Willet Identification - Western or Eastern?

Kayo Roy writes: "On Sunday, June 27, 2004 Alan Smith, Brian Ahara and I found and photographed an adult Willet on the Lake Ontario shoreline of Jones Beach in the Port Weller section of St. Catharines, Ontario. In my rough notes, I recorded the bird as just a very early, dark adult Willet. Once home, and after spending some time looking at the photos, and checking a few bird guides, I determined the bird to be of the Eastern 'semipalmatus' race. My reasoning was that the bird was so heavily spotted and barred dark brown/black, with the barring extending well down the flanks to the very bottom of the underparts. Noted was a whitish spot above the lores, a narrow white eye ring, a dark brown line across the lores, small white chin, coverts showing many numerous small gray/white areas, white undertail coverts, and a white tail with some brownish barring."

Fig. 1. Willet photographed on Jones Beach, St. Catherines, Ontario, Canada, 27 June 2004. Photo copyright Kayo Roy.

"My photos clearly gave me the feeling this adult bird was an 'Eastern' and not the 'Western' form which is considerably paler and with very little flank barring. There has been considerable e-mail discussion with Ontario birders (Alan Wormington, Bob Curry, Ron Pittaway, Michel Gosselin and Ron Tozer), with some varied opinions expressed. It was suggested that ID Frontiers would be an excellent forum to solicit further justified opinions and not just simple votes. We look forward to your comments."

"Everyone should know that there is no confirmed record of 'semipalmatus' Willet in Ontario, however there is one 1985 undocumented sight record from Long Point"

Fig. 2. Willet photographed on Jones Beach, St. Catherines, Ontario, Canada, 27 June 2004. Photo copyright Kayo Roy.

Fig. 3. Willet photographed on Jones Beach, St. Catherines, Ontario, Canada, 27 June 2004. Photo copyright Kayo Roy.

(1) Richard Chandler writes: In reply to Angus Wilson's posting, may I say that I am delighted that this topic has been raised, as I have spent the last twenty years trying to distinguish between breeding inornatus and semipalmatus!

I have photographs in my collection of breeding-plumage individuals that fall into two categories (plus some that cannot be assigned to either group) that seem to correspond to the "classic" plumages of the two races, namely: semipalmatus, with extensive strongly coloured feathers on the upperparts, and fairly extensive bold barring on breast and flanks, and (though it does not show in all the photos) barring on the uppertail. The other group, presumed inornatus, shows a limited number of breeding feathers on the upperparts, more sparse and less prominent barring on the breast and particularly on the flanks, and, where the photos show the tail, a plain grey un-barred uppertail. These photos include presumed inornatus in California, Texas and Kansas in April, and presumed semipalmatus in California (including the bird shown in Rosair & Cotteridge p115I, though none of my photos of this individual show the uppertail), in Texas in April, and in Florida in May and June. A photo of what I regard as "classic" inornatus is shown in, for example, Dennis Paulson's Shorebirds of the Pacific NW (p164). I am puzzled by the occurrence of apparent semipalmatus in California; could this be semipalmatus or is this bird simply showing plumage at one end of the range for inornatus? On this basis, the Ontario Willet, which shows quite numerous strongly marked upperpart feathers, strongly barred breast and flanks, and clearly shows a barred uppertail, should be semipalmatus.

In the old world there are shorebirds such as Common Redshank and European Golden Plover, where the races that breed in the south of the species' range, and therefore breed earlier, acquire rather more non-breeding feathers than individuals of the races that breed further north. Are there any new world examples of this phenomenon? Could the same, or something similar, apply to the two Willet races?

And while we are discussing Willets, may I ask about juveniles of the two races? David Sibley, in his Guide to Birds, suggests that semipalmatus is generally darker on the head and breast, and has less "crisply marked coverts" than inornatus. Can anyone comment on this, please? Richard Chandler (11 Jul 2004)

(2) Dennis Paulson writes: I guess it wouldn't be appropriate for me to transmit the (fairly brief) text on Willet subspecies in my new shorebird book until it's published by Princeton University Press, but I'll say that I've looked hard at the subspecies of Willets for a decade and a half. I had studied lots of them in several museums with large series when doing my 1993 book, and, while doing the new shorebird book, I examined recently collected and beautifully prepared series in the Burke Museum. I can only say they still leave me very puzzled. There are some great average differences, as everyone knows, but it is fairly common to find a specimen from the breeding grounds that could never be distinguished from its counterparts at the other end of the country - and this includes all the "classic" marks mentioned by Richard Chandler.

I think Richard's idea about birds in some areas molting more feathers and thus becoming more heavily marked is a good hypothesis
for Willets. Thus perhaps birds of one breeding population (subspecies) can readily molt to look like the other, even though average differences in the progress of their molt cause them to look rather different most of the time. This is of course only a hypothesis, but there is no doubt that some of the variation among alternate-plumaged shorebirds is caused by their molting more or fewer feathers. Lots of western Willets are more heavily marked than the Oregon bird in my 1993 shorebird book, I should add.

In my opinion, it is impossible to be sure of a supposed out-of-range Willet subspecies. The slight size differences that show up in comparing series of museum specimens would certainly not be detectable in a single bird. I'd also be hesitant at calling a subspecific vagrant just on the basis of being a bit smaller or larger than the other flock members. I think individual variation precludes using plumage as diagnostic either. A combination of heavy markings *and* small size (or light markings and large size) in a specimen should be sufficient evidence, and it would be interesting for someone to look at a lot more Willet specimens than I was able to examine.

Juvenile Willets vary greatly from scarcely marked to heavily marked, more than juveniles of just about any other shorebird that I know of, yet I don't think the variation is geographic. I will have photos of lightly and heavily marked birds in my new shorebird book, the former from the East and the latter from the West. But the juvenile in my 1993 book is from Maine, presumably semipalmatus, and it is strikingly marked (the "crisply marked coverts" David Sibley said were characteristic of western birds).

I should add that from what we know of shorebird movements, there should be no great surprise at an eastern Willet in Ontario in June, but it seems difficult to prove. Thanks. Dennis Paulson (13 July 2004)

(3) Michael O'Brien writes: This Ontario Willet is a good example of a bird in which structure is more distinctive than plumage. While the plumage of this bird may be somewhat equivocal, I believe structure clearly indicates Western Willet. Specifically, the smoothly rounded contours of the breast and back, the thick chest, and the slim, very straight bill are all indicative of Western. Although the amount of barring on body and tail broadly overlap between Eastern and Western, the overall cold coloration of the upperparts contrasting with a buff-tinged belly, the very pale gray ground color to the upperparts (and hence, more contrasty barring), and the neatly-spotted look to the breast are all strongly supportive of Western. While some Willets may be unidentifiable to subspecies, most seem to be identifiable by a combination of structure and plumage. Michael O'Brien (14 July 2004)

(5) Shai Mitra writes: I share the perception of many birders that at least some, probably many, and possibly most Willets can be confidently field-identified to subspecies under favorable circumstances, and that size and structure are particularly useful--even in the field.

My experience with this identification question is in a region (Long Island, New York) where both subspecies occur regularly, but where inornatus (or at least diagnosable inornatus) are much less numerous than locally breeding semipalmatus. Each year around mid-late July, a few huge, pallid, long-legged, long-billed, basic-plumaged birds suddenly appear at favorable sites on the south shore of Long Island (such as Moriches and Shinnecock Inlets), where they wade around, dwarfing the clamorous throngs of alternate-plumaged and juvenile semipalmatus. The combination of multiple distinctive characters, opportunities for direct comparison with up to 50 or more nominate birds, and ample specimen evidence supporting regular occurrence of western birds in the region obviously makes these calls less stressful than those involving the unprecendented occurrence of an isolated individual.

With this forum’s recent thread on quantitative variation in mind, I performed a simple exercise using the only mensural data available to me, those supplied by Hayman et al. My idea was to get a rough idea of the degree of field-diagnosability of the two subspecies. Hayman et al. present ranges of values for wing, bill, and tarsus length for the species as a whole. They furthermore describe thresholds below and above which subspecific identification is ‘almost certain.’ For each of the three variables, the range of ambiguous values between the two thresholds was roughly 10% of the overall range of values, implying that a large proportion of individuals could be identified in the hand, with a reasonable level of certainty, on the basis of individual measurements.

For the purposes of this exercise, I made a number of simple assumptions that I think are reasonable. Those (like Dennis Paulson) with access to actual distributions of values rather than simple ranges could perform similar exercises with far fewer assumptions. My basic approach was to assume that roughly 95% of semipalmatus should show values between the minima cited for the species and the thresholds above which birds are ‘almost certainly’ inornatus; and similarly that roughly 95% of inornatus should fall between the thresholds below which birds are ‘almost certainly’ semiplamatus and the maxima for the species:

1. The Minimum values given by Hayman et al. are ~2 SD below the mean for semipalmatus.
2. The Maximum values are ~2 SD above the mean for inornatus.
3. The thresholds (I’ll call these ‘threshold 1’ for each variable) above which birds are ‘almost certainly’ inornatus are ~2 SD above the mean for semipalmatus.
4. The thresholds (‘threshold 2’) below which birds are ‘almost certainly’ semipalmatus are ~2 SD below the mean for inornatus.
5. An individual semipalmatus would be field-identifiable by a given mensural character, under favorable circumstances, if its value were less than
0.95*threshold 2; that is, if it were more than 5% smaller than the smallest 2% of inornatus.
6. An individual inornatus would be field-identifiable if it measured above 1.05*threshold 1; that is, if it were more than 5% larger than the largest 2% of semipalmatus.

Crunching the numbers, the proportions of individuals of each subspecies that fell more than 5% beyond the extreme ends of the other subspecies’ distributions were as follow

semipalmatus: 38% by wing, 44% by bill, and 67% by tarsus
inornatus: 9% by wing, 51% by bill, and 39% by tarsus

The utility of these characters would be greatly enhanced if they were considered in combination, and identification would be further facilitated by any available clues concerning plumage, molt, date, and behavior. Food for thought. Shai Mitra (14 July 2004)

Although the data are admittedly limited and the assumptions vulnerable to some uncertainty, I think it is an understatement to say that 'some inornatus are bigger than all semipalmatus.' Considering all three linear characters together, the percentage of genuinely intermediate individuals would be very small (and this is without considering plumage, behavior, and the subtle shape clues described by Michael O'Brien).

Putting this in a field perspective, my thinking was that a linear size difference of 5% is plausibly (not obviously) discernible in the field to experienced observers, given opportunities for direct comparison. That's why I set my field-ID threshold 5% BEYOND THE LIMITS of variation in the other taxon. The reason that the western Willets we identify on the East Coast are 'obviously' different-looking is that the ones we identify are not tiny western Willets, the nominate birds standing next to them are not uniformly enormous nominate birds, direct comparison of multiple characters is generally possible, and plumage cues are available for corroboration.

Compare a perfectly average inornatus to an average-small semipalmatus, one SD below the mean (16% of birds should be one SD or more below the mean, meaning that several such birds might be expected in modest group of nominate birds):

wing bill tars
Average inornatus 214 62 65
Semipalmatus (-1 SD) 192 52 53
% larger than semi. 11% 19% 23%

The size and shape disparity between such individuals would be striking.

Yes, there would also likely be several average-large semipalmatus in the flock, but an average inornatus would still exceed even these by almost 10% in both bill and tarsus length. These are actually pretty big differences in the scheme of field identification. The average difference between the two Willets is at least as great as that between Yellow and Myrtle Warblers, which many would agree is often discernible even on lone birds--despite the fact that these two actually have overlapping mensural ranges (believe it or not!).

Compare Hayman et al.'s data for the two yellowlegs. It is true that wing and bill length do not overlap for these two, but tarsus length actually overlaps to an extent comparable to that between the two willets:

Lesser 46-58, Greater 56-70, Overlap/total range 2/24
semipalmatus 50-63, inornatus 61-70, Overlap/total range 2/20

Curiously, semipalmatus has roughly similar tarsus length to Lesser Yellowlegs, whereas inornatus is roughly similar in this feature to Greater Yellowlegs.

Here's one more cautionary note--not a warning against TRUSTING carefully analyzed differences in size, but against DOWNPLAYING them on the basis of the published size ranges (rather than parameterized distributions).

What I wrote above about the warblers was off the top of my head, based on exhaustive familiarity with those two species. I just now checked Curson et al.'s data for aestiva-group Yellow Warblers and Myrtle Warblers:

Yellow Warbler (aestiva group only): male 58-72 (100), female 55-65 (100), pooled 55-72
Myrtle Warbler: male 68-78 (30), female 63-75 (30), pooled 63-78

Pooling sexes (as in Hayman et al.'s Willet data), these two species overlap by a whopping 9mm out of a total range of 23 mm in wing chord! Yes, these warblers are differently shaped (Yellow is relatively short-tailed), but Curson et al.'s tail ranges nevertheless overlap by 6mm out of a total range of 21mm! By the way, my own samples of these species from New York and Rhode Island (thousands of MYWA and 500+ YWAR) closely match Curson et al.'s values, so there is little reason to suspect that their data are biased by pooling disparate geographic populations.

My conclusions are as follows:

1. Although caution is always warranted in all aspects of bird identification (and size presents special obstacles to field discernment), the two Willet subspecies are at least as disparate in several mensural characters as are many taxa that readers of this list would agree are discernibly different in the field.

2. Comparing simple ranges of values, rather than parameterized distributions, can be VERY misleading when one is trying to evaluate the degree of disparity among taxa.

'In the small series in the Leiden Museum, collected by Haverschmidt from September to April, the birds show at most only traces of the breeding plumage. The race inornatus was identified only three times: female 3 March 1976, bill 63.0, tarsus 68.6, wing 214 mm, weight 285 g, female 3 March 1976, bill 65.5, tarsus 74.0, wing 219 mm, weight 310 g, both shot near Nickerie; Sex? 6 September 1980, bill 66.8, tarsus 72, wing 220 mm, shot near Weg naar Zee, north-west of Paramaribo (Spaans & Swennen 1982 Wader Study Group Bull.34:32-34).' [Added 16 July 2004: Arie Spaans just informs me that all his Surinam measurements were taken
from the maximum wing chord. NvS]

>They also give tail length and tarsus length but I could not spare the time to write them down.< May I recommend the use of a digital camera? Quite handy.

A word of warning is, I feel, appropriate when comparing wing-length from various sources: before the 1970s wings were measured as to the natural shape whereas later measuring the fully stretched wing came into fashion. Measuring the same wing by these two methods may give differences of up to 10 mm's!

What about US/Canadian banding results? Are there any? Norman D. van Swelm (15 July 2004)

(18) Angus Wilson writes: Having seen few (if any) fresh alternate-plumage Western Willets, I found the collection of Wisconsin images from John Idzikowski (message 16) instructive. On my computer monitor, the bills of these birds look very dark, almost black. Is this true in life?

Image 3 of John's collection includes a bird that is stretching its wing and shows the central tail feathers nicely. These appear to be quite STRONGLY barred, in constrast to the unbarred or faintly barred of others birds in the group. How typical is this? For ease of discussion, I've included a blowup of this portion of the image as Fig. 5 in the discussion summary that follows Kayo Roy's Ontario 'mystery' willet.

Fig. 5. Detail of the tails of two alternate-plumaged Western Willets in Milwaukee, Wisconsin, 6 May 2001. The central tail feathers of the left hand bird appear to be strongly barred. On the righthand bird the tail is folded but visible feathers are essentially unbarred. Photograph copyright John Idzikowski.

The dark bars look narrower that most Eastern Willets but seem worth noting nonetheless. Something record committee members might tuck away in their memory banks. I'm sure those with access to specimen collections can comment further.

With respect to identification of the Ontario bird, good arguments have been made that this is probably a Western Willet based on structure. Measurements of ratios within the photographs by Ian Mclaren would seem to be the most objective evidence. We have learned that voice is likely to be a strong indicator of subspecies, but surely this of limited value away from the breeding range? There has been less discussion of the Ontario bird's plumage. I get the impression that people are tentatively identifying Western Willets on the Atlantic coast in mid-summer onwards using a COMBINATION of molt state (they are pale gray unlike the much more heavily marked breeding Eastern Willets) and size/structure. Can we conclude from this, that adult Westerns molt earlier than Easterns? Alternatively, are birds reaching the Atlantic coast in mid-summer predominantly non-breeders or failed-breeders that have begun prebasic molt earlier than local Easterns?

One last point that I'm still very much in the dark about, is the movements of Eastern Willets that nest in coastal saltmarshes along the Mid-Atlantic states. Where do they go and when? Do they really vacate the area entirely or do they simply move out into the bays spending more time on open mudflats making them harder to observe? The BNA account speculated that Easterns might make a transoceanic flight to the Caribbean or South America. Is there any new data to support this idea? Field observations by myself and others suggest that plenty of Eastern's do linger and begin molt becoming harder and harder to differentiate from Westerns. But on Long Island NY at least, Willets do disappear eventually and birds from mid-fall onwards are rare indeed.

The willets are a great example of a familiar bird (for North Americans at least) of which we know so little! Angus Wilson (17 July 2004)