Probability in bird identification

One of my regular birding spots is a small farm field near my house in Concord, Massachusetts. I can walk around the entire place in about 20 minutes, but I usually take about two hours and get in some good sparrow-study. The clump of taller vegetation shown here is always a good spot to study birds that come up out of the grass and weeds and perch a little more conspicuously, so I usually spend a lot of time standing here looking at sparrows through my telescope.

On one visit in September 2008, I noticed that the three common species of sparrows – Savannah, Song, and Swamp – were using different parts of the vegetation, and I made a note to watch for this on future visits. It seems to be fairly consistent, so I can now produce a kind of "map" of sparrow-regions within the shrub thicket.

I've divided the thicket into three zones that I'll call the margins, the middle, and the depths. I was seeing mainly Savannah, Song, and Swamp Sparrows, and those are the species I'll discuss here.

• The Margins are occupied by Savannah Sparrows and are rarely used by the other two species (almost never by Swamp)
  
• The Middle zone was used by all three species, although I would say it was preferred by Song Sparrow, and was used rarely by Swamp Sparrow.
  
• The Depths were the home of Swamp Sparrows, which rarely venture into the other zones. Song and Savannah sparrows would also go deep into the shrubs when alarmed (e.g. if a hawk flew by), but their typical behavior when flushed by me was to land on the outer edges of the thicket.

Of course, in real life this thicket is three-dimensional, but the zones I've drawn are more like a cross-section. So a Savannah Sparrow could head for the lower parts of the thicket and still land in the margins of that shrub.

This may seem trivial, but it's really no different from habitat and other "soft" identification clues. You certainly can't identify a bird solely by its chosen perch, but we should never doubt the capacity of the human mind to discern subtle differences in probability (just look at baseball batting averages). The more I watched the more I realized that these are helpful clues to identification, just as date and habitat are helpful clues. A sparrow that flies right up into the highest twigs of the cherry sapling is almost certainly a Savannah, and it doesn't take much more to confirm that. A sparrow that goes straight into the deep cover and then barely pokes its head out is likely to be Swamp (or White-throated, or sometimes Song, but at least we can start with some idea of the likely species).

Check it out in your local area. I'd be interested to hear how it works in other places and with other species. And of course the same principle should be applicable to lots of other situations.

More vocal copying by American Goldfinch, Purple Finch, and Ovenbird

After listening carefully to Pine Siskins this year to pick out the calls of other species (see my earlier post) I have run across several other mimics. In early June in Arlington, Massachusetts I heard a remarkable American Goldfinch song that incorporated several phrases of a Song Sparrow song along with Northern Flicker and I suspect other species were being copied but I couldn't pick them out. This is the first time I've heard an American Goldfinch that included copies of other species in its song, and these were very clear and obvious copies, but as Dan Lane pointed out in comments on my previous Siskin post American Goldfinch seems to be exceptional – the one American member of the genus that does not copy. Apparently they do, at least occasionally.

In western Massachusetts in late June 2009 I heard a male Purple Finch giving its territorial song. This territorial song began with a bright warble like a typical Purple Finch song, then continued into a longer and more varied section – a jumble of short song phrases and copies of other species. I was able to pick out copies of American Robin, Tufted Titmouse, Eastern Towhee, and Hairy Woodpecker, and again I suspect there were more. In this case the copying has been previously reported in the literature, with the BNA account mentioning Barn Swallow, American Goldfinch, Eastern Towhee, and Brown-headed Cowbird.

Several years ago I noticed a few copies in the flight song of an Ovenbird, and I've heard a few others since then. Over several days in late June 2009 I was able to hear many flight songs from one Ovenbird in Colton, New York, which included copies of Red-eyed Vireo song phrase, Robin tut note, and a Purple Finch song phrase. There are very few recordings of the flight song (since it is given at such long intervals) but one on the Stokes Guide to Bird Songs: Eastern includes a lot of mimicry. Listening carefully to that I can hear a snippet of Winter Wren song, the trill of Eastern Towhee song, Red-eyed Vireo song phrase, and ending with White-eyed Vireo song. It's such a jumble of calls that (like the Siskin song) it's very hard to pull out all the copies, so I suspect there are others, and this whole performance – other than the few typical Ovenbird song phrases in the middle – might be copies. This is another instance of vocal copying that has never been reported before, as far as I know. The BNA account for Ovenbird makes no mention of it.

It's interesting that both the Purple Finch and Ovenbird use copies in these sort of extended or auxiliary songs, and not in the more frequently heard "typical" songs.

I should point out a very interesting new blog by Nathan Pieplow about bird vocalizations, with a comment on the Pine Siskin issue here, and lots of other good stuff. And another blog about bird voices by Paul Driver, with a recent post about Mimicry here.

References:

McGraw, Kevin J. and Alex L. Middleton. 2009. American Goldfinch (Carduelis tristis), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America Online: http://bna.birds.cornell.edu/bna/species/080

Van Horn, M. A. and T.M. Donovan. 1994. Ovenbird (Seiurus aurocapilla), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America Online: http://bna.birds.cornell.edu/bna/species/088

Wootton, J. Timothy. 1996. Purple Finch (Carpodacus purpureus), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America Online: http://bna.birds.cornell.edu/bna/species/208

Identification of Belding's Savannah Sparrow

In early March 2009 it was my pleasure to spend a few days in California at the San Diego Bird Festival. On my first morning in San Diego I headed straight out to the Tijuana River National Estuarine Research Reserve hoping to spend a few hours studying Savannah Sparrows (who wouldn't, right?). I had seen some Belding's Savannah Sparrows in March 2008 in Orange County, California, and realized that the treatment of this endangered subspecies in my field guide was not very good, so I was anxious to learn more.

I succeeded in seeing dozens of the resident Belding's, with many birds singing and territorial around the marsh. I also saw quite a few of the migratory northern subspecies of Savannah Sparrows (ranging from grayish to brownish) in weedy areas around the margins of the marsh.

The image below compares two of my photos showing the most obvious differences between the subspecies. Compared to northern birds, Belding's has:

• longer and thicker bill with curved culmen
  
• darker and much thicker streaks on the underside
  
• darker streaks on the upperside and coarser and darker markings on the face
• Belding's never (as far as I could see) raises a crest, while the northern birds often look slightly crested and show a distinct peak on the rear crown
  

Belding's (left) and Northern-type Savannah Sparrows. Photos 5 and 8 March 2009, San Diego, CA 

 copyright David Sibley (digiscoped with Nikon Coolpix 4300 handheld to Swarovski telescope)

Differences in behavior, habitat, and molt were also fairly obvious, and should be very helpful for identification (but bill and plumage features should be checked to confirm). 

• Belding's in early March were singing and territorial, chasing each other across the marsh, while the northern birds were in small, loose flocks moving together widely across dry grassy and weedy areas around the marsh
• Belding's flight was low and weak, with round body and fluttering wingbeats, reminding me of Sharp-tailed Sparrow, unlike the stronger, higher, more "bounding" and undulating flight of northern Savannahs
• in early March the Belding's were all in clean fresh plumage, while the northern birds were molting and scruffy with missing feathers all over the head and body

I could not hear any difference in the song of Belding's compared to Savannah Sparrows from farther north.

I was struck by how much stockier and larger-billed the Belding's looked in comparison to the northern subspecies, in many ways intermediate between northern and "Large-billed" Savannah Sparrows. Apparently some populations intermediate between Belding's and northern birds are resident farther north along the California coast, but alongside the migratory northern subspecies the Belding's Savannah Sparrows in southern California seem quite distinctive.

Vocal copying by Pine Siskins

[10 July 2009 - update added at end]

In July 1990 I spent several weeks birding in west-central British Columbia, where Pine Siskins were one of the most conspicuous birds, and I became convinced that Pine Siskins were copying the sounds of other species of birds and incorporating these sounds into their songs. I was able to identify American Robin (squeal call), Evening Grosbeak (“krrr” call), Warbling Vireo (scold call and a song phrase), White-winged Crossbill (flight call and trill), and even House Sparrow (calls).

Since then I’ve heard this in many other places, and come to expect it whenever I hear siskins singing. From late March to early April 2009 I had the opportunity to listen to several individuals repeatedly in my yard in Concord, Massachusetts, and I was able to identify the following 14 species in the siskins’ songs:

Evening Grosbeak “krrr” call

American Robin, flight squeal and single “pup” call

Eastern Bluebird, rattling flight call

Eastern Towhee, “chewink”

Black-capped Chickadee, a single “dee” note

Northern Flicker, “kew” call

Hairy or other woodpecker, “wika” call

Dark-eyed Junco, soft “tew-tew-tew” call

Eastern Phoebe, song phrase and chase call

Blue-headed Vireo, two syllables of scolding call

Red-eyed Vireo, scold call

Song Sparrow, “jimp” call

Ruby-crowned Kinglet, two-syllabled whistled song phrase

Common Redpoll, flight call

These sounds can be very challenging to pick out, since the song moves along so rapidly that the bird is already singing other sounds while your brain struggles to process and identify the previous sound. The Evening Grosbeak call is commonly used and distinctive enough to stand out each time, but other calls like the Eastern Bluebird rattle simply blend in with the staccato jumble of sounds.

In my yard the same sounds were used repeatedly day after day, so as I became familiar with them I was better able to pick them out, and started to notice other copied sounds. I could not distinguish individual siskins, so I don’t know how many sounds were shared among birds, but the fact that the Evening Grosbeak call and American Robin call were heard virtually every time I listened carefully (and were heard years earlier in BC as well) suggests that most or all siskins use these calls, and probably share a lot of other sounds as well.

If you’d like to hear some of this, check out the recording catalog number 133352 at Cornell's Macauley Library, July 2, 2007, recorded in Newfoundland by Geoffrey Keller. [You’ll have to Search audio for Pine Siskin, and then find this recording]

The bird starts singing at about the 40 second mark, and between 47 and 56 seconds imitates Common Redpoll, Evening Grosbeak, Hairy Woodpecker, Dark-eyed Junco, American Tree Sparrow and American Robin (and probably others). Notice the clear sing-song notes of American Tree Sparrow at about 55 seconds, with the abrupt “pup” of American Robin following immediately after.

Amazingly, this vocal copying seems to have been almost completely overlooked. In Bent’s life histories (online here) copying of Evening Grosbeak and (in a captive bird) Common Redpoll and Canary, are mentioned. A study of vocal copying by Lawrence’s and Lesser goldfinches (Remsen et al, 1982, pdf here) specifically says that the song of Pine Siskin does not include any copied sounds, and the Birds of North America account makes no mention of it.

Presumably the sounds that are learned are simply the ones that are common in the siskins’ environment, so siskins from the western mountains must learn some sounds that are not heard in the east, and vice versa. Another Macauley Library recording from California (catalog number 120288) includes copies of Lesser Goldfinch flight calls, among others. Given how much siskins wander, it would make a really interesting exercise to listen for copied sounds that are “out-of-range”. Hearing, for example, Eastern Towhee in British Columbia or Lesser Goldfinch in Massachusetts would give you some idea of how far that individual siskin has traveled.

Update - in early June 2009 I spent a week at The Nature Conservancy's Pine Butte Guest Ranch near Choteau Montana, and was able to hear several Pine Siskins singing. I was able to hear copies of the vocalizations of seven species there. In addition to the now-expected calls of American Robin, Dark-eyed Junco, Common Redpoll, and Evening Grosbeak, the Montana siskins copied Western Tanager call, Mountain Chickadee dee note, and a Cassin's Vireo song phrase. Clearly these birds had learned their calls in the west, and would stand out in the east. So the challenge remains – to find a siskin that has traveled cross the continent, and recognize it based on the sounds it has learned.

More on identification of Greenland White-fronts

Coincidentally, just days after my last post about the two White-fronted Geese in Concord, two White-fronted Geese showed up in Amherst, Massachusetts (about 60 miles west of Concord, but not the same birds) that appear to be one Greenland and one North American type. James Smith has some discussion and really nice photos on his blog, and he's allowed me to post some of the photos here. 

Two Greater White-fronted Geese - 23 November 2008, Amherst, MA. Photos copyright James P. Smith, used by permission. Click for higher resolution (660 kB). In this montage the upper image shows the apparent North American bird with the Greenland-type on the left. I've resized and pasted in another photo of the Greenland type below for comparison.

This provides a great comparison of the two birds, very well-photographed, in similar light and from similar angles. The Greenland-type bird shows the following differences from the North American-type:

• neck, head and bill larger and bulkier than the North American-type
• neck and head more uniform and darker brownish
  
• scapulars with slightly narrower and less contrasting pale edges
• white border on flank feathers narrower
• pale tips on median and greater coverts narrower and less contrasting, not bright white
• pale edges on tertials slightly narrower and less white

These are all the same differences that I noticed on the two birds in Concord, and that have been described before as identification features for Greenland White-fronts in Europe. One other feature deserves mention:

• bill bright orange (and James says the color is accurate and this bird was truly very orange-billed) which may be useful even though we know how unreliable bill color can be, and the other bird happens to be a fairly obviously pink-billed individual, providing a strong contrast. One thing I noticed on the Concord birds, and again here, is that the North American bird has a fairly obvious whitish tip on the bill, while the Greenland bird does not.

And on 24 November, Taylor Yeager posted photos of another apparent Greenland White-front (scroll through the gallery to see several images of this bird), this one in Sharon, MA about 30 miles south of Concord. This bird doesn't look quite as heavy-billed as the other two, but it still looks fairly thick-necked, the bill is bright orange, and it has the dark head, narrow and low-contrast feather edges on the back, virtually no white on the wing coverts, very narrow white flank stripe, and extensive black on the belly between the legs. I'm comfortable calling this a Greenland White-front also. 

This would all seem to lead to confidence in identifying Greenland White-fronts, and I think we're getting there. But when I browse photos of White-fronts from other areas they show a bewildering range of variation. Identifying Greenland White-fronts will require excellent views and a careful assessment of all the different identifying features, preferably with direct comparison to other White-fronts, and a healthy dose of caution.

Another photo of the two birds in Amherst: Greenland-type in front showing the much larger and bulkier bill and head, and thicker neck, in comparison with the more delicate features of the North American-type bird behind. Photo by James P. Smith, used by permission. 23 November 2008, Amherst, MA.

Finally, one further mystery: It's not at all clear how we can tell that the smaller bird in Amherst is a North American White-front rather than a Eurasian bird. I don't think I've ever seen a discussion of that identification issue, but Eurasian White-fronts show up rarely in Iceland, and could easily wander into North America from there or from Siberia, so it is a potential vagrant to North America. All I can dig up in a cursory search is that Eurasian birds are consistently pink-billed, and may average smaller-billed, and may average less white on the wing covert and tertial edges, but those things are all vague, subjective, and variable. So while I've called the smaller bird in James Smith's photos a North American-type, I can't actually rule out the possibility that it is a Eurasian bird. Any comments or suggestions welcome. 

Identification tips for Greenland Greater White-fronted Goose

The Greenland Greater White-fronted Goose (Anser frontalis flavirostris) is known to be a rare visitor to northeastern North America, but I have never been satisfied with a reliable, objective way of distinguishing it from the North American subspecies of Greater White-fronted Goose, also a rare visitor to the northeast.

Bill color is the most frequently-mentioned field mark – supposedly orange on Greenland birds and pinkish on North American – but this has been refuted multiple times (see Kenn Kaufmann's note in Birding 1994, my blog post here). I suspect this is a mark that works in Europe, where local White-fronts are reliably pink-billed, but not in North America, where lots of the local birds look more or less orange-billed.

Over the last few weeks a single adult White-fronted Goose has been frequenting farm fields near my home in Concord, Massachusetts, and about a week ago it was joined by a second bird (well... "joined" is an overstatement since they pay no attention to each other and keep apart in the big flock of Canada Geese). 

These two individuals, in their details, are really strikingly different, and the differences all lead me to the conclusion that the new arrival is a Greenland bird, and the other a North American bird. On 19 November I was able to get some photos of the Greenland-type bird, and I've annotated a photo here with the differences between it and the North American-type bird. The differences are listed simply from front to rear, not in order of importance. (I'll try to get some photos of the other bird to post here for comparison).

Presumed Greenland Greater White-fronted Goose, 19 November 2008, Concord, MA. Photo by David Sibley

1. bill color slightly more orange, with yellow base (vs. orange with pale pink base on the other bird); but this is very very difficult to judge and essentially worthless as a field mark (see my post from last fall about judging bill color).
   
2. whole head nearly uniform dark brownish, sometimes looking paler on the cheeks but still with solidly dark crown above the eyes (vs. paler gray-brown head and neck with narrow dark crown stripe and contrasting dark border along white front). This is highly dependent on lighting, and would be difficult to judge on a lone bird, but the Greenland-type bird does look consistently darker, and may show a contrasting dark hind-neck that the other bird doesn't, but I need to check that in more lighting conditions.
   
3. neck strikingly thick and heavy, no thinner than the head (vs. neck slender, obviously narrower than the head and looking graceful and slim). This is one of the most obvious differences on these two individuals.
4. margins of mantle and scapular feathers pale brownish, not strongly contrasting (vs. margins whitish, more strongly contrasting).
5. white border on flank feathers narrower and not extending as far forward (vs. broader and longer)
6. pale tips on median coverts narrow and not white, relatively inconspicuous (vs. broader, whitish, conspicuous)
7. greater coverts without pale edges, and with narrow and inconspicuous whitish tips (vs. obvious narrow white edges and broad white tips forming a conspicuous narrow white band across edge of folded wing). This is another obvious difference between these two individuals.
8. black smudges all over the small feathers of the belly between the legs and extending well behind the legs (vs. this area virtually all white). This is separate from the irregular black bands before the legs and up onto the flanks, which are marked similarly in both birds.
9. white tips on the tail feathers seem to be narrower on the Greenland-type, but this is very difficult to see in the field.

Overall the Greenland-type bird appears bulkier, heavy-billed and thick-necked, and darker, without the elegant white fringes that are prominent on the rear upperparts of the North American-type.

Here's another photo of the Greenland-type bird, showing the heavy orange bill, dark and thick neck, and the lack of contrasting edges on the feathers across the shoulders. But none of that is very helpful in a single photo without another bird for comparison, nor is it likely to be very helpful on a single bird in the field. 

I know that some of these things have been described before, and they are still pretty subjective, and it's not safe to generalize from a single individual, but if this is a typical Greenland White-front, then I feel more hopeful about identifying the subspecies.

notes:

Martin Reid's website has some other discussion, photos, and lots references

Lesser Canada Goose in the northeastern US

The primary challenge of distinguishing Canada from Cackling Goose centers on the intermediate-sized 'Lesser' Canada Goose, B. c. parvipes, which reportedly nests across the boreal forest regions of western Canada and interior Alaska (more details on my website here). I've seen a few birds that I thought were 'Lesser' in the northeast over the years, most convincingly one at Brigantine NWR in NJ that accompanied a big flock of Snow Geese rather than flocking with Canadas. But it's a pretty ambiguous subspecies. The identifying features are wholly subjective: anything that seems smaller than the regular Canadas but too big for Cackling is a candidate for 'Lesser' Canada. Recently I saw this bird (the one on the left) that I feel pretty confident is a 'Lesser':

click for a slightly larger image. 

16 November 2008, Great Meadows NWR, Concord, MA - photo by David Sibley. Taken with the 10x zoom on a Canon digital camera. The Canada on the right was a bit farther to the right in the original photo, I cut and pasted it closer for better comparison.

It was distinctly smaller than the Canadas around it, with a very short neck and short bill. In flight going away it appeared no more than 2/3 the bulk of the Canadas next to it. Whenever I scan Canada Goose flocks in the east looking for Cackling Goose, there are always a few birds that grab my attention because they look small. Often they are marginally smaller than the other Canadas, sometimes their smallness was an illusion (happened to be standing in a furrow, next to an unusually large goose, etc). After a closer study these "small" birds fit nicely into the normal range of variation of the local Canadas, and the impression of smallness vanishes. This bird stood out because its smallness held up to closer scrutiny; it was truly small with very short neck and short bill.

I called it a Canada rather than Cackling because it seemed just a little too long-billed, too large overall, and the head/bill too wedge-shaped for 'Richardson's' Cackling Goose. It also looked the same color as the Canadas around it, without the gray sheen on the back or golden cast on the breast that many 'Richardson's' show. 

There are very few reports of 'Lesser' Canada Goose in the northeast. This could be due to the difficulty and inherent uncertainty of identification: we identify only the most clear-cut examples and report those tentatively, if at all. Or maybe it's because few birders really look for them or follow-up on possibles since "it's just a Canada Goose". Or maybe they are truly rare. I have actually seen more 'Richardson's' Cackling Geese in the northeast than convincing 'Lesser' Canadas; but that could be due to any of the reasons above. Given the normal distribution of this subspecies – breeding farther south and west than 'Richardson's' Cackling, with shorter migration – I would expect it to be less frequent than 'Richardson's' on the Atlantic coast, at least in the north. The only way to figure that out is to keep watching for, and documenting, these small geese, in order to develop a better sense of the identification and status of the various subspecies.

So how many do we find?

My last post "How many rare birds do we miss?", was simply getting at the idea that we can miss something glaringly obvious if we are not looking for it. A popular psychology quote goes "I wouldn't have seen it if I hadn't believed it". In the case of the moon-walking bear, since I wasn't looking for it, I didn't see it. 

In the comments a reader pointed to a discussion on his own blog "The Birdist" where he actually tries to answer the rhetorical question in the title of my blog post. So how many rare birds do we actually miss? The Birdist and his interviewee guess that one-third to over one-half of all rare birds are found. I disagree and I would put the percentage much much lower. 

Why do I think so few rare birds are found? Well, for one thing, I look at the very small number that are ever refound. The White-crested Elaenia in Texas last January, or the Variegated Flycatcher in Washington this September, or the European Golden-Plover in Maine in October, or the White Wagtail in the Florida Keys last week, etc etc. all share the fact that none of them was seen anywhere else in North America. Maybe a statistician can comment on the hypothetical numbers involved, but if one-half of all rare birds were being found by birders, I would expect a lot of re-finding of these ultra-rarities. It's likely each of these birds wandered slowly and spent time at many other places where they could have been discovered by birders, both before and after their actual discovery. And there must be countless others that are never found. 

The Jabiru in Louisiana July 31, 2008 could be the same one that was seen in southern Texas August 10th and 20th. But that is a big, conspicuous species that even a non-birder will notice. In fact, as I understand it, it was found in Louisiana by a local hunter who noticed it and took a few snapshots. It was refound by a birder in Texas on August 10th (if indeed the same bird) but then went missing for 10 days, reappearing on August 20th! Add to that the experience in August 2007 when a Jabiru showed up in Mississippi. Birders were alerted to it by the local farmers, but after a few days that bird vanished and was never seen again, despite intensive searching and the fact that its most likely path back to Central America would have taken it through Louisiana, past Houston and Corpus Christi, and over several hawkwatches. Rare birds being refound at a distant location are the very rare exception to the rule. 

Another interesting bit of evidence is what I will call "delayed discovery". There are many examples, as in the 2007 Mississippi Jabiru that was seen by birders only because the landowner was curious and sent out a photo. It turns out the bird may have been in the area for several weeks before that. Or the Least Grebes recently discovered in a park in Boca Raton, Florida. On September 27th self-described "recreational birders" Lee and David Hasse sent an email alerting other birders that "last Sunday" they saw two Least Grebes. This was confirmed by their photos and by "serious birders" the next day. The following day (28th of September) the nest was discovered, and on the next day (29th) the first egg hatched!

The eggs must have been laid about 3 weeks before they hatched, and nest-construction must have started at least a week before that, and probably several weeks for pair formation, etc. So those two grebes must have been at that park for at least 6 weeks, since mid-August. And if the Hasse's weren't there in late September, or hadn't taken the time to report their sighting, would the entire nesting have passed without birders noticing?

So I'll take on the question now: How many rare birds are actually found? At a place like Cape May in the fall, arguably THE most intensively birded location in North America, a Jabiru is almost certain to be seen and identified, I'd say 80 or 90% of the time, and if it sticks around for more than an hour or two the odds go up to near 100%. As we go down the scale of conspicuousness, Fork-tailed Flycatcher is fairly likely to be found (virtually all birders and even a lot of nonbirders would notice one, even from a great distance); Western Tanager is less likely to be found (much less conspicuous - you have to be within a few hundred feet to notice one); Smith's Longspur is even less likely (secretive and nondescript), etc. The odds of finding a Smith's Longspur that only stays for a day are very very low overall (although if it spends that day on the lawn by the Cape May hawkwatch, it's very likely to be found). Staying longer will increase the chances of being found, and so on.

Psychologically, birders at places like Cape May are somewhat more likely to find rare birds simply because, at known vagrant traps, we expect to find rare birds. The same observation that might set a Cape May birder running after a "possible Smith's Longspur" is more likely to be dismissed as "not worth the effort" by a birder in farm fields in upstate New York. And as soon as a historical pattern develops at a location - like Cave Swallows at Cape May in November - the tables turn and birders start actively looking for a specific rare bird.

Back to my point about birds not being refound: if each of these rarities is spending 10 or 20 or 30 days wandering around inside North America, but is only seen once, that's something like a 10% or 5% or 3% chance of being found, and that strikes me as the right range to consider for how many rare birds we actually see.

As a final note, I like to think that this makes me an optimist, since I can always convince myself that the European Golden Plover from Maine is going to show up in my local patch tomorrow, and that there must be loads of rare birds hiding in the next bush.

Variation in Wild Turkey

Here's an example of one of the countless thousands of questions that are still out there waiting to be answered about North American birds. I went to Texas in February (for the opening of the fantastic new Clif Moss Nature Education Center of the Corpus Christi Public Library) and I had a chance to see some Wild Turkeys nearby. As expected they had whiter tips on all the body feathers than northeastern birds, but they also seemed thinner and smaller-headed than the Turkeys that frequent my backyard in Massachusetts, and when I got home I immediately saw what the difference was.

sketch copyright 2008 - David Sibley

The Massachusetts Wild Turkeys have feathering extending all the way up the neck, while the Texas Turkeys have a long stretch of the upper neck bare (here's a quick sketch I did that first morning at home). This is logical, presumably an adaptation to the cold weather in Massachusetts, but in a quick review of some sources (the BNA account) I found no mention of neck feathering as a geographic difference. I thought maybe it was just a winter feature; maybe the Texas birds had already lost their "winter coat", and the Massachusetts birds would lose their neck feathers for the summer. They may still do that, but as of today (2 April 2008) they look the same.

As usual, this just brings up more questions: Is this going to be true year-round? What are the patterns of variation? Do all southern birds - Florida, Texas, Arizona - lack neck feathers? Does it change gradually from north to south or is the change more abrupt? and lots more....

Audubon's mysteries: Carbonated Swamp-Warbler

One of the enduring mysteries of North American ornithology involves several species which were painted by Audubon in the early 1800s but never seen again. The most striking and appealing of these birds is the Carbonated Swamp-Warbler, and since the painting was published ornithologists have debated whether this could be a rare and now-extinct species, a hybrid, or merely a fantasy created by Audubon.

Audubon wrote:

I shot the two little birds here represented, near the village of Henderson, in the State of Kentucky, in May 1811. They were both busily engaged in searching for insects along the branches and amongst the leaves of a dogwood tree. Their motions were those common to all the species of the genus. On examination, they were found to be both males. I am of opinion that they were each young birds of the preceding year, and not in full plumage, as they had no part of their dress seemingly complete, excepting the head. Not having met with any other individuals of the species, I am at this moment unable to say anything more about them. They were drawn, like almost all the other birds which I have represented, immediately after being killed; but the branch on which you see them was not added until the following summer.

Despite his definitive statement that he shot two specimens, and painted them from life (well... death), there has always been speculation that Audubon may have invented these birds, or painted them from memory rather than from specimens. Audubon frequently stretched the truth, and many of his untruths are well-documented, although as Jonathan Rosen suggests in a New York Times book review they are "more like the improvisational ''stretchers'' of Huck Finn than the calculated inventions of a man on the make".

While viewing some of Audubon's original paintings at a New York Historical Society exhibit several years ago I noticed that the paintings ranged from very lovingly detailed (e.g. Carolina Parakeet) to more cursory and simplistic (e.g. Bicolored Blackbird and others painted from specimens brought back by Townsend). It occurred to me that the painting style and quality of details might provide clues to Audubon's mystery birds. The recent launch of a complete collection of high resolution scans of Audubon's plates at the University of Pittsburgh (available here) prompted me to take a closer look at this painting.

A brief survey of the painting reveals several factual errors:

• Dark stripes on the back of the lower bird are pointed at the front, broader and diffuse at the rear - opposite of the normal feather arrangement on songbirds
• Median secondary coverts (the shorter wingbar) should have the lower feathers overlapping those above; both birds have these feathers overlapping incorrectly
• Median secondary coverts on the lower bird are too long
• Too much yellow shows in a broad wedge above the forewing on the lower bird, unlike any known songbird; there is no way the feathers painted on the upper bird could be rearranged to look like the lower bird
• Uppertail coverts of the upper bird are shaped and arranged incorrectly
• There is a general lack of detail, a sort of vagueness of structure to the wingtips and tails, both birds lack primary coverts (which should be visible) and the tails are unusually short with very narrow feathers (unlike any known wood-warbler).

One could give him the benefit of the doubt and allow that these could all be due to carelessness or inexperience. I can imagine that there must have been times when he was distracted from his painting by biting insects, weather, hunger, illness, or other concerns. But checking some of his other early paintings (Chestnut-sided Warbler plate 59, Cerulean Warbler plate 49, Purple Finch plate 4, Song Sparrow plate 25, etc) shows that these are all more detailed, with correct arrangement of coverts and streaks. There are always things to nitpick, but in each of these paintings the overall draftsmanship is painstakingly accurate, and smacks of authenticity. The Carbonated Swamp-Warbler painting is subtly lacking, which leads me to conclude that Audubon was probably not looking at a bird as he painted.

This is all merely speculation and circumstantial evidence - weighing possibilities. There is no way to know for certain what Audubon painted (unless the Carbonated Swamp-Warbler were to be rediscovered and confirmed). It's possible that this species did exist, a hyper-specialized wood-warbler, like Kirtland's, that was extinguished by the first clearing of forests in the early 1800s. It's possible that Audubon did see them, and simply painted them poorly. Or that he saw one and tried to recreate it from memory some time later (lying about the two specimens to bolster the credibility of his painting). But given that the only evidence of the species' existence is Audubon's painting and written account, that evidence needs to meet high standards. I think the quality of the painting casts some doubt on his claim that he was working directly from two specimens, and then that casts doubt on the existence of the species.

Greater Redpoll photos

Maybe I should change the name of this blog to "All about Redpolls", but I've received a few photos that I wanted to pass along, making the point that "Greater" Common Redpoll is not just an eastern specialty. The AOU checklist and the BNA account report that this subspecies winters regularly from Labrador west to Manitoba, and has been recorded from Michigan, Illinois, Indiana, and Ohio, as well as Colorado!

So here are two photos from Seth Cutright of Ozaukee County, Wisconsin showing an apparent "Greater" (on the upper right). Notice the slightly larger size, larger bill, dark color, and heavy streaking especially on the nape, flanks, and throat. If this is really a "Greater", it would apparently be the first for Wisconsin, although that may not be too surprising considering the distribution of other records.

Photo copyright Seth Cutright, used by permission - click to enlarge

And (below) here it is again on the top right, turning away but showing the large size, heavy flank streaking and dark undertail coverts.

Photo copyright Seth Cutright, used by permission - click to enlarge

Cathy Mountain of Fort McMurray, Alberta also sent me a series of photos from her feeder, and among them is another apparent "Greater" Common Redpoll (far right). The large size, dark color, and heavy streaking is pretty obvious. This would apparently represent a first for Alberta, but that should not be too surprising if the subspecies really winters regularly in Manitoba.

Photo copyright Cathy Mountain, used by permission - click to enlarge

So I would urge anyone who sees Common Redpolls to watch for larger birds and try to document potential "Greater" redpolls. This subspecies appears to wander widely to the west of its breeding range, and could seemingly turn up anywhere within the winter range of Common Redpoll (at least west to the Rocky Mountains).

Another note on Hoary:
Trained observers will notice that the photo above shows several Hoaries. In fact it was Cathy Mountain's photos from Alberta that prompted me to check CBC results for my post about winter range of Hoary Redpoll. Fort McMurray is obviously within the core of the winter range of Hoary, as the following photo shows.

Photo copyright Cathy Mountain, used by permission - click to enlarge

I guess there are at least eleven Hoary Redpolls in this picture! (when I said "ten" yesterday that was just an error in addition). The seven in the foreground and one back left are so white they are pretty unambiguous - the kind of Hoaries that are seen only rarely east of the Northern Plains region. (I'm also counting the two central birds with tawny head and faint flank streaking as Hoary, as well as the partial bird visible at the top left.)

Notes:
More of Seth Cutright's redpoll photos

Greg Sargeant reports that there is a "Greater" Common Redpoll in Rhode Island, photo here, which may be the first ever noticed in that state.

Another large pale redpoll

A new photo by Dan Berard shows what looks like an intermediate large redpoll, and should be added to the discussion in my previous post.

photo copyright Dan Berard, used by permission - 7 Mar 2008, Millbury, MA

Dan says both of these birds were about the same size and both were larger than the other redpolls with them. This is interesting on its own because the paler bird looks larger in this photo. Assuming that Dan's observation is a more reliable measure of size, this shows the danger of putting too much emphasis on a single photo when judging size. [update 11 March - Dan points out that the pale bird is at the near edge of the railing, and the dark bird at the far edge, so that could account for the apparent size difference in the photo.]

The bird on the right is dark and heavily-streaked, approaching the "classic" rostrata, with heavy streaks on the breast and flanks. The bird on the left is distinctly paler and Hoary-like, with the pale gray ground color of the head and back, whitish flanks with fairly narrow streaks, etc. If the bird on the left was smaller, I would study it as a potential exilipes Hoary, but it has some odd dense fine streaking around the neck and breast, and the flank streaks are unusually broad for Hoary. It seems too pale to be rostrata, too dark to be hornemanni. Which is it?

Another note: This photo (with two large birds together) demonstrates the point that redpolls seem to travel in more or less cohesive groups within the big flocks - what we might call "subflocks". This is similar to the way geese sort out within a big flock, but the redpoll subgroups are not quite as cohesive. There have been quite a few reports this winter of small groups of redpolls with a high percentage of hoary-types (like 100%). And at Dan Berard's feeders the large birds seem to show up in waves - none in a group of fifty, and then 8 in a group of 30. What this means for identification is that the birds immediately adjacent to each other are more likely to be similar, and that comparing an "interesting" bird to those immediately around it might not be the best measure of how different the interesting bird is from the "typical". You should try to compare an interesting bird with other redpolls across the entire flock, and don't be surprised if studying one interesting bird reveals the presence of several more.

Update 11 March 2008: I wanted to add two more general cautions suggested by Dan Berard.
1) One of the things that makes size difficult to judge is the fluffiness of redpolls. Birds that are relaxed, with feathers puffed out and sitting more upright, can look larger than adjacent birds that are actively feeding. Be sure to consider what the bird is doing and watch a bird for several minutes if possible to get an accurate impression of size.
2) When redpolls fluff their feathers they reveal more white, making the dark flank streaks relatively less obvious. This is particularly obvious where the upper flank feathers overlap the folded wings, and these feathers can also be displaced and stick out over the wing just after a bird has landed. Dan says his eye is sometimes drawn to a flash of bright white, only to discover that it is not a pale redpoll but simply a redpoll with feathers fluffed or a few feathers askew.

Redpoll investigation widens to include "Greater"

Maybe we are just more aware and looking harder for "Greater" Common Redpolls (Carduelis flammea rostrata) this winter in Massachusetts, or maybe it's really an exceptional winter, but there is no doubt that they have come south in significant numbers. I have seen them on a couple of visits to Dan Berard's feeders in Millbury, MA, where he has been studying redpolls almost daily since December. This has confirmed yet again Yogi Berra's comment that "You can observe a lot by just watching".

This subspecies is relatively poorly-known, nesting on Baffin Island and Greenland and wintering irregularly south to the northeastern US. Literature on identification of this subspecies (Beadle and Henshaw 1996; Herremans 1990, Knox 1988, Pittaway 2007) characterizes it as a consistently large and dark bird. Based on those studies and comments from Dan Berard I've put together a list the features that distinguish rostrata from "Southern" Common Redpoll C. f. flammea, listed in order (my judgment) from most important to least important:

• averaging 10% larger and about 50% heavier
• Generally more heavily streaked below (more streaks and wider streaks); blurry streaks extending along flanks to undertail coverts, which average darker than flammea; at least some individuals are more heavily streaked than any flammea Common. Some have the flank streaks converging into a solid dusky patch at the rear end of the flanks.
• Some in MA (as on most specimens here) have a complete necklace of fine streaks below the dark throat patch, which is very distinctive if present.
• At least some have whole face darker, more streaked, with only faint pale eyebrow stripe, and densely streaked malar. These are probably the same birds with the necklace of streaks, very heavy dark flank streaking, and rich brown color, but even on some of the less obvious individuals a general dusky-headed appearance is evident.
• usually appears longer-tailed
  
• bill may appear thicker with curved culmen
• darker and more richly-colored brown overall
• often looks thick-necked and "front-heavy"
• some have more extensive dark throat patch
  
• male typically has less red on breast and lacks red on malar
• Dan Berard reports that he often hears distinctly different "harsher, more rattling" calls when Greaters are present, although he hasn't yet confirmed that these sounds come from the Greaters.
  

Here's a photo of a flock of redpolls with several large dark birds:

photo copyright Dan Berard, used by permission - click to enlarge. Sharp-eyed readers will notice that the top right bird is quite pale, but unfortunately we can't do much more with that.

Three rostrata "Greater" redpolls are fairly obvious here (blue arrows), being larger and darker than the birds around them. And several others are possibly "Greater" but not quite dark enough or large enough to be identified confidently. Features that seem to be especially useful for picking out these birds are the darker head and broader, blurrier streaking on the flanks than other redpolls, but this is not always consistent. Size is the most consistent feature for identifying this subspecies, but one of the first things you notice when you start looking for larger redpolls in a flock is that size is surprisingly difficult to judge (impossible without direct comparisons) and there seem to be some birds that are intermediate in size. Measurements (e.g. in Herremans 1990, Knox 1988, Knox and Lowther 2000) confirm that rostrata and flammea overlap in all measurements, even though the average is significantly different. So I guess it's not surprising that we would see a full range of size from small to large, and it leaves the question - "when is a bird large enough to qualify as rostrata?"

A side note on psychology and perception - This leads to yet another manifestation of the ambiguity issues that I wrote about previously for White-fronted Geese. Because redpoll identification involves subjective judgment of ambiguous variations in color and size, the first few redpolls you look at can set a standard that persists as you scan the rest of the flock. You can scan a flock once and find a dozen that look like candidates for 'Greater' (or Hoary), then scan the same flock seconds later and find no good candidates!

The other interesting observation about rostrata 'Greater' Redpolls is that they may not be as consistently dark as the literature leads us to believe. Here's a male that is quite pale and grayish, not the heavily-streaked and richly-colored plumage that is considered typical of "Greater". But given that it's an adult male the overall color is within the range of Common Redpoll, and the flank streaking pretty heavy for an adult male.

photo copyright Dan Berard, used by permission - click to enlarge. The bird just behind the "Greater" in the center (facing left) is a possible Hoary, so that should be considered when comparing colors.

Among the large redpolls in Millbury, we've seen a few large birds (non-adult-males) that are paler than the typical flammea Common redpolls (in at least some respects).

Here's a picture of one such bird:

photo copyright David Sibley - 22 Feb 2008. click to enlarge

This bird is one of several seen that day that were obviously large (larger than the goldfinch), and that also stood out among the flock of Commons by their pale overall color, with pale and grayish back and pale head. It also had a pale, but faintly streaked, rump. One thing that still points toward rostrata "Greater" Common is the broad blurry streaks on the flanks, and presumably the undertail coverts were also streaked, so maybe this is within the normal range of variation of rostrata, but the overall paleness is not what I expected.

And here's an even paler bird (I wish I had seen this one in life):

photo copyright Dan Berard, used by permission - click to enlarge

This picture may include three subspecies: rostrata "Greater" on the left, a good candidate for hornemanni Hoary in the middle, and an ordinary flammea Common on the right. The middle bird seems about as large as the rostrata on the left, and larger than the flammea at the right. It's very pale overall, and compared to the pale rostrata-types such as in the previous photo it has a paler nape with less distinct streaks, white flanks with narrow dusky streaks, mostly white undertail coverts, smaller bill, and smaller black throat patch. It's just not quite as white and unstreaked as I expected the "classic" hornemanni to be.

Conventional wisdom holds that rostrata "Greater" Common Redpolls are large and dark, and that hornemanni or "Hornemann's" Hoary Redpoll are equally large and very pale. So I have always assumed that the two would be easier to tell apart than the smaller "Southern" flammea Common and exilipes Hoary Redpolls. But the large birds we've seen in Massachusetts this winter do not all fit the model. At the feeders in Millbury the large birds show a wide range of color variation, from very dark (classic rostrata) to quite pale (but so far none of the very pale classic hornemanni color).

Are rostrata "Greater" Commons more variable than previously thought with some relatively pale individuals? It appears so. But how close can they get to hornemanni Hoaries? And are hornemanni just as variable? Knox (1988) says that female and immature hornemanni can match exilipes in plumage, so can some individuals be relatively dark like the darker Hoaries we've been talking about this winter? Knox (1988) mentions one mixed pair of rostrata and hornemanni from Greenland. He found no other convincing evidence of hybridization, but it is still possible that these two large subspecies are interbreeding, producing a whole range of intermediate plumages. Any or all of these scenarios are possible. The only thing that seems clear is that the plot has thickened, questions abound, and identifying these large redpolls may be more challenging than we thought.

Thanks to Dan Berard for permission to use his photos, for hosting the birds and the birders, and for discussions about redpoll identification.

References:
Beadle, D. and B. Henshaw. 1996. Identification of “Greenland” Common Redpoll Carduelis flammea rostrata . Birders Journal 5: 44–47.

Herremans, M. 1990. Taxonomy and evolution in redpolls Carduelis flammea-hornemanni ; a multivariate study of their biometry. Ardea 78: 441–458.

Knox, A. 1988. The taxonomy of redpolls. Ardea 76: 1–26.

Knox, Alan G. and Peter E. Lowther. 2000. Common Redpoll (Carduelis flammea), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America Online: http://bna.birds.cornell.edu/bna/species/543

Pittaway, R. 2007. Redpoll Challenge: 4 subspecies. Online http://www.jeaniron.ca/2007/Redpolls/redpolltext3.htm

Where do Hoary Redpolls winter?

In considering the true status and distribution of Hoary Redpoll, I realized that I had overlooked one significant source of information - the Christmas Bird Count results. While Common Redpoll nests throughout the boreal forests across the continent, Hoary is virtually unknown as a breeder east of Hudson Bay (ignoring the "Hornemann's" Hoaries in Greenland and Baffin Island). Therefore the source of virtually all Hoary Redpolls seen in North America in winter must be northwestern Canada and Alaska, and one might expect that Hoary would be less common eastwards in winter. The CBC results show just such a pattern.

Hoary (right) and Common (left) Redpolls recorded on CBCs in two "invasion" years representative of recent CBCs, 1993-94 and 1999-2000.
From National Audubon Society (2002). The Christmas Bird Count Historical Results [Online]. Available http://www.audubon.org/bird/cbc [4 Mar 2008]

There may still be a lot of identification issues with these data, and the coverage in most of Canada and Alaska is spotty or nonexistent, but one pattern seems clear: The proportion of Hoary to Common redpolls is much higher in the west (Alaska, Alberta, and Saskatchewan) and decreases eastwards.

Redpoll age variation and ID

The question of age variation has come up repeatedly in redpoll discussions, so I finally tried to find an answer to the question of just how important age-related variation in plumage might be. And based on published studies the answer is ... not very important.

Seutin et al. (1992) studied redpolls at Churchill, Manitoba using a character index, and found only one statistically significant age-related plumage difference (aside from the obvious red/pink breast of adult males) - flank streaking on adult males was lighter than on any other age/sex class. All other features - undertail covert streaking, rump color, forehead color, poll color, etc. were statistically the same on adult male, immature male, adult female, and immature female. This near-complete absence of age-related variation differs from what Knox (1988) reported - that in European redpolls immatures are darker than adults. I haven't read the Knox paper, so I don't know how different the results actually are, but regardless of what he found in Europe, the paleness of a redpoll in North America does not seem to be linked to the bird's age, and therefore ageing a redpoll is not helpful to the identification process.

The lack of age-related variation does make sense given the lack of differentiation in redpolls at all. Going back to the species question: Seutin et al. analyzed their data on plumage variation by searching each age/sex class for any sign that the distribution of scores was bimodal; e.g., did each age/sex class show any hint of sorting into paler and darker groups on any of the measured plumage characters? Only adult males showed a tendency to sort into two groups, and that was slight. Immature males showed a suggestion of sorting into pale and dark groups only when sophisticated statistical tests were applied to the data. Attempts to sort females of either age class into pale and dark groups failed completely.

Combining this with Troy's similar finding that redpolls form a continuum from pale to dark, and Troy's data showing that intermediate-colored birds were also intermediate in skeletal measurements, implies that observers should expect to find lots of intermediate birds that cannot be confidently assigned to a species.

It's possible that characteristics not measured by Seutin or Troy - like scapular color, or forehead fluffiness, or voice, or something else (?) - might lead to better separation of two groups, but their inability to distinguish two clear groups does not bode well for identification of non-adult-male redpolls.

Urging caution when identifying Common Redpolls

Redpoll identification is challenging because Hoary and Common Redpoll seem to show an unbroken continuum of variation from pale to dark, and there are no fully reliable differences. So birders have to rely on a subjective assessment of overall color and struggle to define the threshold for confident identification. Virtually all birders see redpolls only in the winter, where identification is reduced to an utterly one-sided question: "Where can we draw the line so that we are sure the accepted records of Hoary Redpolls are correct?"

Since the goal is high confidence, this leads to a narrow definition of Hoary Redpoll - only the palest (and smallest-billed) birds are named and reported as Hoaries. This leads to two problems (the second more serious). First, we only identify part of the population - some darker Hoary Redpolls are excluded. Second, many of those intermediate birds are simply lumped into our broad definition of Common, and receive no extra attention. There's nothing wrong with excluding some intermediate birds from being called Hoaries, as long as birders understand that some Hoaries are being excluded. This is conservative. But most people give redpolls very uneven treatment by demanding that Hoaries meet high standards, and then calling everything else Common. This is not conservative. We should at least be fair and apply equally strict criteria to our Common Redpoll identifications.

By excluding some darker Hoary Redpolls (and calling them Commons) this uneven treatment prevents us from understanding Hoary Redpoll as a population.

Some might argue that this approach is correct, but it is not supported by Troy's (1985) research. Troy found that all redpolls breeding in taiga/forest habitat were uniformly in the darkest one-third of all redpolls he studied (with no intermediates at taiga sites). Redpolls breeding at tundra sites, on the other hand, were primarily in the paler two-thirds (from intermediate to very pale). If we start with the hypothesis that Common Redpoll is a dark southern population with its core nesting range in the forest and Hoary is a paler tundra-breeding population, then the intermediates, which Troy found only in the tundra habitats, seem most likely to be either hybrids or darker variants of Hoary. Calling these intermediate birds Commons would imply that Common Redpoll is represented by dark birds in the taiga and paler birds in the tundra. This seems less parsimonious and, if true, suggests that we should consider the possibility of two subspecies of Common Redpoll along those lines. I don't know of any research that would help resolve the status of these intermediate birds.

So we need two different approaches to identification. The traditional conservative approach is fine if we acknowledge its limits and we are just trying to be confident about adding Hoary Redpoll to a list. But if we are trying to understand Hoary Redpoll as a species, we need a more inclusive definition that will identify the whole population. In the short term, I think we need to use the "intermediate" label more, calling lots of birds "intermediate" until we get a better understanding of where to draw the line.

I realize that the birds in my redpoll photos are outside of the traditional "safe" zone for Hoary Redpoll, but it's misleading and obscuring to say that these birds are just "pale Commons". I am convinced that the birds I photographed are not just Commons and they presumably come from tundra habitats where Hoary nests. Whether they are good Hoaries or hybrids/intergrades is the next question. But we can't even begin to answer that question if we just dismiss them as non-Hoaries. We need to spend more time studying these intermediate birds, and resist the urge to stick arbitrary labels on them, if we are to have any chance of answering the bigger questions about redpolls.

Notes:
1) Matt Sharp suggested that the differences between two of the redpolls I photographed and called Hoary-type and Common, are simply age-related. I have very little experience judging tail feather shape so I don't want to attempt ageing, but if there is a difference between these two birds I would reach the opposite conclusion to Matt: it looks to me like the paler bird has more pointed outer tail feathers than the darker bird and is therefore more likely to be a 1st-winter, ... but I'll defer to banders' opinions on feather shape!

Regardless of feather shape, I can't accept that age alone explains the plumage differences. This pale bird was one of two or three similar individuals among 75 to several-hundred Redpolls I saw that day. These pale birds stood out subtly but distinctly from the darker brownish Commons. There was certainly a range of variation in color among the Common Redpolls, but these few birds still stood apart. They were paler overall than any female-type Common, with a "colder" gray tone, had slightly finer and paler streaking on the flanks (index score 3+), mostly white rumps with just a few dusky marks (index score 4+), less distinct streaking on the head and neck, and less brown on the back. And they often gave the impression of a slightly broader (fluffier) neck and fluffier forehead than the typical Common.

2) James Smith has posted photos of some other pale redpolls he saw along with the previously-reported Hoaries on Cape Ann; and Marshall Iliff photographed one obvious and five probable Hoary Redpolls on Cape Cod last week. In both cases these are birds that, like my New York photos, stand out among Common Redpolls but are not quite pale enough to reach everyone's threshold of "safe" Hoary. So let's call them intermediates until we have a better understanding of variation. I think they're closer to Hoary than to Common, so it would be wrong to call them Common with no further discussion. If anyone with more redpoll experience is willing to offer an opinion on any particular bird the discussion could be very helpful.

For comparison, here is a Swedish bird carefully-diagnosed as a 1st winter Hoary (so I'm not the only one calling such dark birds Hoaries).

3) On bill size as a distinguishing feature of redpolls: look at this page of variation in bill size of European Hoary Redpolls (scroll down for Hoary). Here is photographic evidence that bill size is not very useful for identifying Hoary and Common Redpoll, and this answers the question I posted previously about bill shape - Hoary can have a bill just as slender as the typical Common.

4) I agree with Ron Pittaway's point on ID-Frontiers that Redpoll identification is especially challenging in photos, since the differences are so slight and are strongly affected by lighting and image processing. The most useful photographs are often those with multiple birds rather than full-frame portraits, and I would urge photographers to try to include several redpolls in each image so that any Hoary candidates can be compared to adjacent birds under identical conditions.