Thayer's Gull (thayeri)
(last update: January 22, 2013)
Thayer's Gull adult December
below you will find part 3 of the paper: Hybridization and changes in the distribution of Iceland gulls
"we" in the text below refers to the original authors. If any errors occur in this text, please let me know and mail to marsmuusseatgmaildotcom.
Hybridization and changes in the distribution of Iceland gulls
British Columbia (BC) and Washington are the main wintering area for thayeri (Campbell et al., 1990; Morgan et al., 1991). It is ‘abundant’ or ‘very abundant’ in BC with 15 counts of concentrations of 500-2000 feeding birds (Campbell et al., 1990). In the Great Lakes it was a minority of wintering Iceland gulls in southern Ontario (Pittaway, 1992) and at a Québec site it was 3% of 1392 winter Iceland gulls in 14 years (Steeves, Holohan & Bathurst, 1989). It is rare on the east coasts of Canada and the USA (American Ornithologists Union, 1983; Godfrey, 1986). We know of only two pelagic records off eastern Canada (Brown, 1986; J. K. Jensen, in litt.).
Few early or recent records of kumlieni from Pacific coasts are now accepted (Weber, 1981; Godfrey, 1986; Campbell et al., 1990). Most Iceland gulls wintering in the Great Lakes region are kumlieni, i.e. in southern Ontario (Pittaway, 1992). At a Québec site it was 97% of 1392 winter Iceland gulls in 14 years (Steeves et al., 1989). In coastal/insular eastern Canada and the north-eastern USA it is widespread and at least locally common (Powers, 1983; Brown 1986; Godfrey, 1986; Zimmer, 1991; A. J. Gaston, in litt.). Coastally and offshore ‘white-winged’ adults might be either pale kumlieni or glaucoides but usually were distinguished in sight records from ‘typical’ kumlieni. The abundances of white-winged birds and typical kumlieni at sea in winter between Canada and west Greenland were of the same order (Brown, 1986), but white-winged birds were almost unknown among Iceland gulls off north-east US coasts (Powers, 1983). Earlier, white-winged adults were said to reach those coasts mainly in invasions (Bent, 1921).
Only three specimen records of winter adult glaucoides from eastern Canada were accepted by Godfrey (1986). One of those is now missing from the NHM (this study). A subsequent specimen from southern Ontario in 1974 was accepted by Godfrey (Pittaway, 1992). We traced none from the USA. Only 1% of 630 recoveries of Iceland gulls ringed as chicks in west Greenland were from Canada or the USA (Salomonsen, 1967, 1979b). It is clear that glaucoides rarely reaches the North American mainland.
The taxonomic composition of earlier and more recent samples of winter adult specimens are compared from Greenland to the British Isles (Table 4). Samples up to 1915 were 75-100% glaucoides, but those after 1950 were 33-80% kumlieni. The large differences apply to all areas (see below). Statistical tests were inappropriate because of the very small samples and qualifications noted in Table 4. The post-1950 samples from east of Greenland are also mainly from storm-driven invasions, when kumlieni is more frequent than in other years (Ingolfsson, 1967; Fjeldså & Jensen 1985; Weir et al., 1995).
Wintering Iceland gulls were common on the west and south coasts of Greenland (Hagerup, 1891; Winge, 1898). Greenland specimens, which were probably all glaucoides (see above), may have included winter birds. West Greenland birds mostly winter there; 98% of 630 recoveries of those ringed as chicks were from Greenland (Salomonsen, 1967, 1979b). Birds moved south in winter, with northern coastal limits at about Disko Bay (Salomonsen, 1967) and south of close ice or in polynyas at sea (Brown, 1986). Wintering gulls at a large polynya off east Greenland included no Iceland gulls (Hjort, Hakansson & Stemmerik, 1983). In September, ship surveys in the north-west Iceland – east Greenland - Jan Mayen Island triangle, only 3.2 Iceland gulls/year were recorded and all from near east Greenland coasts (I. Petersen, 1995, and in litt.). Two immature specimens in winter from south-west Greenland in 1898 and 1905 were attributed to kumlieni (Salomonsen, 1967), but the 1898 bird was later attributed to thayeri (Boertmann, 1994). Identifications of variable immatures of unknown origin are speculative. At sea between Greenland and Canada adult kumlieni was common in winter (Brown, 1986). Coastally it was not given as a winter visitor by Salomonsen (1950, 1967), but was regarded as regular by Olsen (1991) and Boertmann (1994).
The Iceland gull was the only gull wintering in Iceland that did not breed there (Saunders, 1882-1884). The little-known east Greenland population was thought to be the source of most birds wintering in Iceland. They are common there (below) and only 1% of 630 recoveries of birds ringed as chicks in west Greenland were from Iceland (Salomonsen, 1967, 1979b).
Eleven winter adult Iceland gulls collected by I. Dinesen in north-east Iceland in 1907-13 and now in ZMK, NHM and NMS, included one kumlieni, the first known from Iceland. Samples listed in Table 4 suggest that kumlieni was as common or commoner than glaucoides after 1954, but the proportion of kumlieni may be substantial, as in the 1966 sample (Table 4) and the present relative abundance of the two taxa cannot be assessed (A. Petersen, in litt.).
Christmas Bird Counts on 5% of all Icelandic coasts averaged c. 1000-3000 Iceland gulls of all ages per year, but these coasts included many which were favoured by Iceland gulls (A. Petersen & Hjartarson, 1989, 1991, 1993; A. Petersen, in litt.). The total winter population of both taxa combined may exceed 10 000 birds in some years (A. Petersen, in litt.). Usually, small numbers of pale immature and mainly female Iceland gulls winter in the Faroe Islands (J.-K. Jensen, in litt.). The first known kumlieni specimen was in 1905 and no more were known until the adults, which predominated in the 1983 storm-driven invasion (Fjeldså & Jensen, 1985).
In the British Isles the Iceland gull formerly was a rare winter visitor (MacGillivray, 1824, 1852). Only one of four adult or subadult specimens for before 1950 was kumlieni from 1869 (Weir et al., 1995). Most sight records of adults in the storm-driven 1983 and 1993 invasions were of white-winged birds and four of seven NMS 1993 adult specimens were glaucoides (Weir et al., 1995). Numbers wintering in the Faroe Isles and British Isles combined probably exceed 1000 only during some invasions.
‘Evidence’ of hybridization between Iceland gull subspecies was formerly from specimens of unknown origin with wingtip melanism which was thought to be unusual for the location (Dwight, 1906, 1917, 1925; Taverner, 1933; Rand, 1942; MacPherson, 1961). Definitions of taxa from wingtip melanism were not agreed upon between workers (Rand, 1942). Biometrics were few; e.g. MacPherson (1961) could find measurements of only one supposed smithsonianus x thayeri. In the field complete reproductive isolation between smithsonianus, thayeri and kumlieni was reported by Smith (1966), but this aspect of his study has been discredited (Knudsen, 1976; Gaston & Decker, 1985; Snell, 1989, 1991c).
Within Iceland gulls sympatric thayeri and kumlieni hybridize freely and pairings are more or less random in relation to wingtip melanism and iris colour (Knudsen, 1976; Gaston & Decker, 1985; Snell, 1989). There was suggestive evidence of hybridization between thayeri and glaucoides at the only site where sympatric breeding was supported by specimens (Table 3). ‘White-winged’ birds in east Baffin Island were sympatric with thayeri at 73°N and freely interbred with thayeri at 69°N (Reynaud et al., 1981; Snell, 1989). ‘White-winged’ birds were sympatric with kumlieni in east Baffin Island as far back as 1878 (Kumlien, 1879) and the two taxa are now sympatric in Greenland (see above). The palest-winged kumlieni cannot have persisted for c. 150 years and spread if kumlieni only hybridizes with darker thayeri. Moreover, kumlieni tends to be darker in the west and paler in the east (above), which also suggests kumlieni has resulted from hybridization between thayeri and glaucoides.
While the replacement of glaucoides by thayeri was eastwards, distributions of other large gulls of the region shifted mainly north. The southern limits of glaucous gulls moved north in eastern Canada (MacPherson, 1961). Great black-backed gulls L. marinus spread north in Greenland (Salomonsen, 1950, 1979a), and in Arctic Russia where they became sympatric with or displaced glaucous gulls (Uspenskii, 1984; Kalyakin & Ponomareva, 1999). Birds of the herring gull complex are recent colonists of Arctic Siberia (Uspenskii, 1984). They colonized Iceland from the 1920s (Ingolfsson, 1970) and south Greenland from the 1970s (Boertmann, 1979). The exception seems to be Canada where thayeri was already present further north. Herring gulls bred as far north in western Foxe Basin in the 1820s (Table 3) as in the 1980s (Forbes et al., 1992). In Baffin Island, they bred about as far north in the 1900s (Hantzsch, 1930) as in the 1980s (Snell, 1989).
Iceland gull distribution shifts involved 40° of longitude. There were limited east-west or west-east components in the herring gull colonizations of Iceland and Greenland (above), but other longitudinal shifts involved only small numbers of large gulls in North America. Lesser black-backed gulls L. fuscus from Europe crossed the Atlantic and now appear to breed in eastern Canada (Godfrey, 1986; Steeves et al., 1989). There is slight introgression of western Pacific slaty-backed gulls L. schistisagus into western and south western Alaskan populations of glaucous-winged gulls (Kessel, 1989; Petersen, Weir & Dick, 1991; Bell, 1996).
Distribution shifts by Iceland gulls differed so greatly from those by other large gulls of the region that they could not be due to the same set of external variables. We believe that thayeri spread east and interbred with glaucoides, giving rise to the variable hybrid kumlieni. As the eastward shift continued, the hybrid zone moved east and introgression may now extend to the eastern limits of the greatly reduced range of glaucoides.
Relatively simple criteria are widely considered to distinguish unstable, hybrid populations from stable, valid taxa (Otte & Endler, 1989). Five criteria identify kumlieni as an unstable, hybrid population:
(1) specimen data confirm that the postulated parent taxa formerly bred sympatrically and may have interbred;
(2) known range contraction by one parent taxon corresponds in space and time to known or apparent range expansion by the other parent;
(3) non-assortative breeding by the hybrid with one parent is known and is strongly suggested with the other;
(4) the hybrid is variably intermediate in wingtip melanism and iris colour and this variation tends to be geographical with respect to the parent taxa;
(5) the present range of the hybrid is limited between those of the parents and overlaps both near the Low/High Arctic boundary. Also the hybrid has not spread into the High Arctic where thayeri has directly replaced glaucoides.
One genetic study shows that thayeri, kumlieni and herring gulls are very closely related, but other northern large gulls are too, and variability is greater between colonies than between species (Snell, 1991a). Although more thayeri material is still needed for another molecular study, preliminary indications are that thayeri differs somewhat from kumlieni, glaucoides and herring gulls (M. Willcox, in litt.). This does not preclude hybridization between thayeri and glaucoides, given its frequency in large northern gulls (Snell, 1991a,b, 1993; Bell, 1996). Also thayeri and glaucoides need not necessarily be one species; it is only necessary that kumlieni is intermediate between them.
Possibly thayeri bred in a High Arctic `Banksian Refugium' of western Canada during the Wisconsonian Glaciation, and glaucoides in a Low Arctic refugium of the western Atlantic (Salomonsen, 1972). Wherever thayeri bred formerly, in Canada the Cordilleran and Laurentide Ice Sheets parted and circumferentially shrank early in the present inter-glacial (Pielou, 1989). This would have exposed western Arctic coasts and an overland route from the Pacific along which thayeri migrates today (Johnson & Herter, 1989). Salomonsen (1950, 1967) treated glaucoides as Low Arctic, but it bred far into the High Arctic only 150 years ago and its current restriction to more or less the Low Arctic cannot indicate a simple climatic limit.
Populations wintering in the Pacific and the Atlantic might have advanced and retreated several times in the Arctic in accordance with climatic change during the present inter-glacial. They need not have hybridized during every meeting and when they did, gene flow might have been in either direction. It appears that the brief systematic ornithology of the Arctic happened to coincide with the most recent hybridization between thayeri and glaucoides, an eastward shift in distribution by thayeri of c. 40° of longitude and further introgression by the hybrid into the remaining range of the Atlantic parent. Compared with other hybrid and parent distribution changes in northern large gulls (Snell, 1991a,b, 1993; Bell, 1996), the extent and speed of change in Iceland gulls has been extreme.
Hybrid zones are common in birds. They may be very wide and most are explained in northern regions by secondary contact during the present interglacial, but most do not move far (Hewitt, 1989). Hybrids are presumed adaptively less fit and tend to occur in density troughs between the parental peaks (Barton, 1989). In the Iceland gull there were rapidly shifting distributions of both parent and hybrid taxa, but kumlieni is now probably more numerous than thayeri, and glaucoides may be 10 times more abundant than either of them.
Other workers (e.g. Dwight, 1906; Snell, 1989) have suggested that Kumlien's gull is a hybrid and we present strong new evidence to support this. Further genetic studies are still desirable, although separation of large northern gulls is extremely difficult using mitochondrial DNA (Snell, 1991a; Bell, 1996). A study of microsatellite DNA will probably be required. Whatever taxonomic treatment is adopted, Kumlien's gull is a useful common name for the hybrid population and it commemorates a man who made an outstandingly comprehensive study of Arctic natural history in a single year.
We warmly thank staff of the institutions listed in Table 2. Some of them and other workers who provided valuable additional data, comments or both were Drs W. R. P. Bourne, Aberdeen, J. Fjeldså & I. K. Petersen, Copenhagen, V. N. Kalyakin, Moscow, A. Petersen, Reykjavik and J.-K. Jensen, Nùlsoy, Faroe Isles. We also thank the Royal Danish Consulate, Edinburgh, The Institute of Scandinavian Studies and C. Duxbury (Department of Russian) of Edinburgh University and the Scottish Ornithologists' Club (SOC) for translations from the Danish, Russian and German.
This paper benefited greatly from two important Arctic natural history literature bequests to Edinburgh libraries, the Harvie-Brown Bequest to the NMS and that of George and Irene Waterston to the SOC. We thank the Editor and referees for valuable comments to the paper; Dr A. J. Gaston, Québec, helped immeasurably from his wide field knowledge of Canadian breeding dispersions of thayeri and kumlieni, which he generously shared with us.