Friday, March 28, 2008

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.


David Sibley said...

Scott Weidensaul comments:

An interesting new take on one of the enduring mysteries of American ornithology. But I would raise one caution -- remember, you are not analyzing Audubon's art, but two other artists' interpretation of his art -- in this case, the craftsman at Robert Havell's shop in London who engraved the plate in 1829, and the colorist who then tinted it.

The only handy reference to the original watercolor (aside from asking pretty please at the New-York Historical Society) is the 1966 two-volume set, "The Original Water-color Paintings by John James Audubon for The Birds of America," published by American Heritage.

Looking at the image of the original watercolor, you can see that, while several of your criticisms are valid, several "mistakes" are clearly artifacts of the engraving process. For instance, the watercolor seems to show a more typical, if poorly defined, uppertail covert structure on the upper bird, which I think the engraver misinterpreted. It's also clear that the large area of yellow on the lower warbler isn't part of the wing, but rather neck feathering. The flight feather structure on the lower bird is much better defined, and more typical of warblers.

On the other hand, the mistakes in wing covert shape and overlap are as you observed, the rear body structure on the lower bird is crude, and the tails are undeniably short for a wood warbler. Ditto the problem with the back streaks, which are the reverse of what's typical for most warblers.

Audubon claimed to have shot the birds in 1811, and we know that in 1812, rats destroyed almost all of his stored paintings, forcing him to redo many of them. While this was fortunate in the long run, since his later works were much stronger artistically, he may have been forced to redo some from memory - and it seems likely this was one of them.

Like you, I doubt he had a specimen in hand -- but does that mean he hadn't killed a couple of intriguingly odd warblers in Kentucky in 1811? I'm willing to give the old guy the benefit of the doubt.

David Sibley said...

Hi Scott, thanks for your comments. You make a good point about the engraving being a couple of steps removed from the original. It sounds like the engraving is pretty faithful to the original, but of course the next step in this research would be to examine the original painting, and that could reveal other clues as well.

I wasn't suggesting that the large yellow patch was supposed to be part of the wing, just that the feathers above the bend of the wing should always be the scapulars and mantle, which are shown dark and streaked in the upper bird, so there's no way for a big yellow wedge to appear there unless some birds have yellow scapulars (which would be really odd).

The fact that he recreated some paintings destroyed by rats in 1812 is another really interesting bit. Maybe this painting is a replica of one he remembered. That could explain the anatomical flaws, but then how much faith can we have in his memory of the plumage?

In the end, this is all speculation, and it's possible that Audubon did see a couple of odd warblers that were the last of their kind. I'd just feel a lot better about that if this painting was as detailed and accurate as his other paintings.

cyberthrush said...

I believe there can be such a thing as analyzing something to death... I've never really much cared for Audubon paintings because compared to say the later work of Fuentes, they seem primitive, almost childlike, to my non-artist eyes. MANY of them, I think, show inaccurate proportions, colors, postures, anatomical lapses, or over-simplicity. But I've never had any doubt that this painting represents 2 birds Audubon actually saw and had in hand at some point -- and in fact I think it has significant detail in it, if you don't deliberately focus on the flaws; personally, I always wrote these birds off as flukish hybrids or mutations, but we'll likely never know. Though fun, it seems a little risky, to try to second-guess this far in the future the ol' guy's veracity here, and like Scott, with John James not around to defend himself, I'd give him the benefit of the doubt (...until a long-lost journal shows up saying it was all an April Fool's joke! ;-)

David Sibley said...

Hi Cyberthrush, I guess if you have no doubt that Audubon saw these birds there's no point in trying to open your mind to other possibilities ;-) but I'll restate my case anyway.

This painting does have quite a bit of detail, but it's lacking the sort of precise, intimate details that Audubon put into nearly all of his paintings (and that must have come from close study of the actual birds). One of the characteristics of Audubon's work is that he seemed to be fascinated by the shapes and patterns of tails. He usually posed his birds with tails twisted and fanned in artful ways, and then he drew the tails carefully. So why are the tails of these two Carbonated Swamp-Warblers so vague and boring?

Maybe he was looking at specimens, and was just uninspired that day. But I think the fact that this painting is one of his less detailed efforts, with errors and omissions of a sort that I would expect from drawing without a model, leaves open the possibility that he was working from memory.

This is still just speculation that can't prove anything, but as we weigh the possibilities I don't think we should ignore it.

Another angle: Audubon writes that he thought these two birds were immatures, and his description could lead to the conclusion that these were fledglings in transition from juvenal to 1st-basic plumage. He mentions short tails, fluffy, formless feathers, and plumage not fully-developed except on the head. The date (May) is early for fledgling warblers, and I can't think of any species that fits, but we should expand our search to include the possibility of juveniles and not just breeding-plumaged males.

cyberthrush said...

Hi David, you're the professional artist, so I respect that what you see as major differences in Audubon's works (and barely register on my untrained radar) certainly MAY mean something here. I'm simply worried that there are way too many variables we can't know about 200 years later... I didn't even know of the early destruction and re-do of much of JJA's work, 'til Scott mentioned it. Or another simple variable: presumably the 2 birds were shot; I would think birds THAT small being shot could significantly alter their true appearance when one goes to paint them; maybe that can account for some (not all) the oddities you find (yes, I realize it would apply to all his warblers).
Speculation is fine, and Audubon was certainly known to engage in exaggerations or falsehoods on occasion (usually when it served some specific purpose), but I just think the default position here ought be to accept Audubon at his word unless more solid evidence says otherwise. Frankly, it is just ALL-too-easy to speculatively cast doubt on the claims of individuals, whatever they may be ;-)
Many yrs. ago Stephen Jay Gould (who I admire BTW, R.I.P.) drew up a speculative case against Teilhard de Chardin as instigator of the Piltdown hoax -- a case that I believe has been largely demolished since, but it demonstrated the ease with which such can be done at a distance.

(p.s. just to insure that I don't confuse anyone out there, my "Fuentes" reference in 1st comment was meant to be Louis Agassiz Fuertes, NOT FueNtes.)

Louis Bevier said...

With all due respect, I think both Scott Weidensaul and “cyberthrush” focus too much on the messenger, David, and therefore overlook some valuable lessons. Many birds are known only from “types” based on illustrations alone. It has long been important to evaluate these. The details of the description and the associated plate are frequently used to located missing specimens or series, often resolving long-standing taxonomic mysteries. Clem Fisher, daughter of well-known ornithologist James Fisher and curator of birds and mammals at the Liverpool Museum, has made a career of this, with several publications based on careful and critical review of early illustrations, showing their taxonomic significance. The digitization of many of these old, rare plates that usuallly reside in rare book collections inaccessible to most, is a great advance. I like the idea of David exploring these. Another collection of important digitized plates can be found at the Academy of Natural History's (Philadelphia) library. It is fascinating and valuable to explore these early naturalists’ writings and paintings.

Fresh review instead of "hands-off" reverence, as advocated by Scott and "cyberthrush," seems most worthwhile to me. David's inquiry stands as something opposed to the usual quick embrace of only the most current information and digs deeper. In an interesting twist to this Carbonated Warbler story, it was Sir William Jardin who warned about this practice in his preface to a new edition of Alexander Wilson's American Ornithology in 1832. He said, "In looking at our present knowledge of the natural history of any vast country, we generally lose sight of a very important circumstance." Jardin feared that Americans habitually undervalued the work of "early naturalists, and their sources of information." Alexander Wilson's reputation had fallen with the publication of Audubon's work, and Jardin wrote a wonderful biographical sketch to Wilson for the updated 1832 edition of Wilson's work. Jardin added and annotated many of the birds published by Audubon subsequent to the first publication of Wilson's "American Ornithology," and Jardin was probably the first to report the Carbonated Warbler as most likely just a poorly remembered, immature male Cape May Warbler.

I think David could have noted some of the history on the thinking surrounding the Carbonated Warbler and Audubon's mystery birds. For example, Ken Parkes reviewed Audubon's mystery birds in a story published in Natural History magazine back in 1985. Parkes did not say much about the Carbonated Warbler, but did feel this was one mystery that had a highly plausible answer—first Alternate male Cape May Warbler. Personally, I have a difficult time seeing these as immature Cape Mays, but many features do suggest that. Given the vagueness of Audubon's painting and description, perhaps because they were based partly on memory, the various problems David points out make sense and should make us reevaluate Parkes's and Jardin's suggestion. (Coues suggested hybrid Cape May x Blackpoll). Douglass Morse, in his book "American Warblers," reviewed the situation too and noted that Audubon had worked from memory on a number of plates at the time of the Carbonated Warbler illustrations (rats!). Morse gives some other possibilities too, and readers may wish to consult that volume for further information.

In sum, I think the possibilities that David offers are worth consideration, but "Sylvia carbonata" of Audubon is probably not as much of an enduring mystery by comparison. Now, Townsend's Bunting, that's interesting one...

David Sibley said...

Thanks Louis, for adding some background and some important points. The Jardin/Coues/Parkes suggestions that these are poorly-painted Cape May Warblers or Cape May X Blackpoll hybrids is of course an attempt to make sense of what we see in the painting, but it should carry the qualifier "If the painting is a faithful representation of the birds...."

We know that Audubon worked on some paintings from memory at that time, and I think this painting is consistent with "from memory" work. As any artist knows, working from memory will get some details correct and others very wrong. It's impossible to perfectly recreate a memory as a drawing, because memories tend to be fragmentary (drawings must be complete), and as the drawing appears on the paper it interacts with the memory, altering and recreating the remembered image (n.b. - the same thing happens with field sketches, thus the need to sketch and double-check details while watching a bird, not after a sighting).

If this painting was as intimately detailed and fundamentally accurate as most other Audubon works I would be arguing that he did have specimens and that this was a real species. Unfortunately my study of the painting leads me to conclude that there is a good chance this is not an accurate representation of the birds. It's unlikely that we'll ever know for sure either way, so we just have to look at all of the evidence and weigh the possibilities.

Scott Weidensaul said...

Just a quick response to Louis's comment...far from criticizing David for taking a fresh look at this old mystery, I commend him for it, and I thought I'd made that clear. My only recommendation was to consider what effect the engraving and tinting process may have had on what Audubon originally painted. "Hands-off reverence" is hardly what I'm encouraging (especially given my friendship with Clem Fisher, and her work with Gould's birds).

Fascinating conversation...

Louis Bevier said...

This is an interesting topic, so pardon me for stepping in again. Thanks for the clarification, Scott. It seems I projected the wrong meaning onto "give the old guy the benefit of the doubt" as used by both you and "cyberthrush."

Some additional points-- I agree that we cannot know what Audubon had in mind, or in hand, unless someone finds a specimen that can be tied to his plate and description (both are important). Without that independent verification, the name simply remains in limbo. Questioning the reliability of what we have doesn't diminish Audubon or the name Carbonated Warbler. Suggesting explanations is valuable, and I see David's points as adding to the case already made long before: that the birds were probably drawn from memory.

Robert Mengel said just that in 1965 in his "Birds of Kentucky" (pg. 523): "Although Audubon suggests to the contrary, the plate may have been based on memory, which in this instance served poorly (see A.O.U. Check-List, 1931:373)." The text in that Fourth edition Check-List says: "...As a number of his drawings of birds obtained about this time were later destroyed it is possible that the published plate may have been based to some extent upon memory."

In short, most of what has been written on this blog has been said before. Although I didn't specifically point it out, I think it is notable that Jardin questioned the validity of carbonata in 1832, just one year after publication of the description in 1831. So, it is not as though David's current critique is without long precedence.

The point about whether Audubon, his engravers, or both could have contributed to the confusion is a good one. This happened to some extent with another of Audubon's mystery birds, the "Bird of Washington," which most now accept as likely just immature Bald Eagles. Some copies engraved and tinted by Lizars had the name Falco washintoniensis, whereas all the others engraved by Havell had the name washingtonii, which is the name Audubon used in his text. Robert Mengel reviewed the case of this form too. He weighed all the possibilities that apply to the warbler: this could have been a gigantic, never-again-seen species of eagle (Audubon gave the wingspread as 10 feet 2 inches) that once bred in Kentucky (Audubon said he saw a pair of these brown eagles at a cliff nest); or the lost specimen could have represented a genetic freak (but how to explain the "pair"); or, lastly, that Audubon's information is just unreliable in this case. Mengel even put it delicately, saying, "…we do not know how (and let me suggest, reluctantly, if) Audubon actually made his measurements." (Wilson Bulletin 65: 148, 1953).

One final comment on David's critique and the idea of examining the original watercolors-- I think the patch of yellow above the bend of the wing can be explained as the breast feathers fluffed out over the wrist and scapulars. Having had fresh specimens in hand (ahem!), this happens commonly in smaller birds during preparation. Audubon illustrated his male Chestnut-sided Warbler with a similar broad patch of chestnut over the bend of the wing and spreading onto what should be the scapulars and illustrated the female normally. Given Audubon provides a reasonable description of his Sylvia carbonata, I think he did have some bird or birds in hand at one time. But if he redid his artwork without the specimens or sketches made at the time the birds were collected, I don't think the "original" watercolors will help determine what these birds really looked like.

I still think that Ken Parkes had this about right. The short tailed look, "acute" bill, "lower back dull yellowish-green" [= rump?], "tips of the second row of coverts white, of the first row yellow" along with other aspects of Audubon's description and plate, fit Cape May Warbler tolerably well, assuming the cheek patch was absent (or forgotten). Kinda like this.

David Sibley said...

I think this is a really good discussion, so by all means keep "stepping in". I only disagree with one of your points Louis:
The yellow wedge above the bend of the wing simply can't be the breast feathers. On the Chestnut-sided Warbler Audubon painted the chestnut breast-sides of the male overlapping the bend of the wing in an extreme way, but the structure is still clear and makes sense (more or less). The yellow wedge on the Carbonated Warbler is not drawn as over-fluffed breast feathers (they would hide the bend of the wing) and has to represent the scapulars and mantle feathers. Those feathers are clearly olive with dark centers on the upper bird, it would be really odd for any warbler to have a yellow patch there, and Audubon was always more careful than that in his drawing of structure.

I think this mistake is one of the most telling.

Paul said...

David, your observations on the Carbonated Warbler plates are interesting for what they add to a persistent mystery. I have always thought that particular painting was not up to the standards of many of Audubon's other works. Of course, you are right to remind us that the inconsistencies may simply be due to his painting from memory rather than specimens.

The reason I am writing is there is a possible explanation for the mysterious plates that I can relate from personal experience, and I believe Audubon may have been painting an actual bird, even if he was not painting from the specimen. An account of the event I am referring to was published in the journal of the Minnesota Ornithologists' Union, The Loon (Vol 67, no. 3, p. 167).

In 1991, my brother and I found a strange warbler in western Minnesota that we could not identify because of its odd combination of clean black cap and yellow underparts. It also had black streaking on the sides and two white wingbars. We took careful field notes, and over the next year discussed the identification many times with the problem being an absence of any known species with that particular combination of marks. Eventually, we were left with a best guess of it being an aberrant Blackpoll Warbler with yellow pigment where the body feathers ordinarily are white.

It was a year or two later, when we happened upon a reference to the Carbonated Warbler that it occurred to us how similar our bird was to Audubon's painting. We did not know about the Carbonated Warbler at the time of the observation, but the similarities were so striking, we immediately thought of our bird. I don't mean to suggest that we have confirmed the existence of the Carbonated Warbler, rather, if our bird was indeed an aberrant, yellow Blackpoll, then perhaps that is what Audubon painted, or something similar. Our published account includes an image comparing what we documented side-by-side with Audubon's plate. I've put a copy of this image at, if there is any interest.

Paul Hertzel