Bits and Pieces Part 3: Internalized Beliefs, How They Got That Way, and What the Record Really Shows

Part 1 is here. Part 2 is here. It looks like this series will end up being a five-parter. Part 4 should follow later this week and Part 5 sometime thereafter.

Most of us with an interest in the ivorybill have internalized a set of beliefs about what constitutes “suitable” habitat. To a large extent these beliefs treat the Singer Tract as a model – a vast tract of “virgin” bottomland forest dominated by oaks and sweetgums, with abundant, moss-draped cypress for atmosphere (although ivorybills seem to have avoided cypress in the Tract). The habitat description in Stephen A. Shunk’s excellent Peterson Reference Guide to Woodpeckers of North America clearly expresses some of these beliefs, which have influenced the overwhelming majority of modern search efforts and which are embedded in the minds of most searchers (myself included) to the point of being a default:

Virgin bottomland forest almost always below 100 ft. (30 m) elevation. May also have occurred in uplands but by 1900 restricted to areas downstream of pine-bald cypress interface. Requires large tracts of contiguous forest with very large-diameter trees and adequate dead and dying trees to provide forage and nest sites.

Screen Shot 2017-10-09 at 1.54.44 PM

The Guide, which draws on an 1891 article by Hasbrouck for the 100 foot elevation, goes on to list three different habitat types described by Tanner – sweet gum-oak dominated forests in the southeast (with species of oak varying depending on location) outside of Florida; river swamps in Florida dominated by cypress, black gum, and green ash; and creek swamps in Florida characterized by cypress, red maple, laurel oak, black gum and cabbage palmetto, with feeding in adjacent pine woods.

This is not to criticize the Guide or its author – I recommend the book highly and the overall treatment of the ivorybill is thorough and evenhanded. The quote is intended to point out the pervasiveness of these ideas about habitat requirements, ideas that Tanner reinforced, especially in later years. They’re so pervasive in part because the myth of the “virgin forest” has shaped ivorybill lore since well before Hasbrouck and has influenced almost all habitat assessments since Tanner.

The virgin forest myth is a topic for another day; the central point is that ivorybills have been found in more diverse habitat types than most have believed. It’s worth bearing in mind that Tanner himself asserted that ” . . . at present the only suitable habitat for ivorybills is in tracts or areas of virgin timber”, a narrow, almost lawyerly, and largely conjectural conclusion – one not entirely supported by fact. Ivorybills bred in at least one Singer Tract area (Mack’s Bayou) that was predominantly regrowth, and as has been discussed in previous posts, Tanner became more dogmatic and blinkered about habitat requirements in later years, dismissing the John Dennis Texas recording because a Pine Warbler was captured on the tape.

As should become clear, Ivory-billed Woodpeckers were found in a variety of different habitats and did not always require extensive tracts of contiguous forest. Large diameter trees, their importance, and what Tanner meant by “large” are discussed in another post, but even if one accepts that large trees were preferred in the Singer Tract, much of the feeding sign Allen and Kellogg documented in Florida was on small, fire-killed pines.

My plan for this post was to focus on Ivory-billed Woodpecker records from outside the historic range as delineated by Tanner in his oft-reprinted 1942 map and adaptations thereof.

Screen Shot 2017-10-08 at 4.10.51 PM

US Fish and Wildlife Service range map based on Tanner, including locations of selected post- Singer Tract reports

The purpose behind this original plan was in part to show that the historic range of the ivorybill was considerably more extensive than is commonly believed, but more importantly to show that ivorybills inhabited more varied habitats than is commonly believed. As time went on, my focus has shifted even more toward the question of habitat diversity, though the fact remains that the historic range was considerably more extensive than the Tanner map, or the one drawn by Hasbrouck in the 1890s, might lead one to expect.

Screen Shot 2017-10-08 at 4.05.13 PM

Hasbrouck’s Ivory-billed Woodpecker Range Map (1890)

In recent Facebook comments, several ornithologists have suggested that without physical evidence, no record should be accepted. As I see it, this standard is, to some extent, a kind of ahistorical overkill, since it has never been applied in the past. Post-1939 Singer Tract observations don’t meet it, so the last record date is pushed back by several years, which can impact statistical analyses. At the same time, the parsimonious approach is not entirely meritless, since it eliminates false positives. The problem is that there’s no purely objective standard for evaluation of historic (and pre-contact) records, even when it comes to specimens. In many cases, location information for specimens is non-existent or ambiguous; for example, one specimen in Cornell’s collection (1896) is listed as coming from the “Florida Keys”; “Key” in this context more likely refers to an island of forest surrounded by the Everglades than to the islands offshore.

Thus, in this post and the next, I’ll be looking at many reports from within the recognized historic range, as well as some from beyond those boundaries. I will be focusing on reports accompanied by physical evidence or published accounts stating that physical evidence was obtained but will include or mention a few additional ones that seem particularly credible based on the source or amount of detail.

The more surprising of these reports are unsupported by physical evidence. These come from the Eastern Shore of Maryland, Swedesboro, New Jersey, and Pennsylvania. Most date to the 18th-century, a time when ivorybills were reported to have fed on trees girdled for clearing. The most interesting of these come from Peter Kalm, a Swedish naturalist and student of Linnaeus. They have been discussed in several articles by Benjamin Leese, who has also written about early records from Ohio, Indiana, and Kentucky (for which the case is stronger). Most of these papers do not seem to be available online.

Just under 200 of the 418 specimens (including Cuban ivorybills) listed by Hahn provide no location information or merely identify the state, or country in the case of Cuba, where the collection took place. There are multiple cases in which specimens were reportedly collected but have not been found, and in several cases, there’s no way to correlate the claim of collection with an actual specimen. In one instance, not catalogued by Hahn, a pair of ivorybill specimens was mislabeled as Pileated Woodpecker until the error was corrected in the 1960s.

When it comes to pre-Columbian sites, there’s no way to be certain whether material collected from graves and middens involves trade goods or locally killed animals, although tarsometatarsi are likely local, especially east of the Mississippi, where there’s little evidence to suggest that woodpeckers had ceremonial value. Leese addresses this subject in a paper on Native American uses of ivorybill parts. I have included records involving tarsometatarsi and one from a West Virginia midden that involves parts of two lower mandibles.

I’ve created a google map showing the locations of the records from habitat types that don’t fit the ‘large tracts of contiguous bottomland forest’ paradigm. I’ve provided some details for each location. The map draws on Appendix E or the US Fish and Wildlife Service’s Recovery Plan, Tanner, and Jackson. I’ll discuss the records from coastal areas, south Florida, the upper Mississippi and locations east of the river in the next post. I’ll conclude this one with a look at the records from the western edge of the range, since they relate to the Nebraska report and why I thought it might be credible; the Nebraska location is actually somewhat east of the records from the southern plains.

One record that I find compelling does not include a specimen; it’s from the 1820 Long Expedition, the first scientific exploration of the American West. Screen Shot 2017-10-08 at 8.08.14 AM

The ivorybill’s call is described, and Pileated Woodpecker is distinguished and described as common in the area. These facts lend credibility to the report, as does the fact that Thomas Say was the expedition’s naturalist. While Say is best known as an entomologist, the expedition produced the first descriptions of a number of bird species, and an entire genus of flycatchers was named in his honor.

The location of this record is approximately south of Tulsa on the Canadian River, near the 96th Meridian. This is farther west than the location of the erroneous southeastern Nebraska report and well into the eastern Great Plains. It is also well beyond the range of the bald cypress. The relatively narrow floodplain would have been dominated by cottonwoods and willows, as it is to this day. This image, from ca. 1920s shows “Standing Rock” a geological feature now flooded that was discovered by the expedition a day after the ivorybill encounter. It should add a visceral sense of the area’s appearance to supplement the description above.

Screen Shot 2017-10-09 at 2.45.16 PM

Even if one opts to reject this record for lack of physical evidence, there are several others from approximately the same longitude that do involve specimens, and some are from the 20th century.

The Recovery Plan suggests that there are two records from west of Tulsa, a specimen was “probably” collected by Woodhouse along the Cimarron River, Pawnee County in 1849. Per Jackson, the specimen was sent to the Philadelphia Academy of Natural Sciences, which has four specimens without location information in its collection. The second is from House Creek in Pawnee County, also Woodhouse 1849, and also reportedly sent to the Philadelphia Academy but not found there.

There are several Plains records from Texas. A specimen currently in the Dallas Museum was collected on Bois d’Arc Island, just southeast of Dallas, elevation 400′, in 1900. There were multiple reports from the area through 1910, and an additional bird may have been collected in 1918. A bird was reportedly “caught in a trap” in nearby Kaufman County in 1927 and examined by an R.E. Huck but not preserved. An additional Texas record, from farther south but west of the 96th Meridian, comes from New Braunfels County, south of Austin and east of San Antonio. There were multiple reports ca. 1900, with a collection reported but no specimen preserved.

Although only one record from the eastern Plains can be attached with certainty to a currently existing museum specimen, there’s proof that Ivory-billed Woodpeckers were found in this region as recently as 1900 and considerable circumstantial evidence for their presence along riparian corridors on the plains of Texas, and possibly Oklahoma,  into the 20th-century. The habitat involved is markedly different from what so many have believed ivorybills require. I’m not suggesting that ivorybills persist at the western edges of their historic range, but as will be discussed in the final installment of this series, I think their ability to exploit these relatively narrow, willow and cottonwood dominated floodplains can help explain how the species could have persisted into the 21st century.

Stay tuned for Part 4.


The Two Faces of James T. Tanner – A Study in Selective Memory and Blindspots

In a 1936 letter to James Tanner before Tanner began his survey of possible ivorybill habitat in the southeast, Herbert Stoddard wrote, “The area where they (Ivory-billed Woodpeckers) may occur at present is simply tremendous, not restricted as many believe.” Stoddard continued, “. . . if I had the rest of my life for the purpose, I doubt I could cover adequately half the ground I now think worth investigating. When I say adequately I have in mind five days I spent looking especially for these birds with men such as Bob Allen and Alex Sprunt on an area of some ten thousand acres known to be frequented by several pairs of these birds without seeing one. Of course this was due to the element of luck, as others have gone in the same area for a few hours and seen one or two. But it indicates the time one would have to spend in these great river valleys to really be reasonably sure that the birds were absent or even extremely rare therein.” Tanner was one person, and he clearly didn’t have the time to devote to the thorough survey Stoddard suggested, but the later record suggests that he failed even more deeply to take Stoddard’s words to heart. Stoddard’s perspective leaves room for the possibility that Tanner grossly underestimated the ivorybill population in his monograph, something that might lead to very different projections about the likelihood of survival today. It also sheds additional light on the difficulties in searching for ivorybills and on what appear to be some significant blindspots in Tanner’s thinking about the species.

I recently wrote a Facebook post in which I stated Tanner had a “hard time” finding Ivory-billed Woodpeckers. A very prominent American birder and ivorybill skeptic promptly shot back asking for citations, which I provided; I wouldn’t play along with the snarky response, but the exchange gave me the impetus to write this post. To most birders, Tanner is a kind of heroic figure, a pioneer ornithologist, and the author of the only in-depth, definitive study of the Ivory-billed Woodpecker, a bird that, in Tanner’s telling (and even more so in the retelling), was a super specialist that depended on virgin forest for its survival. Tanner is also rightly seen as having played a central role in the development of modern environmentalism, fighting to save the Singer Tract itself and later helping to create Congaree National Park, one of the few remaining substantial old growth stands in the south.

Those who are more deeply steeped in ivorybill lore are likely to have a somewhat different perspective, even as they admire Tanner’s good qualities. In later years, Tanner became increasingly dogmatic about ivorybill habitat requirements and was usually harshly dismissive of  reports from others – including John Dennis in Texas and Agey and Heinzmann in Florida. (This dismissiveness began to take hold in the early 1950s.) By no later than 1985, and probably much earlier than that, Tanner had become convinced the ivorybill was extinct, a conviction that he usually leavened with the statement that he’d love to be proven wrong. Ivorybill aficionados may also raise questions about the 1930s surveys he conducted in the southeast, and the rapidity with which he dismissed almost every place he visited as being unsuitable, sometimes based on a visit lasting a day or two. Most of these visits were made outside of breeding season, at times when finding ivorybills was considerably more difficult.

Tanner may have been right to dismiss many of these areas, but the possibility that he might have missed several populations cannot be ruled out, especially in light of the letter from Stoddard, an established ornithologist with first-hand knowledge of the species. At minimum, Stoddard’s letter put Tanner on notice that ivorybills could be very difficult to find.

Were preconceived beliefs driving Tanner from 1937-1939; was he being cavalier in his dismissals; or was he simply doing his best to accomplish what Stoddard said would take a lifetime in a few weeks spread over three years? It was probably all of the above, but whatever the reason, his mindset became an even bigger problem as time went on, leading him to foment the false impression that ivorybills should be easy to find, a notion that informed the views of my interlocutor on Facebook and of many those who believe the species is extinct.

The fact is that when Tanner was in the Singer Tract, the only birds he could find on a regular basis were the “John’s Bayou Family”, the group first studied in 1935. Their nesting and roosting grounds were approximately 1 mile from Sharkey Road. From 1935-1939, the birds nested in the same general vicinity. Three of the four nests were in very close proximity to one another, and the greatest distance between any two nest sites was 3/4 of a mile. Despite this clustering, it took Tanner five days to find a nest in 1939, when he didn’t have J.J. Kuhn (who had been forced out of his job as Singer Tract Warden) to assist him.

While Tanner paid respect to and credited Kuhn, the passage of time has made it increasingly clear that he downplayed the degree to which he depended on Kuhn to find ivorybills. Kuhn had been observing the birds for several years before the 1935 expedition (which also took several days to find ivorybills), and it’s reasonable to infer that the John’s Bayou family was already somewhat habituated to human presence by 1935. Certainly after ’35, they were accustomed to human activity in the immediate vicinity of their nest sites.

The only other ivorybill Tanner saw was one he named Mack’s Bayou Pete, presumably a bird hatched in 1935 from a nest that Kuhn found. As with the John’s Bayou pair, this bird was found within a home range that had been roughly delineated in 1935; Tanner believed he could recognize Pete’s voice, but the bird was far more frequently heard than seen.

Edited to Add: I’ve been reminded that the 1935 Mack’s Bayou Nest failed, and that Pete more likely fledged in 1937. The fate of the adult Mack’s Bayou pair is unknown.

Tanner continued to correspond with informants in the Singer Tract until 1948 or ’49 and was advised that one or two birds remained until that time. As Tanner’s temporal distance from his experiences in the Singer Tract grew, so did the distortions in his memory. These distortions are perhaps most dramatically illustrated in an article he wrote entitled “A Forest Alive”, which was published posthumously in Birdwatch (2001), a British magazine. Prominent ivorybill skeptic Martin Collinson mentioned the article and quoted it at some length in a 2007 blog post:

Collinson concluded his post with this observation: “As others have pointed out… you can only rationalise the failure of the current searches if you don’t include the words ‘active and noisy’, ‘called frequently’ and ‘easier to see and follow’, in your dataset.”

Of course, Tanner’s observations applied only to the John’s Bayou birds studied in close proximity to a nest. More importantly, Tanner’s recollection was incomplete and inaccurate, and he corrected it in another, humbler, posthumous piece that appeared in Birdwatcher’s Digest (2000) under the title “A Postscript on Ivorybills”. Tanner began by observing that, “Like everyone else’s memory, mine forgets the annoying and unpleasant things and remembers the pleasing; the mosquitoes go but the birdsong remains. Mine is also likely to forget the important and regular events and retain the trivial and unusual.”

The article makes it clear that finding even the John’s Bayou birds (which had a known territory and roosting ground) could be difficult and was even harder outside of nesting season. Tanner concluded with this reminiscence: “Hunting in other areas of the Singer Tract for ivorybills was even more difficult and discouraging. My journal is full of such comments as ‘saw old sign, lots of impenetrable vines, and no ivorybills.’ The only other ivorybill we ever found outside of the one nesting pair and their offspring was a single male that we dubbed Mack’s Bayou Pete because of the area he usually inhabited. He was hard to find and once found was soon lost . . . ”  And to reiterate, Mack’s Bayou Pete’s home range was at least something of a known quantity.

By 1985, Tanner was advising the US Fish and Wildlife Service to classify the ivorybill as extinct. He gave the following reasons, to paraphrase:

1) The reports have been unconfirmed and they often describe behavior that is characteristic of Pileateds (flushing from stumps and logs).

Tanner described this as being rare in ivorybills, which may or may not be true, but it’s certainly not unheard of. Allen and Kellogg observed a female feeding on the ground. Moreover, not all reports involved this behavior.

2) There’s so much recreational use of areas where ivorybills used to be present that people should be hearing and seeing them, and there are no extensive areas of former IBWO habitat that are remote and unvisited.

What’s omitted from this analysis is the fact that the beginning of nesting season and the end of deer season are roughly congruent, at least in Louisiana. This means that there’s actually very little human traffic in many prime nesting areas during the times when birds are most active and vocal. And of course, the John’s Bayou area was not particularly remote, although other parts of the Singer Tract were considerably more difficult to reach.

3) Ivorybills are fairly conspicuous birds that, “if present . . . are not hard to find. Furthermore, the feeding sign they made (scaling bark from extensive areas of not long dead limbs and trees) could be easily recognized as a clue to their presence.”

With regard to the feeding sign, I have to wonder what Tanner would say about the sign we’ve found that appears to match this description perfectly. With regard to being easy to find, Tanner’s field notes and his reminiscence in “A Postscript . . .” make it clear that ivorybills were extremely difficult to find, except in extraordinary circumstances – in close proximity to a nest and perhaps with a good deal of habituation to human presence.

Tanner’s story is amazing. His monograph is seminal though not gospel. The in-depth examination of his work in Stephen Lyn Bales’s Ghost Birds is essential reading. Many of his later pronouncements, however, have done a disservice and continue to sow confusion more than two decades after his death.

I’m indebted to Fredrik Bryntesson for sharing Herbert Stoddard’s 1936 letter and reminding me of Tanner’s missive to USFWS. This piece is informed by years of discussion with Frank Wiley. He also made a number of constructive suggestions about the specific content. Direct descendants excepted, Frank was undoubtedly J.J. Kuhn’s greatest admirer.

Apples to Apples and Sweet Gums to Sweet Gums

I’ve been corresponding with Mississippi-based searcher Christopher Carlisle both privately and on Facebook, and our conversations have inspired some additional thoughts on bark scaling and led me to revisit Cornell’s 2006-2007 final report, which includes two interesting photos of Pileated Woodpecker work on sweet gums taken by Martjan Lammertink in Congaree National Park. I was familiar with the document, which is available here but had forgotten about the images. Scroll to page 30 (some additional images of interest from Texas appear on the following page.)

While it’s a tiny sample, there may be some value in comparing the Congaree photos with the work on two heavily scaled sweet gums in the Project Coyote search area.

Correction: the tree in the second image is a hickory.

Photo by Steve Pagans

Photo by Steve Pagans


While it was not possible to examine this scaling up close, the bark appears to be tightly adhering on both trees, and the decay state is likely comparable to the Congaree sweet gum. The work in both cases is on the boles, and the size of the trees involved seems to be roughly comparable. Beyond that, the scaling is dramatically different in a number of ways. Most obviously, the suspected IBWO work from our search area is far more extensive than the PIWO work from Congaree (I suspect the small patches of scaling in the lower photo are the work of a Hairy Woodpecker). As I’ve discussed, PIWOs are not well-suited to scaling bark, anatomically, and it’s not a preferred or efficient feeding strategy – accounting for 23% of observed foraging behavior in Tanner and 7% in Patricia Newell’s more recent and PIWO-focused study.

The most significant difference though is in the appearance of the edges. When Pileated Woodpeckers scale tight-barked hardwoods, they typically remove the bark in layers, as in these images. This layered scaling is also very apparent in the photos from Congaree, and it’s absent from our sweet gums as well as from the extensively scaled oaks and hickories that we found in 2013-2014. Our trees show very extensive scaling with no indication that bark has been removed in stages; the edges are clean and incised and very large areas have been stripped, down to the sapwood. If Pileated Woodpeckers were the cause, it would be reasonable to expect that there would be some sign of layering, especially given the surface area involved, which far exceeds that shown in the Congaree photos or any others I know of (on hardwoods) from the Cornell searches.

The high branch work from Texas is intriguing, but the resolution of the images is insufficient to determine whether this clean scaling or whether there’s some excavation, especially on the lower portions. In addition the work is not nearly as extensive as some of the high branch work we’ve found.


Edited to add: I’m more intrigued by extensive work below the crown, as in these images. One tree (in the third photo is), definitely a sweet gum is from the old search area. The other tree, which I believe is an oak, is within a hundred yards of where I recorded kent-like calls in 2013, although the photos were taken a year before.



There are several other images of sweet gum scaling from Congaree in this report from the mobile search team, but this work does not involve the removal of bark from large contiguous sections of trunk, and the resolution is not sufficient to tell whether the bark has been removed in stages, although it appears to be in at least one of the four photographs. If I were to find it, I would assess this work as being mildly interesting but would not get excited about it in the absence of other indicators.

To return to my exchange with Chris, it gave me the opportunity to revisit this material, to give some more thought to my hypothesis about feeding sign, and to make some adjustments:

  • I think scaling on pines, even live ones, is physically possible for a pileated, although the bark will often show signs of having been removed in layers, meaning the edges will not appear as clean. Nonetheless, I do not think there’s a way of reliably determining what species has scaled a pine in the absence of a direct observation and suspect that even HAWOs and RBWOs can scale extensively even on recently dead pines.
  • For the work that I think is diagnostic, the species that I think are most reliable are oaks, sweet gums, and hickories. I think that the distinction is an easy one to make once you’ve seen the work firsthand.
  • In the old Project Coyote search area, we had persimmons, hackberrries, and honey locusts that had very suggestive scaling and lots of it. But the bark on honey locusts tends to loosen early in the decay process. Hackberry bark fractures, and while persimmon bark is thick and tight, it tends to loosen when the wood is still very hard. These qualities complicate the analysis, though I’m confident some of that work was done by IBWOs. There was a vast difference between the size of bark chips from known PIWO foraging on honey locusts and suspected IBWO foraging on the same species.
  • Extensiveness and quantity (concentrations) are important to look for too, especially when there are a lot of pines or an abundance of more easily scaled hardwoods. For this analysis, I include looking at high branch work but otherwise don’t ascribe a great deal of importance to “Tanneresque” sign, unless there’s a lot of it, and it’s associated with what I think is diagnostic.

I am not suggesting that this is the only way that Ivory-billed Woodpeckers forage. There may be seasonal and regional variations, but I suspect that some of this type of work will be present in any area where IBWOs are resident. It’s not necessarily easy to find. I agree with Fangsheath, from, that the failure to do so should not be treated as evidence of absence, but I’m convinced that finding this very specific type of work is compelling evidence that IBWOs are present.