|
Photo © Roy Harvey |
|
 |
|
|
IDENTIFICATION OF ADULT VEGA GULL: FIELD OBSERVATIONS FROM JAPAN, FEBRUARY 2003
Chris Gibbins
Introduction
In February 2003 I spent 7 days looking at gulls at Choshi, a fishing port on the island of Honshu, Japan. My main interest was in the phenotypic variability of Vega Gull L. vegae. This taxon winters in large numbers at Choshi and was the most abundant large white headed gull present during my visit. In this paper I aim to describe and analyse patterns of variability in adult vegae relative to other taxa within the fuscus-argentatus-cachinnans complex. Before discussing the characteristics of vegae, an overview of field methods is necessary.
Methods: assessment of colours, grey tones and wing-tip pattern of Vega Gull
Ideally these parameters should be assessed from birds in the hand. I did not have this luxury and so attempted to be as objective as possible, within the constraints imposed by field observation. Birders may well argue that it is important to have an understanding of whether taxa such as vegae are diagnosable in the field; consequently, characters that are only observable in the hand or biometric measures that are apparent only at the population level may be perceived as being of little practical value. In this context, my field observations of vegae using the methods described below may be of value.
To produce relative values for the upperpart tones of vegae I used a grey scale based on percent saturation (Figure 1). Although graded at 5% intervals, in the field I found it difficult to work consistently to such a fine resolution; categories were therefore amalgamated to describe grey tones at 10% intervals (e.g. 10-20, 20-30 etc). I was careful to assess tones in neutral light conditions with all birds side-on.
I used a colour chart to assess leg and bill colours in the field (Figure 2). The colour range illustrated in Figure 2 is a subset of the complete colour pallet available on computer graphics software; numbers in the Figure are catalogue codes. The pattern of black on the bill of individual vegae was coded according to the categories presented in Figure 3.
Iris pattern was assessed using the scheme presented in Figure 4. In vegae, as with some other gull taxa, the creamy or yellow iris is variably peppered with dark spots. In the field, these spots can make the eye look anything from pale (few or no spots) through greyish (medium spot density) to blackish (high spot density). I used the scheme in the field to allocate each bird to one of the categories shown in Figure 4. As discussed by Yésou (2001) with respect to Mongolian Gull L. (cachinnans) mongolicus, assessment of eye pattern and colour is extremely difficult in the field. While the scheme I used is by no means ideal (e.g. spots are not distributed evenly as in Figure 4 but often coalesce to form dark patches), it is more objective than simply referring to eyes as, for example, 'dark' or 'pale'. The data give an index of the appearance of the iris pattern of vegae in the field.
Upperpart tones and bare part colours and patterns were assessed for a sample of adult birds in the field. Sample sizes for assessment of each parameter are given in subsequent figures. Adult birds were taken as those showing no trace of brown in the upperparts. As emphasised most recently by Howell and Mactavish (2003), such assessments should be seen as approximate, providing simply a relative guide to field appearance. Digital images and video footage taken of individual birds were also used to assess bill structure and primary pattern. All photographs were taken with a Sony DCR PC120 digital video camera, combined with a Leica APO telescope. Moult stage was assessed for a sample of both flesh and yellow legged birds.
Figure 1. Grey scale used in assessment of upperpart tones of adult Vega Gulls
Figure 2. Colour chart used to assess bill and leg colours of adult Vega Gulls
Figure 3. Scheme used to assess patterns of black on the bill of adult Vega Gulls. Because of more or less continuous variation in bill markings, few individuals matched exactly the patterns illustrated here; individuals were simply allocated to one of the four general descriptive categories and the percentage of birds in each category was then calculated. The categories were labelled: (1) no black, (2) black present on lower mandible, (3) black present on both mandibles, and (4) black markings forming ring across both mandibles.
Figure 4. Categories of spot density used to classify iris patterns of adult Vega Gulls. Percentage values are the proportion of the iris covered by spots.
Field characteristics of adult Vega Gull
Plate 1. Vega Gull, Choshi, Japan, 18/02/03. This bird illustrates a very typical adult vegae. Note the bird's structure, extent and intensity of head streaking, dark-looking eye, leg colour and upperpart tones. Each of these characters is discussed in detail with respect to the birds shown in subsequent plates. The frequency of different upperpart grey tones is summarised in Figure 5. Grey tones ranged from 15-40% saturation, although the vast majority of birds (83%) had grey tones of 20-30% saturation. This compared to 30-35% saturation for kamtschatschensis Common Gulls, 40-45% for Black-tailed Gulls and 45-50 % for the Slaty-backed Gull present alongside. Herring Gulls range from approximately 5-10% (argenteus) to 20-30% (argentatus). Thus, the most frequent grey tone in vegae equates to the darkest of Herring Gulls. Because 10% intervals were used in the field assessment of birds at Choshi, it is difficult to be any more precise than this; differences between Herring and Vega Gulls may be more apparent if a 5 % interval scale was used and/or specimens were examined in the hand. There may also be differences in colour hue between the two, but assessment of this would require birds to be side by side.
Figure 5. Upperpart tones of adult Vega Gull present at Choshi (n = 59)
Plate 2. Vega Gull, Choshi, Japan, 18/02/03. Another fairly typical bird. As with the bird in Plate 1, note the general proportions of this individual. It lacks bulk and has a rather gentle feel; the head is high and rounded and the bill rather short. Together with the dark looking eye, these traits give an impression not dissimilar to some larger kamtschatschensis. The legs of this bird are dull pinky-flesh; this basic tone was shared by 93% of vegae, with only 7% having yellow type legs (n = 101; see Plates 16 and 17 for a more detailed discussion of leg colour).
Plate 3. Vega Gull, Choshi, Japan, 18/02/03. Note that the structure, leg colour and dark-looking eye of this individual are similar to the birds in Plates 1 and 2. Basally the bill has the same greenish cast as the birds in Plates 1 and 2 and becomes more yellow in its distal portion. Bill colour is discussed further with respect to Plate 5.
Plate 4. Vega Gull, Choshi, Japan, 18/02/03. This bird's bill has a black spot within the red gonys area. The extent of black on the bills of vegae varied, although data suggest a dominant 'type', with most birds (68%) having no black (Figure 6). As with this individual, 28% of birds had a black spot confined to the lower mandible while birds with black on both mandibles were rare (<10 % of individuals).
Figure 6. Patterns of black on the bills of adult Vega Gull during February (n = 103).
Plate 5. Vega Gull, Choshi, Japan, 19/02/03. An example of a bird with black on both mandibles (i.e. Type 3 from Figure 3). Also note that its bill colour is rather orange-yellow. Around 50% of birds had broadly yellow toned bills, 33% greeny-yellow and 17% orange-yellow bills (n = 101). This variability may represent transition from winter to summer condition.
Plate 6. Vega Gull, Choshi, Japan, 18/02/03. Note the pale-looking eye of this individual. In fact all birds appeared to have some degree of pigmentation on the iris; with individuals such as this, pigmentation was limited to such a small number of spots that from a distance the eye essentially looked pale. The frequency of different degrees of iris spotting is shown in Figure 7. Allowing for the difficulties of assessing the exact extent of spotting in the field, data suggest that most Vega Gulls have light (1-24%) to medium (25-49%) spot densities. In the field, this gave most birds a facial expression that was quite unlike Herring Gull.
Figure 7. Degree of eye pigmentation (% of iris covered in spots) in adult Vega Gulls (n = 100)
Plate 7. Vega Gull, Choshi, Japan, 18/02/03. Note that on this individual the red of the gonys does not reach the cutting edge of the bill. This pattern was unusual, occurring in only 5-10% of birds (n = 102). On most birds (~90%) the red reached the cutting edge of the bill, while only a very small number (~2%) had red extending onto the upper mandible. These percentages are given as approximate values because it was very difficult to assess the true extent of the gonys spot in the field. The bird in Plate 7 is a clear example of an individual where the red does not reach the cutting edge but in many the upper edge of the red faded gradually to orange yellow, so deciding whether the gonys spot reached the cutting edge was problematic. The general conclusion to draw from these data is that in most birds the red reaches the cutting edge, while individuals with either limited or extensive red are infrequent at this time of year. It remains to be determined whether this is seasonally variable; individually marked birds need to be tracked to see whether, as might be expected, the red becomes more vivid and extensive in the breeding season.
Plate 8. Vega Gull, Choshi, Japan, 22/02/03. Another typical individual: note the general structure, pink legs, medium dark eye (26-50% spot density) and a red gonys spot confined to the lower mandible.
Plate 9. Vega Gull, Choshi, Japan, 21/02/03. A relatively weak-billed bird; compare this individual (bill length:depth ratio 2.36) with those in Plates 10 and 11.
Plate 10. Vega Gull, Choshi, Japan, 23/02/03. Bill slightly more robust than average (length:depth ratio 2.28) and rather michahellis-like. Note the steep curvature at the tip of the upper mandible (resulting from a relatively small radius) compared to the bird in Plate 9.
Plate 11. Vega Gull, Choshi, Japan, 23/02/03. A very large, robust (male?) bird with a bill length:depth ratio of 2.18. The gonys angle is particularly marked. Also note the greyish looking iris, resulting from an even peppering of spots (assessed in the field to be 26-50% density), and the typical deep red orbital ring of vegae.
Plate 12. Vega Gull, Choshi, Japan, 18/02/03. As with this bird, most vegae sampled (68%) had black on 6 primaries (i.e. black extended inwards as far as P5). The extent of black on P5 varied, with the most common pattern being similar to a michahellis, with a broad band extending across both webs (as in this plate). Birds were also observed with isolated black patches on each web of P5 while on some the black was confined to the outer web of P5; a very small number (<1%) lacked black on P5.
The extent of black in the wingtip of the vegae from Choshi is compared to data for mongolicus given by Yésou (2001) in Figure 8. It is clear from this comparison that, at the population level, vegae has black on fewer primaries than mongolicus. Moreover, observations from Choshi indicate that in those vegae with black on P4 (25 % of birds), the black is confined to the outer web of the feather. Although Yésou does not discuss this specifically, Plate 98 of his paper shows a mongolicus with black extending across both webs of P4, suggesting that at least some can have more black on P4 than found in even the most extreme vegae. No birds with black on P3 (i.e. black on 8 primaries) were observed at Choshi whereas 11% of the mongolicus in Yésou's sample had black on P3. (In addition to surveying the 114 birds included in Figure 8, I routinely checked the primary pattern of preening vegae and did not see any birds with black on P3.) Overall, data suggest that average or population-level differences between vegae and mongolicus exist. The key question concerns whether my sample of 'vegae' included some mongolicus. This is discussed later.
Figure 8. Extent of black in the primaries of a sample of vegae and mongolicus.
Plate 13. Vega Gull, Choshi, Japan, 18/02/03. Note the typical pattern on the outer primaries, with a mirror on both P10 and P9. On some birds the mirror on P10 was much larger and separated from the white tip by only a narrow black band, as for example with the bird shown in Plate 14.
Plate 14. Vega Gull (with Black-tailed Gulls), Choshi, Japan, 18/02/03. Note the narrow black band on P10 (visible on the underside of the far wing) separating the large mirror from the white tip.
Plate 15. Adult Vega Gull with a 'thayeri type' P10 pattern, Choshi, 22/02/03. The extremes and some intermediate patterns of P10 observed at Choshi are illustrated in Figure 9. A random sample of 30 birds showed that 23 (76%) had a mirror isolated from the white primary tip by a complete black band (bottom two birds in Figure 9), 4 (13%) had the mirror separated from the tip by an incomplete black band (middle and second upper birds in Figure 9), while 3 (10%) had the white mirror and wingtip completely merged (upper bird). In addition to this survey of 30 birds, many hundreds of P10 patterns were routinely checked. Only 2 individuals were seen with a thayeri type pattern, where the tongue on the inner web of P10 merged with the white of the feather-tip (upper bird in Figure 9). Plate 15 shows one of these two birds. Also note that this individual lacks black on P5. Other than their wing-tip pattern, these two thayeri individuals looked typical.
Figure 9. Variation in P10 pattern observed in adult Vega Gull at Choshi.
Plate 16. Unidentified gull, Choshi, Japan, 19/02/03. Note the leg colour of this bird, matching colour 712/3 in the colour chart (Figure 2). Non-pink legged birds such as this made up around 7% (n = 101) of the Vega-type gulls present and had legs assessed as matching either colour 122/1225, colour 149 or colour 712/3; another bird had legs that I described as grey flesh with a greeny cast to the front of the tarsus. These compare to the various pink and flesh colours of the majority of vegae (highlighted in Figure 2). No birds with truly yellow legs (again, as highlighted in Figure 2) were seen.
Plate 17. Unidentified Gull, Choshi, Japan, 21/02/03. Another non-pink legged bird, with legs assessed in the field as matching 712 in the colour chart (Figure 2). Note however that the reproduction of this image makes the legs look flesh coloured. On this individual P10 is still regrowing. Of the sample of pink legged birds assessed for moult (n = 127), 9.5 % were still regrowing P10 whereas 88% of the non-pink legged birds seen (a total of only 9) were still re-growing P10. Unfortunately it is not possible to be confident whether this apparent difference in moult stage is real or simply a sample size artefact.
Plate 18. Unidentified Gull (centre) with 2 cy Slaty-backed Gull (left) and adult Vega Gull (right), Choshi, 23/02/03. Compared to the vegae, note the upperpart tone of this bird (assessed at 40% saturation) and its rather slender build. It has a complete black band across P5 and black on the outer web of P4. In some ways this bird fitted expectations of 'taimyrensis', but it had flesh legs.
Discussion
Although many taxa within the fuscus-argentatus-cachinnans complex have been the subject of recent interest, vegae has remained somewhat neglected. For example, although they indicated that they collected data on vegae, Leibers et al (2001) omitted this taxon from their paper describing the genetic and phylogeographic relationships between gulls within the complex. While information on the phenotypic characters of other taxa increases at pace, much of what is known of vegae is either old or not widely accessible to European ornithologists. Note for example that in his discussion of vegae, Yésou (2002) cited research mainly from the 1920's and 30's, with more recent references listed in his paper relating to either non-English language material or unpublished data. My general aim with the current paper was to provide further descriptive information on vegae. More specifically I have attempted to add some detail as to the variability of vegae to the summary information provided by Malling Olsen in his recently published gulls monograph.
Subsequent to my visit, Nial and Charlie Moores provided a wealth of photographs and descriptive information on 'the Herring Gull assemblage' in South Korea, including discussion of the identification of immature birds (http://www.wbkenglish.com/lwhgid.asp). However, their discussion of the separation of adult vegae from mongolicus was based largely on subtleties in jizz and head streaking; they provide no real assessment of relative wing-tip patterns, upperpart tones, eye pigmentation, bill markings or leg colours. My observations, summarised in the plates and figures presented above, hopefully help fill this information gap.
The bare part (eye and bill) colours, upperpart tones, bill structure and wing-tip patterns described above suggest that, even allowing for imprecision associated with field observation, adult vegae is consistently diagnosable from most taxa within the fuscus-argentatus-cachinnans complex. In a west European context, Herring Gull is perhaps most similar to vegae. Differences in iris pigmentation, orbital ring colour and the number of primaries with black pigmentation should allow separation of typical individuals of these two taxa. In particular, the darkest Herring Gulls (i.e. those closest in upperpart tones to vegae) should be those with the most limited amount of black in the primaries (i.e. L.a. argentatus), so in most cases identification should be straightforward. A bird that combines dark upperparts with a peppered iris, a broad black band across both webs of P5 and some black on P4 would, as far as I am aware, be out of the range of argentatus. Clearly, the major source of confusion for vegae is with mongolicus (Mongolian Gull). Further comparative work on these two taxa is needed before their systematic relationship can be elucidated. Yésou (2002) suggests that until such work is undertaken, mongolicus may be best treated as a subspecies of Vega Gull (following Helbig et al., 2002). While acknowledging that mongolicus is probably best regarded as a full species, Malling Olsen (2003) actually treats it as a subspecies of Caspian Gull L. cachinnans. Similar uncertainties surround the identification and taxonomic position of vegae relative to taimyrensis and birulai; all of these issues are discussed below.
The obvious problem with my data on 'vegae' is that the origin of the birds cannot be proven and, hence, statistics represent a meaningless amalgamation of various wintering taxa. While for the purist, all descriptive work on gulls should be undertaken on the breeding grounds, this is often only possible for those fortunate enough to be engaged in professional research. At Choshi, there is the real problem of describing vegae from a sample that contains an unknown number of mongolicus. Two points argue against this possibility. First, mongolicus is considered to be white headed in winter but, with the exception of one or two birds not included in any of my sample data, I saw no white headed birds during my stay (Figure 10). Second, the frequency distribution plot for the number of primaries with black pigmentation (Figure 8) shows clear separation of the two data sets (mine and that of Yésou, 2001). If my 'vegae' sample contained a mix of both vegae and mongolicus, there would be some birds from Choshi with black on 8 primaries and hence the frequency plots would not separate in the way that they do. While there are statistical nuances which mean that this approach is not absolutely watertight, it is a strong indication that the Choshi sample is from a single 'population' (i.e. vegae) that is different to that of Yésou (mongolicus).
In his review of the fuscus-argentatus-cachinnans complex, Yésou (2002) argues that taimyrensis and birulai are invalid taxa. He suggests that taimyrensis has likely been described erroneously from either hybrids (between heuglini and vegae) or by scientists describing populations from the range of heuglini when in fact they were further east and within the range of vegae. With respect to taimyrensis, Yésou's assertion raises the question as to what the darker-backed and yellow-legged birds wintering in places such as Japan and Hong Kong actually are. As yellow-legged but otherwise identical birds occur in populations from across the range of vegae, he argues that to accord these birds subspecies status (i.e. naming them birulai) is incorrect. Yésou's paper rather suggests that King and Carey (2000) were being a little ambitious with some of their identifications of birulai in Japan (they described birulai as being generally a little larger, with a heavier, more angular structure, slightly darker upperparts, darker pink legs, a darker iris, heavier head streaking and less black in the wingtip than vegae).
The vegae type gulls at Choshi with non-pink legs presented a real identification challenge (Plates 16 & 17). They could simply be 'yellow' legged vegae (i.e. the invalid 'birulai' of Yésou) or what have previously been defined as 'taimyrensis'. Some were rather dark (hence fitting 'taimyrensis'), others were not. I saw few of these birds so it was difficult form any general conclusions. What was interesting however was their moult was consistently later than the average pink legged vegae (see caption to plate 17). Conversely, some extremely dark birds had flesh or pink legs (e.g. Plate 18), but again small numbers limited the possibility of reaching any firm conclusions as to their identity. I have seen such birds labelled as hybrids or intergrades between heuglini and vegae by some authors.
Figure 10. Extent of head streaking in adult Vega Gull in late February (n = 102). Examples of the most frequent type are illustrated in Plates 1-8. An example of a bird with streaks predominantly around the eye and nape is shown in Plate 13. Only 1 % of the sample had unstreaked heads; these individuals were omitted from analyses summarised in Figures 1-9.
Conclusions
Although variable, the average adult vegae appears distinct from other taxa within the argentatus-cachinnans-fuscus group. The extent of black in the primaries and the extent of winter head streaking seem to offer the best chances of separation from mongolicus, the most similar taxa. Differences in jizz between mongolicus and vegae, recently suggested by some authors, may also help separate these taxa, but with my limited experience I would not like to comment on this. Small numbers of birds present at Choshi (around 5 % of individuals) fell outwith the normal range of variability of vegae, being too dark and/or non-pink legged; the majority of these birds had late primary moult. Identification of these individuals is not clear. My suspicion is that in the past these would have been labelled either taimyrensis or birulai. However, Yesou (2002) argues that these taxa are invalid. Clearly much remains to be learnt about the gulls wintering in Japan, and in particular the origin of darker-backed and/or non-pink legged birds.
References
Howell, N.J. and Mactavish, B. (2003). Identification and variation of winter adult Kumlien's Gulls. Alula 1 (9), 2-16.
King, J. R. and Carey, G.J. (2000). Gulls in Japan. Birding World 13, 160-163.
Liebers, D. and Helbig, A.J. (2002). Phylogeography and colonisation history of Lesser Black-backed Gulls as revealed by mtDNA sequences. Journal of Evolutionary Biology 15, 1021-1033.
Liebers, D., Helbig, A.J. and de Knifff, P. (2001). Genetic differentiation and phylogeography of gulls in the Larus cachinnans-fuscus group (Aves: Charariiformes). Molecular Ecology 10, 2477-2462.
Olsen, K-M. and Larsson, H. (2003). Gulls of Europe, Asia and North America. Pica Press. Yésou, P. (2001). Phenotypic variation and systematics of Mongolian Gull. Dutch Birding 23 (2), 65-82.
Yésou, P. (2002). Systematics of Larus argentatus-fuscus-cachinnans complex revisted. Dutch Birding 24 (5), 271-299.
Acknowledgement
Thanks in particular to Asoa and Michiaki Ujihara for help with site information and some very useful discussions of Japanese gulls in the field. Thanks also to Brian J Small for useful comments on an earlier draft of this article.
Chris Gibbins - email: c.gibbins AT abdn.ac.uk
|
|
|
|
