1996. Band-recovery information for Common Eiders banded in Newfoundland and Labrador were also summarized, including data from a release program of hand-reared ducklings in northern Newfoundland from 1988–1996. The composition of the eider harvest varied across the province. In northern and eastern areas, borealis Common Eiders made up the bulk of the harvest. King Eiders were also taken in these areas. In southern areas, most Common Eiders taken were dresseri. Sex ratios tend to be near 1:1, and immature birds comprised most of the harvest. Recovery distributions showed that Common Eiders breeding in Newfoundland and Labrador were mainly taken within the province, but also contributed to harvests in Québec and Nova Scotia, and to a lesser extent, Maine. Hand-reared ducklings had direct recovery rates of 0.047 and were most likely to be recovered within the province in the year of banding, but contributed to harvests in Atlantic Canada and New England as they aged. The harvest of eiders in Newfoundland varies regionally and seasonally; therefore specific harvest regulations could be implemented to manage the variety of populations present in the province throughout the year. Introduction Somateria mollissima L. (Common Eider) is a wide-ranging circumpolar species that has a long history of utilization by people. The largest of the sea ducks, Common Eiders have been, and continue to be, an important source of meat, eggs, and down to northern aboriginal peoples. More recently, recreational harvests have developed in the southern areas of their range and sustainable down-collection industries have been developed at some breeding areas (Bédard et al. 2008). In eastern North America, Common Eider numbers were depleted at the turn of the 20th century, and were one species specifically considered in the Migratory Bird Treaty negotiated between the United States and Canada. Unlike most ducks, Common Eiders vary across their range in size and other morphological features, leading to the recognition of 7 subspecies worldwide. In eastern North America, 2 of these sub-species are present: the 1Canadian Wildlife Service, Environment Canada, 6 Bruce Street, Mount Pearl, NL A1N 4T3, Canada. 2Wildlife Research Division, Environment Canada, 6 Bruce Street, Mount Pearl, NL A1N 4T3, Canada. *Corresponding author - scott.gilliland@ ec.gc.ca. 502 Northeastern Naturalist Vol. 16, No. 4 S. m. dresseri (American) and S. m. borealis (Northern) subspecies (Mendall 1980). The American Common Eider breeds from southern Labrador south to New England, and winters within this range, with most of the northern breeding populations of American Common Eiders moving south to some degree in winter (Goudie et al. 2000). The Northern subspecies breeds in Labrador and north throughout the eastern Canadian Arctic and Greenland; some move south ahead of the ice to Québec and Newfoundland in the nonbreeding season, while others move east to winter in Greenland (Mosbech et al. 2006). Along the coast of Labrador, an intermediate form (intergrades) of the 2 subspecies is found breeding (Mendall 1980, 1986). The coast of insular Newfoundland is an important wintering area for both Somateria spectabilis L. (King Eider) and Common Eiders (Reed and Erskine 1986), and the Common Eider is one of the most important game birds in Newfoundland and Labrador (Wendt and Silieff 1986). The size of the Common Eider population breeding in insular Newfoundland is small (7000 pairs of dresseri; S.G. Gilliland, unpubl. data); hence, most of the 100,000+ eiders that over-winter in Newfoundland waters (Gilliland et al. 2009; Canadian Wildlife Service, St. John's, NL, Canada, unpubl. data) must originate from breeding areas along the coasts of Labrador, northern Québec, and the eastern Canadian Arctic. Since 1975, when regular harvest surveys began, the harvest of Common Eiders in Newfoundland has ranged from 4500 to over 23,000 birds, averaging 11,500 birds, and is usually the largest harvest of Common Eiders of all the provinces in Canada (Gendron and Collins 2005). Due to inherent biases in the these surveys (Chardine et al. 2008, Wendt and Silieff 1986), these estimates likely underestimate the true level of harvest. Previous descriptions of eiders wintering in Newfoundland were based on anecdotal observations (see Gillespie and Learning 1974), but indicated that of the two subspecies of Common Eiders present, the borealis subspecies comprised the majority of wintering birds (Reed and Erskine 1986). Gillespie and Learning (1974) indicated that borealis were found primarily off the north, east, and southeast coasts of Newfoundland, whereas smaller numbers of dresseri wintered off the south and southwest coasts of the island. Information on the species and subspecies composition of eiders wintering in Newfoundland is important as dresseri and borealis Common Eider and King Eider populations face different environmental conditions and anthropogenic pressures (Reed and Erskine 1986). The breeding population of dresseri in Newfoundland is currently depressed but is showing signs of recovery, while borealis birds breeding in Canada and Greenland are showing different regional trends and face different harvest pressures (Chaulk et al. 2005, Christensen and Falk 2001, Gilliland et al. 2009, Merkel 2004a). It is also not possible to distinguish subspecies during aerial surveys, so estimates of subspecies composition are essential for determining the size of these wintering populations from aerial survey data. The purpose of this paper was to summarize several data sets useful for management of eiders in Newfoundland and Labrador, with the overall 2009 S.G. Gilliland and G.J. Robertson 503 goal of obtaining a better understanding of how Common Eider populations are structured and utilized in eastern North America. We compiled information from collections of eiders of both species shot by hunters, and information from recoveries of Common Eiders banded in Newfoundland and Labrador. Specifically we describe the spatial and temporal distribution of the age, sex, species, and sub-specific composition of eiders harvested in Newfoundland. The spatial and temporal distributions of band recoveries of eiders banded in Newfoundland and Labrador were analyzed to determine the contribution of the local breeding populations to the harvest. We also estimate juvenile recovery rates and adult survival rate from a data set of hand-reared pre-fledging Common Eiders released in northern Newfoundland as part of efforts to re-build local populations, and comment briefly on the contribution of these hand-reared birds to harvest. Methods Examination of harvested bird carcasses Demographic and morphological measurements from 1672 Common Eiders and King Eiders were collected in insular Newfoundland, primarily of birds taken between 1981 and 1996, with a focus during 1985–1987 (n = 972) and 1994–1995 (n = 364), while a small portion of the collections occurred before 1980 (n = 12). The data were compiled from several sources: birds collected under various scientific collection permits (n = 20), confiscations of birds that were taken illegally (n = 137), and eider heads collected from hunters (n = 1535). For the bulk of the sample that came from hunters, individuals in communities where eider hunting is common were contacted and asked to supply heads from birds they had shot. Therefore, the sample cannot be treated as randomly collected, but opportunistically collected from a representative set of communities that regularly hunt eiders. The legal eider-hunting season in Newfoundland spanned approximately 107 days from late November to 10 March in 1980–1997; in 1998, the season was reduced to approximately 78 days, ending at the end of February. Daily bag limits were 12 eiders from 1980–1990 and 6 eiders from 1990–2000 (with a further reduction to only 3 eiders in the month of February from 1997–2000). Possession limits were twice the daily bag limit. Where possible, we recorded the following information for each bird: species (King or Common Eider), sex, age of males based on plumage (adult: after second year, immature: second-year or hatch-year), and date and location of the collection. Females cannot be reliably age based on plumage alone. We also took the following complete set of bill measurements from 1182 birds: culmen midline, exposed culmen, nostril extension, and frontal extension (Mendall 1980, 1986). Data were recorded on standardized data sheets to help ensure repeatability of measurements, and the bulk of the data was collected by 3 experienced biologists familiar with eiders. Information was incomplete for some birds; hence, sample sizes vary across the various summaries of the database. 504 Northeastern Naturalist Vol. 16, No. 4 Following Mendall’s (1986) key, we used 4 bill measurements to classify Common Eiders into 1 of 4 subspecies classes: dresseri, borealis, borealis-dresseri intergrades, and “borealis or intergrade.” Because our primary interest was to determine if Common Eiders originated within the breeding range of S. m. borealis (High Arctic and Low Arctic population groups; Reed and Erskine 1986), or within the breeding range of S. m. dresseri (Atlantic population breeding group; Reed and Erskine 1986), we collapsed the three S. m. borealis subspecies classes (borealis, intergrades and “borealis or intergrades”) into a “borealis-type” class. We assumed that all Common Eiders classified as dresseri originated from the Atlantic population breeding group. The data collections were made from several locations. For the purposes of our analysis, we pooled samples into 5 coastal areas: 1) Northeast Coast, 2) Bonavista, 3) Avalon, 4) Burin, and 5) South Coast (Fig. 1). Additionally, to maintain sufficient sample sizes in examining temporal patterns in the subspecies composition, we further pooled the three northeastern coastal areas (Northeast Coast, Bonavista, and East Avalon; the Northeastern Coastal Zone), and the 2 southern coastal areas, (Burin and South Coast; the Southern Coastal Zone). Band recoveries Banding and recovery data for all the Common Eiders banded in Newfoundland and Labrador were provided by the Bird Banding Office, Canadian Wildlife Service, for recoveries up to January 2000. Between 1983 and 1996, 857 adult female Common Eiders were banded in Groswater Bay and Table Bay in Labrador, and St. John Bay, Hare Bay, and Grey Islands in northern Newfoundland from 1983–1996. During 1988–1996, 3413 hand-reared ducklings were banded and released at Hare Bay and Sacred Bay in northern Newfoundland, and 156 hand-reared ducklings were banded and released in 1990 and 1991 at Oderin Island, in southern Newfoundland (Fig. 1; more on this hand-rearing program below). Our analyses were restricted to recoveries of these birds shot between September and April. We summarized the temporal and spatial distribution of these recoveries. A Common Eider propagation project was initiated in 1988, and was intended to increase the relatively small number of Common Eiders that nested in Hare Bay, Newfoundland. Smaller initiatives were undertaken at Sacred Bay and Oderin Island. Eider eggs were collected from colonies in Newfoundland, Nova Scotia, and Québec, and incubated to hatch. The ducklings were dry-brooded until they were four-weeks old, moved to a holding facility adjacent to a small eider colony, and allowed access to salt water for one to two days before being released. For the hand-reared ducklings released at Hare Bay, Sacred Bay, and Oderin Island, we determined the sex ratio at release and tested for differences in spatial patterns of direct recoveries (shot in the year immediately following banding) and indirect recoveries (shot one or more years after banding) (recovery type) using a chi-square test for goodness of fit (Zar 2009 S.G. Gilliland and G.J. Robertson 505 1984). We ensured data were homogeneous before pooling across release sites and years, and determined if there were different spatial patterns Figure 1. Locations of Common Eider banding sites in Labrador and Newfoundland, primary locations where eiders were collected by hunters, and delineation of coastal areas used for data summaries. 506 Northeastern Naturalist Vol. 16, No. 4 of recovery type using a heterogeneity χ2 test (Zar 1984). We then used recoveries of hand-reared ducklings released in Hare Bay and Sacred Bay to estimate juvenile recovery rates (f ’) and adult (1+ year old) survival (S), using Program MARK (White and Burnham 1999). Because there were only a small number of hand-reared ducklings released from Oderin Island in southern Newfoundland, we excluded them from this analysis. As we were constrained by a data set that only included juvenile marked birds, juvenile survival rates (S’) and adult recovery rates (f) could not be calculated (Brownie et al. 1985). However, if juvenile survival is held constant over time, adult survival rates are estimable (Barker and Buchanan 1993). Computationally, this is achieved by combining juvenile survival rates and adult recovery rates in a single parameter θ (Barker and Buchanan 1993). Model fitting was done by sequentially fitting models with fewer parameters. The global model allowed for sex and time (t) variation in all rates (S, f ’, and θ). Akaike’s information criterion (corrected for small sample sizes; AICc) was used to assess model fit. Reduced parameter models of juvenile recovery rates (f ’) were built first as they do not influence calculations of the other parameters in the model. Once the best-fitting parameterization for f ’ was found, θ was modeled. Finally the parameter of interest, adult survival S, was modeled with the best-fitting parameterizations of f ’ and θ. Due to the relatively sparse data, we could not calculate meaningful estimates of the variance inflation factor (ĉ). Sufficient data were not available to estimate adult survival rates from birds banded as adults, and their recovery rates were previously reported in analyses by Krementz et al. (1996). Results Examination of harvested bird carcasses King Eiders constituted 10% of the birds examined (n = 1672), and most were taken in the Northeast Coastal Zone (Fig. 2, and specifically in the eastern portion of this zone [Fig. 1]). Males constituted 54.7% of the population, and the sex ratio did not significantly deviate from 1:1 (χ2 =1.5, df = 1, P = 0.22), or vary across the non-breeding season (χ2 =4.3, df = 5, P = 0.50). The majority of male King Eiders were immature birds (85%, n = 91). The sub-specific classification for Common Eiders was: 25% dresseri, 20% borealis, 2% intergrades, and 53% “borealis or integrade” (n = 1205); 285 birds were not classified because of missing data. The combined “borealis- type” class comprised 75% of the sample. The sex ratio of harvested birds did not vary detectably over the non-breeding season (χ2 = 3.8, df = 6, P = 0.70), with 49% males, which did not detectably differ from a 1:1 ratio (χ2 = 0.48, df = 1, P = 0.49). The majority of male Common Eiders were immature birds (60%, n = 706). Along the Northeast Coast, Bonavista, and the Avalon areas, the majority of Common Eiders were the borealis-type (Fig. 2), primarily taken from January to March (Fig. 3). Along the South Coast and Burin coastal areas, the majority was dresseri (Fig. 2), and the largest harvests were in November 2009 S.G. Gilliland and G.J. Robertson 507 and December (Fig. 3). In the Northeastern Coastal Zone, borealis-type eiders dominated the harvest from October to April, while in the Southern Coastal Zone, dresseri was dominant in all months except April (Fig. 4). Figure 2. Species composition of eiders (King and Common Eider) and subspecific composition of Common Eiders in each coastal area for eiders hunted in Newfoundland, 1980– 1996 (sample sizes in parenthesis). 508 Northeastern Naturalist Vol. 16, No. 4 Figure 3. Seasonal distribution of Common Eiders collected from hunters in Newfoundland from 1980–1996. 2009 S.G. Gilliland and G.J. Robertson 509 Band recoveries Up until 2000, there had been 243 recoveries of Common Eiders banded in Newfoundland and 57 recoveries of birds banded in Labrador (includes all wild and hand-reared birds of all ages). These recoveries occurred in coastal areas of eastern Canada and the New England States (Fig. 5). The majority of the recoveries were on the island of Newfoundland; 54% (31 of 57) of recoveries from birds banded in Labrador and 79% (192 of 243) of recoveries from birds banded in Newfoundland, were on the island of Newfoundland. Within Newfoundland, 60% of recoveries were within the Northeastern Coastal Zone, while the remaining 40% were in the Southern Coastal Zone (n = 165). Recovery of hand-reared ducklings. The sex ratio at release did not deviate significantly from a 1:1 ratio of males to females (χ2 = 0.518, 1 df, Figure 4. Seasonal distribution of the Common Eider harvest by sub-species, for the Northeastern and Southern Coastal zones for eiders hunted in Newfoundland from 1980–1996 (sample sizes in parenthesis). 510 Northeastern Naturalist Vol. 16, No. 4 P = 0.472, n = 3569) and was homogeneous across years and release sites (χ 2 = 11.1, df = 9, P = 0.94, n = 3569). Recoveries of ducklings released from the Northern Peninsula occurred from Table Bay, Labrador south to Massachusetts (Fig. 5). The majority of recoveries (80%) were in Newfoundland and St. Pierre and Miquelon, with smaller numbers occurring in Labrador Figure 5. Recovery distributions of hand-reared Common Eiders released at Hare Bay (a) and Sacred Bay (b), Newfoundland, and wild female Common Eiders captured at breeding colonies in southeastern Labrador (c) and northern Newfoundland (d). 2009 S.G. Gilliland and G.J. Robertson 511 (1%), Québec (5%), Nova Scotia (8%), and the New England States (5%). Within Newfoundland, in the Northeastern Coastal Zone, most recoveries were made late in the fall and early winter, while recoveries in the Southern Coastal Zone were similar from November to March, except for a notable peak in December (Fig. 6). Recoveries from Newfoundland, St. Pierre and Miquelon, Labrador, Quebec, and the New England States were pooled (Nova Scotia was kept separate) to examine regional patterns of recovery type, and differed among areas (χ 2 = 37.9, df = 3, P > 0.001, n = 226). The primary difference in the pattern of direct and indirect recoveries was that a larger proportion of the indirect recoveries occurred in Nova Scotia and the New England States than expected (Fig. 7). Juvenile recovery rates of hand-reared ducklings banded only in northern Newfoundland were best modeled with time variation, with little evidence of a difference between the sexes (Table 1). The mean juvenile recovery rate (based on a random effects model) was 0.047 ± 0.007 (SE). θ was best modeled as a constant. Models with time variation or constant adult survival were similarly well supported, while there was only weak support for a sex difference. Due to sparse data and poor estimation of time-specific survival rates, model averaging was not useful, so the constant adult survival rate of 0.715 ± 0.051, with a 95% profile likelihood interval of 0.605 to 0.803, is presented. Direct recovery rates of hand-reared ducklings released at Oderin Island in southern Newfoundland in 1990–91 were high, at 0.147 ± 0.028 (23 of 156 banded ducklings). Figure 6. Seasonal distribution of band recoveries reported by hunters for Common Eiders banded in Newfoundland and Labrador, and recovered in the Northeastern Coastal and Southern Coastal zones of Newfoundland. 512 Northeastern Naturalist Vol. 16, No. 4 Discussion The composition of eiders harvested in Newfoundland varies considerably across the province. In northern and eastern areas, King Eiders are taken in some number, with borealis-type Common Eiders making up the bulk of Figure 7. Geographic distribution of direct (hatch-year) and indirect (after-hatchyear) band recoveries for hand-reared Common Eiders released in Newfoundland (sample sizes in parenthesis). Table 1. Model selection results for hand-reared Common Eiders released at 4 weeks of age in northern Newfoundland, 1998–1996. Model Number of ModelA AICc ΔAICc likelihood parameters Deviance St, f ’t, θ 2135.1 0.00 1.000 20 84.7 S, f ’t, θ 2135.3 0.20 0.927 11 103.0 Ssex, f ’t, θ 2137.3 2.10 0.347 12 103.0 Ssex*t, f ’t, θ 2145.7 10.5 0.005 29 77.0 Ssex*t, f ’t, θsex 2147.1 12.0 0.003 30 76.4 Ssex*t, f ’t, θt 2160.3 25.2 0.000 38 73.2 Ssex*t, f ’t, θsex*t 2172.5 37.4 0.000 47 67.0 Ssex*t, f ’sex, θsex*t 2181.1 46.0 0.000 40 89.9 Ssex*t, f ’, θsex*t 2183.6 48.5 0.000 39 94.5 Ssex*t, f ’sex*t, θsex*t 2184.4 49.3 0.000 56 60.3 AS = adult (1+ year) survival rate, f ’ = juvenile recovery rate, and θ = product of juvenile survival rate and adult recovery rate. 2009 S.G. Gilliland and G.J. Robertson 513 the harvest. Along the south coast of Newfoundland, dresseri-type birds predominated. However, both regions had mixes of the two sub-species in their harvest—notably, some proportion of borealis were taken throughout the season along the southern coasts of Newfoundland. Sexes appeared to be equally represented in the harvest of King and Common Eiders in the province. In many ducks, males comprise a larger portion of the harvest, and often there are more males in the population (Johnson et al. 1992). Sex ratios in the long-lived sea ducks, notably eiders, are probably closer to even (Swennen et al. 1979). Data from the Species Composition Survey (SCS) for Common Eiders taken in Newfoundland 1972–2006, show a slight female bias in the harvest (45% male; Gendron and Collins 2005). In Greenland, sex ratios of hunted birds were slightly male biased (57%) for Common Eiders, and even for King Eiders (Merkel 2004b). Similarly, males comprised 58% of the Common Eider harvest in Denmark (Noer et al. 1995). Large portions of the harvest comprised immature birds. This was expected, as immature waterfowl, including eiders, tend to be more susceptible to harvest (Joensen 1974). The SCS for Common Eiders taken in Newfoundland 1972–2006, show 60% of male eiders taken were immature (Gendron and Collins 2005), the same value found in our study. Similarly, hunters in Greenland took mostly immature birds (76–96% of Common Eiders and 56–92% of King Eiders harvested; Merkel 2004b), while 45% of the harvest in Denmark comprised immature Common Eiders (Noer et al. 1995). The temporal, spatial, species, subspecies, age, and sex distributions based on information collected from hunters, will reflect to a degree, the overall distribution of animals, but temporal and spatial patterns in hunting effort and sampling intensity, as well as differences in vulnerability to hunting among cohorts (e.g., age or sex) will lead to biases. We suspect the sex, species, and sub-specific composition of the harvest to reasonably reflect the overall eider populations in the region. The age structure of the harvest is not likely to reflect the population due to the higher vulnerability of immature birds. Temporal patterns in harvest, however, are more difficult to interpret, as they will reflect bird movements, local abundance of birds, and the behavior of hunters. The eider duck hunting season has generally lasted from late November to 10 March or late February over the study period. Because so few of our collections occurred outside the hunting season, the composition of eiders observed during the early fall and the spring may not fully reflect the distribution of eiders during these periods. The weather and sea-ice limits, and in some circumstances enhances, hunting opportunities in Newfoundland. Generally the weather, and its effect on sea conditions, is too severe to allow unrestricted access to areas suitable for sea-duck hunting. This limitation is especially true along the eastern portions of the Avalon Peninsula in January and February, so samples from this region of the Avalon were limited. The sea-ice that forms along the coast of Labrador and the eastern Canadian Arctic moves southward along northeastern Newfoundland in January through 514 Northeastern Naturalist Vol. 16, No. 4 April. Sea-ice both affects the duck hunter’s access to specific hunting areas and the locations of eider concentrations along the Northeast Coast. Sea-ice can serve as a barrier to hunters, but also has a calming effect on seas and will concentrate the eiders, making them more vulnerable to hunting if they can be reached. It is not known whether the composition of eiders harvested during heavy ice conditions would be the same in conditions with less ice. Nevertheless, the geographical distribution of our sample generally reflects previously described distributions of wintering eiders (Canadian Wildlife Service, St. John's, NL, Canada, unpubl. data; Gillespie and Learning 1974) and hunting effort (Wendt and Silieff 1986) in Newfoundland. Affiliations and movements American Common Eiders are harvested throughout the hunting season in Newfoundland and are the dominant subspecies of Common Eider taken along the South Coast, Burin Peninsula, and the French islands of St. Pierre and Miquelon. While they comprise a smaller component of the take along the Northeast Coast, this area over-winters the largest concentration of eiders in the province, and significant numbers of dresseri-type Common Eiders likely over-winter in this area. The temporal patterns of band recoveries for the hand-reared eiders released from the Northern Peninsula suggest that these birds begin their southern migration by November, with a pulse of birds that moves through the Northeast Coast during the fall. Some birds appear to stop and winter along the Northeast Coast, but most move to the South Coast and French islands to winter. As these birds aged, they were more likely to be recovered in locations south of Newfoundland, likely a function of a developing migration strategy and integration with the wild populations as they matured. In spite of being hand-reared, these birds did migrate seasonally, as is typical of wild eiders. The eiders that breed along the south coast of Labrador have similar movement patterns to the birds from the Northern Peninsula, although a larger proportion of these birds move to wintering areas around the Mingan and Anticosti Islands in Quebec, and the Maritime Provinces (Bourget et al. 1986). There is less information from which to determine the timing of migration, or the breeding affiliations for borealis and King Eiders. A large number of Common Eiders were classified in the “integrade” or “borealis or integrade” categories, suggesting many of the birds harvested in Newfoundland may be breeding in Labrador. However, Mendall’s key was developed with a limited number of samples of northern birds from a restricted number of sites, so this conclusion is somewhat tenuous. For birds breeding farther north in Arctic Canada, some move south to Newfoundland waters to winter, but many winter in Greenland (Lyngs 2003, Mosbech et al. 2006), while Common Eiders marked as breeding in Greenland have not been recovered in Newfoundland. These patterns suggest that many of the borealis Common Eiders wintering in Newfoundland (and Québec) are likely from more southerly breeding areas for the subspecies, but samples are limited and more work (banding and telemetry) is needed to fully understand the 2009 S.G. Gilliland and G.J. Robertson 515 sources of Common Eiders wintering in Newfoundland. Northern Common Eiders arrive at wintering areas in Newfoundland by November, but the temporal patterns of the eider harvest (assuming our data reflect relative level of harvest) suggest most arrive in northeastern Newfoundland, their primary wintering area, between December and March. As for King Eiders, Mehl et al. (2004) has shown that this species may range widely in winter, and although sample sizes were limited, they had evidence that individual birds wintered in different ocean basins (Pacific and Atlantic) in subsequent seasons. The harvest of King Eiders in eastern North America is very small overall; only in Greenland are significant numbers taken (Merkel 2004b). Direct recovery rates of hand-reared eiders averaged close to 5%, and varied annually. Adult female Common Eiders breeding in Atlantic Canada had recovery rates around 1% (Krementz et al. 1996), considerably lower than juvenile birds. Once again, higher susceptibility of juveniles to harvest was expected, and these data show juveniles may be 5 times more susceptible to harvest. However, this analysis was based on hand-reared birds, which may have higher than usual recovery rates. Interestingly, direct recovery rates of birds released at Oderin Island, which unlike Hare Bay and Sacred Bay, does not have a local breeding population, had very high direct recovery rates of 15%. With no adult birds to accompany, these young birds may have been more vulnerable to harvest. Although this study was not designed to assess the relative importance of releasing hand-reared birds in to an area with or without local wild populations, the difference in recovery rates would suggest releasing birds to an area with a remaining wild population may be beneficial. Krementz et al. (1996) documented a survival rate of 0.87 for adult breeding females in Atlantic Canada, so the adult survival rates of 0.71 shown in our study are low for Common Eiders, although not unheard of in over-hunted populations (H.G. Gilchrist, Environment Canada, Ottawa, ON, Canada, pers. comm.). However, the survival rate has wide confidence intervals, and a number of assumptions were required to obtain this estimate. The survival rate assumes that survival reaches a plateau after one year of age, which is likely not the case for a bird that shows delayed maturation. The constraints imposed in our modeling efforts may have biased estimates of adult survival rates; however, Barker and Buchanan (1993) showed that any bias introduced by constraining juvenile survival rate to be constant did not bias adult survival rates too much, if juvenile survival rates did not fluctuate widely. Finally, released birds were hand-reared, and for a variety of reasons may not have survival rates comparable to wild adult birds once they reached maturity. As such, we would not recommend using the value of 0.71 as the current estimate of adult survival of Common Eiders breeding in Newfoundland, but rather suggest this estimate may indicate that prebreeding age classes have reduced survival rates and that harvest rates or natural mortality of hand-reared birds may have been relatively high over the period of study (1988–2000). 516 Northeastern Naturalist Vol. 16, No. 4 Management implications Currently, the species, age, and sex composition of the duck harvest in Canada, including eiders, is estimated from the Species Composition Survey. This survey uses duck wings, collected from hunters, to determine the composition of the harvest. However, there are no criteria for classifying the sub-species of Common Eiders using wing characteristics. The subspecific composition of the eider duck harvest is an important demographic parameter to monitor for the management of Common Eiders in eastern North America and Greenland. The results of this study show that considerable demographic information on the eider duck harvest can be obtained from the collection of eider heads, and we believe a regular, and statistically valid, annual collection of eider heads would be worth considering. The timing and relative contribution of borealis and dresseri to harvest were different between the Northeastern Coastal and the Southern Coastal zones. Because the size and productivity of these populations are different, they may benefit from different harvest management regimes. Depending on management goals, seasons could be varied within these zones to modify the sub-specific composition of the harvest. For example, if protection of locally breeding dresseri was the priority, the hunting season in the Northeastern Coastal Zone could be delayed; possibly allowing a large portion of migrating dresseri to pass through this area before the hunt begins. Similarly, if protection of borealis was of concern, the seasons could be adjusted. For example, the eider season in the Northeastern Coastal Zone could be closed in the latter part of the winter, when hunting mortality will likely have the largest effect on population growth, while the season in the Southern Coastal Zone need not be affected. Acknowledgments We thank Ian Goudie, who provided unpublished data, and Jean-Pierre Savard, who reviewed an earlier version of this paper. Pierre Ryan coordinated the collection in Newfoundland, Canada, and Rodger Etcheberry the collection in St. Pierre and Miquelon, France. Lance Woolaver measured all the heads, and Rina Nickols compiled the initial data summaries. Finally, we thank all the hunters who provided the heads used in this study. Literature Cited Barker, R.J., and I.M. Buchanan. 1993. Estimating adult survival rates of New Zealand Black Swans banded as immatures. Journal of Wildlife Management 57:549–555. Bédard, J., A. Nadeau, J.-F. Giroux, and J.-P.L. Savard 2008. Eiderdown: Characteristics and harvesting procedures. Société Duvetnor Ltée and Canadian Wildlife Service, Québec Region. Sainte-Foy, PQ, Canada. 46 pp. Bourget, A., P. Dupuis, and W.R. Whitman. 1986. Les eiders hivernant dans le golfe du Saint-Laurent : Effectifs et distribution. Pp. 94–99, In A. Reed (Ed.). Eider Ducks in Canada. Canadian Wildlife Service Report Series Number 47. Ottawa, ON, Canada. 177 pp. 2009 S.G. Gilliland and G.J. Robertson 517 Brownie, C., D.R. Anderson, K.P. Burnham, and D.S. Robson. 1985. Statistical inference from band recovery data: A handbook. United States Fish and Wildlife Service Resource Publication Number 156, Washington, DC. 305 pp. Chardine, J.W., G.J. Robertson and H.G. Gilchrist. 2008. Seabird harvest in Canada. Pp. 20–29, In F. Merkel and T. Barry (Eds.). Seabird Harvest in the Arctic. Circumpolar Seabird Group (CBird), CAFF Technical Report No. 16 CAFF International Secretariat, Akureyri, Iceland. 77 pp. Chaulk, K.G., G.J. Robertson, B.T. Collins, W.A. Montevecchi, and B. Turner. 2005. Evidence of recent population increases in Common Eiders breeding in Labrador. Journal of Wildlife Management 69:805–809. Christensen, K.D., and K. Falk. 2001. Status of the Common Eider breeding population in the municipality of Avanersuaq (Thule), northwest Greenland. Polar Research 20:109–114. Gendron, M.H., and B.T. Collins. 2005. National Harvest Survey web site Version 1.2. Migratory Bird Populations Division, National Wildlife Research Centre, Canadian Wildlife Service, Ottawa, ON, Canada. Available online at http://www. cws-scf.ec.gc.ca/harvest/default_e.cfm. Accessed 17 October 2007. Gillespie, D.I., and W.J. Learning. 1974. Eider numbers and distribution off Newfoundland. Pp. 73–78, In H. Boyd (Ed.). Canadian Wildlife Service Waterfowl Studies in Eastern Canada, 1969–73. Canadian Wildlife Service Report Series Number 29. Ottawa, ON, Canada. 106 pp. Gilliland, S.G., H.G. Gilchrist, R.F. Rockwell, G.J. Robertson, J.-P.L. Savard, F. Merkel, and A. Mosbech. 2009. Evaluating the sustainability of harvest among Northern Common Eiders in Greenland and Canada. Wildlife Biology 15:24–36. Goudie, R.I., G.J. Robertson, and A. Reed. 2000. Common Eider (Somateria mollissima). No. 546, In A. Poole, and F. Gill (Eds.). The Birds of North America. The Birds of North America, Inc. Philadelphia, PA. 32 pp. Joensen, A.H. 1974. Populations and shooting utilization of migratory ducks in Denmark, with particular reference to the eider duck (Somateria mollissima). Pp. 252–261, In Proceedings XIII International Congress of Game Biologists, 1973. Stockholm, Sweden. Johnson, D.H., J.D. Nichols, and M.D. Schwartz. 1992. Population dynamics of breeding waterfowl. Pp. 446–485, In B.D.J. Batt, A.D. Afton, C.D. Ankney, D.H. Johnson, J.A. Kadlec and G.L. Krapu. (Eds.). Breeding Ecology and Management of Waterfowl. University of Minnesota Press, Minneapolis, MN. 635 pp. Krementz, D.G., J.E. Hines, and D.F. Caithamer. 1996. Survival and recovery rates of American Eiders in eastern North America. Journal of Wildlife Management 60:855–862. Lyngs, P. 2003. Migration and winter ranges of birds in Greenland. Dansk Ornitologisk Forenings Tidsskrift Number 97. Copenhagen, Denmark. 167 pp. Mehl, K.R., R.T. Alisauskas, K.A. Hobson, and D.K. Kellett. 2004. To winter east or west? Heterogeneity in winter philopatry in a central-Arctic population of King Eiders. Condor 106:241–251. Mendall, H.L. 1980. Intergradation of eastern American Common Eiders. Canadian Field-Naturalist 94:286–292. Mendall, H.L. 1986. Identification of eastern races of the Common Eider. Pp. 82–88, In A. Reed (Ed.). Eider Ducks in Canada. Canadian Wildlife Service Report Series Number 47. Ottawa, ON, Canada. 177 pp. Merkel, F.R. 2004a. Evidence of population decline in Common Eiders breeding in Western Greenland. Arctic 57:27–36. 518 Northeastern Naturalist Vol. 16, No. 4 Merkel, F.R. 2004b. Impact of hunting and gillnet fishery on wintering eiders in Nuuk, Southwest Greenland. Waterbirds 27:469–479. Mosbech, A., G. Gilchrist, F. Merkel, C. Sonne, A. Flagstad, and H. Nygaard. 2006. Year-round movements of Common Eiders breeding in the Arctic Canada and West Greenland followed by satellite telemetry. Ardea 94:651–665. Noer, H., I. Clausager, and T. Asfreg. 1995. The bag of eider Somateria mollissima in Denmark 1958–1990. Danish Review of Game Biology 14:1–24. Reed, A., and A.J. Erskine. 1986. Populations of the Common Eider in eastern North America: Their size and status. Pp. 156–162, In A.Reed (Ed.). Eider Ducks in Canada. Canadian Wildlife Service Report Series Number 47. Ottawa, ON, Canada. 177 pp. Swennen, C., P. Duiven, and L.A. Reyrink. 1979. Notes on the sex ratio in the Common Eider Somateria mollissima (L.). Ardea 67:54–61. Wendt, J.S., and E. Silieff. 1986. The kill of eiders and other seaducks by hunters in eastern Canada. Pp. 147–154, In A. Reed (Ed.). Eider Ducks in Canada. Canadian Wildlife Service Report Series Number 47. Ottawa, ON, Canada. 177 pp. White, G.C., and K.P. Burnham. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46:120–139. Zar, J.H. 1984. Biostatistical Analysis, 2nd Edition. Prentice-Hall, Englewood Cliffs, NJ. 718 pp.