Taxonomic History
The early authorities, for instance Bolle (1857) and Godman (1872), classified the prevailing form of gull in Macronesia as L. argentatus. Later they were assigned to L. cachinnans by Meade-Waldo (1893): 'Larus Cachinnans. Mediterranean Herring Gull (Gaviota). This is the common resident Gull. It breeds in all the islands. We never saw the true Larus argentatus'. Bannerman (1919) also classified them as L. cachinnans noting that the 'Yellow-legged Herring Gull is the common gull of the Mediterranean, Black and Caspian Seas. It breeds in all the north Atlantic islands. In winter it ranges down the west coast of Africa to Senegambia'.
The Yellow-legged Gulls of Macronesia were first classified at the subspecies level by Dwight (1922) as a race of the Lesser Black-backed Gull L. fuscus atlantis. Dwight had made an extensive study of skins obtained principally from the Azores (but with two from the Canaries), noting for atlantis: 'Compared with affinis [graellsii], the nearest race [of L fuscus] both in color and in distribution, the mantle of atlantis is a clearer, paler, bluer gray without any of the brownish tinge that marks all the other races even in perfectly fresh plumage, and, furthermore, when adult males are compared with males, and females with females, the average larger size is obvious'. His later classification (1925, p.214) was based on the 'yellow feet and a wing-pattern with very little white or gray...A single white mirror [is found] on the tenth (outer). Rarely does a second mirror develop on the ninth'. Dwight (1925, p.216) re-emphasised the heavier structure of atlantis compared to graellsii: 'It is not surprising for this island-form, like many others, to have a heavy bill, tarsus, and feet'. However, it appears he gave priority to the wing-tip pattern in making the final assignment. Another feature Dwight (1925 p.211) noted for atlantis was the similarity of its juveniles, other than being paler, to those of graellsii.
Chavigny & Mayaud (1932) studied a number of atlantis from the Azores and reported that there was never a mirror on P9 and that the mirror on P10 was smaller than that on michahellis. Teyssèdre (1983) broadly agreed with this stating that only three out of 23 specimens examined from the Azores (including Dwight's) had a white mirror on P9 and in each case it was small. All sources agree that atlantis is darker than Mediterranean Yellow-legged Gulls. Dwight (1925, p.122) described it as deep neutral grey (grade 5 on his darkness scale), darker than Kittiwake Rissa tridactyla (neutral grey, grade 4) and Yellow-legged Gull L. cachinnans (light neutral grey, grade 3) but paler than graellsii rated as dark neutral grey (grade 6). There are quite a number of other accounts of atlantis but they are reduced in value because they do not mention the islands on which the observations were made. It would appear that the most frequently reported wing-tip patterns, such as those shown by Jonsson (1998), with extensive black on P5 and P4 and little white on P10 and P9, prevail on the Azores. Wilds & Czaplak (1994) studied the timing of the primary moult using specimens from the Azores. For one specimen collected on 17th August, they found P1-P6 new, P7 growing and P8-P10 old; for further specimens collected from 6th-7th September, they found variation from P1-P7 new, P8 growing and P9-P10 old to P1-P8 new, P9 growing, P10 old.
The classification by Dwight as a race of fuscus did not become well-established. Chavigny & Mayaud (1932) placed the Azores birds under cachinnans because they thought the Azores birds resembled michahellis closely in all respects other than the amount of white on P9 and P10. Volsøe (1951) classified the Canary Islands birds under argentatus because their behaviour did not seem to deviate from other Herring Gulls. He had obtained a number of specimens from Tenerife in the Canary Islands and found that, compared to typical specimens form the Azores, they had more white on P10 (with one showing only a small dark subterminal bar on the outer web) and some had a small mirror on P9. Volsøe concluded that atlantis was not as sharply defined as supposed by Dwight. Bannerman (1963) also considered eventually that atlantis should be classified as a Herring Gull rather than as a Lesser Black-backed Gull because of its characters. He described atlantis on the Canary Islands as having a large mirror on P10, usually no mirror on P9 and only a narrow black band on the 6th primary (P5) with a black spot on the 7th (P4). He cited the mantle colour for specimens from the Canary Islands as light neutral grey, two shades lighter than graellsii. He also gave the bill as deep chrome with vermilion patch on lower mandible and the legs and feet as pale chrome yellow or yellow ochre. Bannerman continued: 'L. fuscus affinis [graellsii] is more easily confused with it, ... but it has usually a brownish tinge instead of the clearer, paler blue-grey back of L. atlantis; ... In all forms of L. fuscus, the bills and tarsi are much more slender than in examples of L. atlantis ... In size the Atlantis gull is rather larger than the lesser black-backed gulls ...'.
More recently divergence in classification has persisted with Teyssèdre (1983) considering atlantis as a form of Yellow-legged Gull and Moreno (1988) as a form of Herring Gull. When the Yellow-legged Gull was first claimed as a species in Britain in 1993 (Brit. Birds 86:1-2), atlantis was assumed to be a race of this new species although the taxonomic position in truth remained confused.
The relationship between atlantis and the nearby michahellis on the African and European coasts is a long-standing problem. Dwight (1925, p.208-210) said atlantis (Azorean Lesser Black-backed Gull) is resident on the islands and coasts of northwestern Africa including the Azores, Madeira, Canary and perhaps the Cape Verde Islands. Hence Dwight included the Azores, Canaries, Madeira and coastal Morocco and Western Sahara in the range of atlantis. Dwight's original papers on atlantis are available at Dwight. Shortly after Dwight's analysis, Ticehurst & Whistler (1928), also classified the Yellow-legged Gulls of Galicia as a race of the Lesser Black-backed Gull. Volsøe (1951) thought that those on the Canary Islands showed 'transition to the neighbouring race michahellis from the Mediterranean'. Dwight had noted (1925 p.207) the same transition from the opposite perspective that the Mediterranean form of cachinnans approaches atlantis on the western limits of its range with a tendency in the wing-tip for only one mirror (P10) and for the white to be replaced by grey. Bannerman (1963) thought that the gulls breeding at Berlengas in south-west Portugal and those found on the Morocco coast in winter were the same as those found in the Canary Islands. Stegmann (1934) had also come to this conclusion, grouping all the Yellow-legged Gulls of Macronesia, Portugal and the Atlantic coast of Morocco together as atlantis because of their common mantle shade (paler than brittanicus [graellsii] but darker than those on the western border of the michahellis) and wingtip pattern (extensive black, white spot on P10 and at times on P9). Vaurie (1965) noted that, on wing-tip patterns, the birds of the Canary Islands, classified as L.argentatus atlantis, showed a tendency to michahellis with 'occasionally a white spot, or ... some traces of white, on the [ninth] primary'. He concluded that this tendency to be whiter was emphasised more in the birds of the Atlantic coast of Morocco but these (south to Mogodor at least) are more similar to michahellis than they are to atlantis. Etchécopar & Hüe (1967 p.275) considered that the birds breeding on the island of Mogador, about 150km north of Agadir, were intermediate between atlantis and michahellis. Bernis (1954) thought that the gulls nesting on the Atlantic coast of the Iberian peninsula belonged to michahellis but also showed slight tendencies towards atlantis. Skornik (1997) considered that the birds on the Atlantic coast of Spain showed characteristics intermediate between cachinnans and argentatus. Jonsson (1998) thought that birds breeding on the Atlantic coast of the Iberian peninsula and Morocco stand morphologically closer to atlantis and are better treated as such.
Summarising the above, in the past atlantis has been classified as a Herring Gull (because of its behaviour), a Lesser Black-backed Gull (because of its wingtip pattern) and a Yellow-legged Gull (because of its leg colour). As we shall see in the ensuing accounts, the Atlantic Yellow-legged Gull can be regarded rather appropriately as a mixture, in varying proportions, of all these three species. The Yellow-legged Gulls on the Atlantic continental coasts have usually been classified as michahellis/cachinnans rather than atlantis but a number of observers have noted a close relationship between atlantis and these forms, particularly in Morocco and Portugal.
Copyright © Nick Rossiter 1999-2002. All rights reserved.