This post is inspired in large part by an exchange of emails with Chris Sharpe, an ornithologist who is working on an IBWO literature review. Our correspondence revolved around the IUCN’s species account, which describes the ivorybill as “possibly extinct” and cites recent statistical analyses that suggest extinction is likely, as well as one that indicates survival is possible and another that concludes “very large search efforts are needed to detect small populations.”

Chris pointed out that there are many other species on the Red List that fall into a similar category with many unverified reports but in more remote habitats and nowhere near the search effort that has been expended on the ivorybill. While there’s some validity to this assertion, I think the reality is considerably more complex and that the ivorybill is in fact sui generis.

Many of the specifics of ivorybill history are little-known, and the statistical studies seem badly flawed. One focuses on collection rates but may not adequately address changing attitudes toward conservation in the early 20^th century among other factors. Another is perhaps even more problematic for a number of reasons – most importantly its focus on “verified sightings”, which is a particularly complicated issue when it comes to the ivorybill. It omits numerous controversial sightings and does not include post-Singer Tract instances in which physical evidence was obtained, although the authenticity of that evidence has been contested. “Sidewinder” posted the abstract and a good summary of the findings on ibwo.net a few years ago:

Roberts, D. L., C. S. Elphick, and J. M. Reed. 2010. Identifying anomalous reports of putatively extinct species and why it matters. Conservation Biology 24:189–196.

As species become very rare and approach extinction, purported sightings can stir controversy, especially when scarce management resources are at stake. We used quantitative methods to identify reports that do not fit prior sighting patterns. We also examined the effects of including records that meet different evidentiary standards on quantitative extinction assessments for four charismatic bird species that might be extinct: Eskimo Curlew (Numenius borealis), Ivory-billed Woodpecker (Campephilus principalis), Nukupu`u (Hemignathus lucidus), and O`ahu `Alauahio (Paroreomyza maculata). For all four species the probability of there being a valid sighting today, given the past pattern of verified sightings, was estimated to be very low. The estimates of extinction dates and the chance of new sightings, however, differed considerably depending on the criteria used for data inclusion. When a historical sighting record lacked long periods without sightings, the likelihood of new sightings declined quickly with time since the last confirmed sighting. For species with this type of historical record, therefore, new reports should meet an especially high burden of proof to be acceptable. Such quantitative models could be incorporated into the International Union for Conservation of Nature’s Red List criteria to set evidentiary standards required for unconfirmed sightings of “possibly extinct” species and to standardize extinction assessments across species.

Here are the Ivory-billed Woodpecker sighting data they used:

Physical evidence: 1897, 1898, 1899, 1900, 1901, 1902, 1904, 1905, 1906, 1907, 1908, 1909, 1910, 1913, 1914, 1917, 1924, 1925, 1932, 1935, 1938, 1939

Independent expert opinion added: 1911, 1916, 1920, 1921, 1923, 1926, 1929, 1930, 1931, 1933, 1934, 1936, 1937, 1941, 1942, 1943, 1944

Controversial sightings: 1946, 1948, 1949, 1950, 1951, 1952, 1955, 1958, 1959, 1962, 1966, 1967, 1968, 1969, 1971, 1972, 1973, 1974, 1976, 1981, 1982, 1985, 1986, 1987, 1988, 1999, 2004, 2005, 2006

Data were from Tanner (1942), Hahn (1963), Jackson (2002, 2004), Fitzpatrick et al. (2005), Hill (2006), and Floyd (2007).

Some conclusions: For the Ivory-billed Woodpecker, the prior sighting record suggests that even by the time of the first controversial sighting, the species was relatively unlikely to remain extant (ca. 21% chance), regardless of the level of evidence (physical or independent expert opinion) used…the effect on the predicted extinction date will depend on the details of the sighting record. Including controversial sightings will, by definition, move expected extinction dates forward in time. An ever-increasing burden of proof should be required with increasing time since the last verified sighting. The burden of proof also should be greater when there is a pattern of frequent sightings prior to the last accepted record and lower when long periods between sightings are common in the historical record.

A second paper published in 2012 reached the same conclusion using 39 sightings “classified as certain and 29 classified as uncertain”.

The insistence on verification is problematic because it’s founded on an appeal to authority, and for much of the time frame in question, the primary if not sole authority was James T. Tanner. Tanner had a strong predisposition to dismiss every post-Singer tract report he received and was somewhat cavalier about reports he investigated in the late ’30s as well. This likely resulted in his underestimating the population at the time. At minimum, he missed six pairs in Mississippi, according to Jackson ( In Search of the Ivory-billed Woodpecker, pp. 60-62.)   These details reinforce the point made by “Fangsheath” of ibwo.net that, for the ivorybill at least, demanding verification leads to “a rather vicious circularity”. In a recent email, Fangsheath also pointed out that “the Roberts et al. scheme fits tidily into the narrative created by Tanner, that the Singer Tract was the last hope of the ivory-bill, and that when that: wilderness was lost the bird essentially became extinct . . . Tanner searched across several states for the bird and found none elsewhere, yet he himself believed that the bird survived in a number of areas. And so we have a curious contradiction – on the one hand an appeal to authority is being used to exclude many sightings, and on the other hand that same authority is saying the bird exists in areas where sightings are excluded. The mythos that the bird had very specific habitat requirements clouded every search effort and every sighting evaluation, before and after 1944.”

Fangsheath’s observation about the impact of Tanner’s narrative on future search efforts is profoundly important. Tanner’s ideas about habitat, which were in large part a product of his cultural background and the “myth of the frontier”, hardened over time. In later years, he ignored some of the caveats he set forth in his own monograph. But even his approach to reports from the 1930s reflected a set of beliefs about habitat requirements that had no scientific basis. Many people, myself included, are prone to reflexively accepting these assumptions about “habitat quality” because the conventional wisdom is so deeply engrained.

Fangsheath also reminded me that in the years before Tanner, the species was being written off and had been for decades. Florida was believed to be the last stronghold; the Singer Tract and Atchafalaya basin were not considered, nor were the large tracts of overcup oak/water hickory forest (where Beyer found ivorybills in 1898), many of which were untouched until the 1940s. Unlike the Singer Tract, these areas were often roadless and very difficult to penetrate; the Tract was bisected by a road, had few deep bayous, and was largely free of undergrowth, making it much easier to search. I’ve already discussed Tanner’s difficulty in finding ivorybills anywhere besides John’s Bayou. In this context, it’s worth noting that Bick’s 1941 sighting (from his car on Sharkey Road) involved two birds feeding in a lower lying “ash flat” in which overcup oak predominated.

In addition to these conceptual flaws, the papers grossly underestimate the number of post-Singer Tract encounters within the historic range. While it is impossible to quantify the controversial reports, it’s clear that the 26 or 29 referenced in the studies are the tip of the iceberg. There are well over four times that many on record for the 1944-2003 time period. And no doubt, numerous encounters never made it into the literature.

I initially posted a brief comment about this on Facebook and ibwo.net, with a link to Jerome Jackson’s 2002 Birdwatching Daily article listing 20 pre-Arkansas and post-Singer Tract encounters (some just auditory and one from Cuba) and a reference to Michael Steinberg’s Stalking the Ghost Bird catalogue of 85 sightings during the same time frame. A commenter wrote, “Yes but no confirming photos”. That also led me to think it would be worthwhile to explore this subject in a somewhat more depth, since there is physical evidence, albeit contested, related to several post-Singer Tract and pre-Arkansas reports.

Before turning to the physical evidence, it’s worth reiterating that the 1944 date for the last “verified sighting” is fundamentally flawed and arbitrary. The 1944 date is for the “Say Goodnight” encounter that involved artist Don Eckelberry and two local boys, Billy and Bobby Fought, and the purported last lone female IBWO in the Singer Tract. This poignant story was retold in The Race to Save the Lord God Bird, and it has become central to the popular lore about the species. There’s no doubt that the encounter took place, and Eckelberry no doubt believed that the bird was the last of her kind. Roger Tory Peterson apparently received and accepted a report that a single bird, presumably the same one, was still present in 1946. Tanner was likely aware of this and had an additional reason to think that birds persisted in the Tract well after 1944. His papers include a letter stating that Singer Tract game warden Gus Willett saw a pair in November 1948 at North Lake #1 (I have been unable to identify this lake). According to the letter, the “[b]irds are moving over a much larger area than formerly.” The letter mentions that there had been several other sightings during this time period. There’s no reply from Tanner in the archives and no further correspondence about the Singer  Tract birds. It appears that this is the letter from Tanner’s former student, Arthur MacMurray, that is referenced in Jackson, but Jackson’s account does not mention Willett, who would have been familiar with ivorybills. The strong possibility that a pair of birds remained in the Tract for more than four years after it was cut should itself raise questions about Tanner’s narrative.

Jackson’s list of 19 US reports between 1944 and 2003 was undoubtedly not intended to be comprehensive (and his book includes many others). Instead, it focuses on sightings by professional ornithologists and/or people who were familiar with the species. These include: Allan Cruickshank, John Terres, Herbert Stoddard, John Dennis, Davis Crompton (Dennis and Crompton studied ivorybills in Cuba), Whitney Eastman, and William Rhein. Jackson doesn’t mention it, and was perhaps unaware when he wrote the article, but Rhein had filmed the Imperial Woodpecker in Mexico several years prior to his 1959 Florida ivorybill sighting. It strains credulity to think that every one of these experienced observers, most or all of whom were familiar with the species and all of whom knew pileateds well, would be mistaken. Jackson points out that Terres kept his sighting to himself for more than 30 years out of “fear of being scorned.” Such was the climate surrounding ivorybill claims, even in 1955.

Steinberg lists 85 sightings between 1944 and 2003. His list includes most, if not all, of Jackson’s reports, breaking them down into individual incidents. Jackson treated repeat encounters in the aggregate. Nonetheless, most of Steinberg’s reports do not appear on Jackson’s list. Many of them are from less illustrious sources and quite a few are anonymous, but some of them are from game wardens, field biologists, and graduate students in ornithology.

The Ivory-billed Woodpecker Recovery Plan, prepared by the U.S. Fish and Wildlife Service and completed in 2010, includes an even more detailed compilation of records, some of which are of course also included in Jackson and Steinberg. It lists approximately 100 reports of varying quality between 1944 and 2003. Some of these are aggregates, involving multiple observations. The number of observations is sometimes enumerated and sometimes characterized as “numerous”, “multiple”, or “several”, so it’s impossible to arrive at an accurate tally, but the total almost certainly exceeds 150.

Although the Steinberg and the Recovery Plan compilations are more extensive, they are undoubtedly far from comprehensive. For starters, the climate of intimidation around reporting ivorybills was strong enough to deter John Terres in 1955, and that climate of intimidation only grew more toxic over time, as the Big Thicket and Fielding Lewis incidents, not to mention the battle over the Arkansas reports, make clear. In addition, there’s good reason to believe that countless reports from local people, hunters, and amateur birders have been discounted, dismissed, or ignored by authorities, at least prior to the “rediscovery”. Steinberg writes of his first visit to the Louisiana Department of Wildlife and Fisheries Baton Rouge office in 2003,“ . . . few people took the book, or more important the larger issue very seriously. The typical response from many in the office, other than Nancy Higginbotham, seemed to be, ‘Don’t you have anything better to do?’” (Higginbotham claimed two sightings, one of male in 1986 and one of a female in 1987, both in the Pearl River area.) We’re personally familiar with several instances in which reports were dismissed or ignored. One involved the landowner in our old search area, who came forward after the Kulivan sighting and was deeply offended by the way he was treated. As far as I’m aware, no one has done a comprehensive review of records committee submittals within the historic range of the ivorybill; I know of one such submittal from Texas in 2002 that came from unsuitable habitat, outside the historic range, but there may well be others that are more robust.

I’d like to turn to three examples of physical evidence obtained in the post-1944 era. Some of this is fairly well known – the Agey and Heinzmann feather and the Fielding Lewis photographs. The other is somewhat more obscure, but no less interesting for being so.

The Agey and Heinzmann observations took place between 1967 and 1969, in Polk County Florida. They obtained two recordings that Tanner dismissed. There was evidently some very poor communication about it that was compounded by an obvious error on Agey and Heinzmann’s part. They mistakenly thought that calls on their first clip were consistent with some ivorybill sounds from the Singer Tract; they clearly are not. The second clip, which is dominated by a Red-shouldered Hawk, recorded March 3 1968, does have some faint kent-like sounds, but the quality is extremely poor. They are most easily heard on the amplified version that begins at 3:14. The RSHA calls were not what interested Agey and Heinzmann, but they failed to make this sufficiently clear.

I believe Tanner wrote the notes that accompany the recording:

“Well, I’m not sure what to say here. As far as I can determine, there are only four original sound clips here. The recordings at 0:04-0:32 and 0:49-0:57 are certainly flicker-like, especially the continuous series at 0:04-0:21 and 0:26-0:32, but I am less certain about the latter part of the first series (0:21-0:23), which appears to have been recopied twice at 0:41-0:48. A similar vocalization is included at 0:49-0:57 (recopied at 0:59-1:06). Given the very different quality of these sounds relative to those in LNS #6784, combined with the close similarity of the calls to those of Colatpes auratus, leads me to doubt that any of these sounds were given by C. principalis. The recordings from 3 March 1968 (2:45-3:07) represent the calls of Buteo lineatus. Quality unchanged (2;3 – the signal is not bad for the first part, but terrible for the rest, most of which seems to represent copies of various recordings)”

Agey and Heinzmann found several feathers near a cavity, and one was identified as the innermost secondary of an Ivory-billed Woodpecker. Oddly, Jackson wrote that “ . . . some shadow of doubt is cast over these records because Agey and Heinzmann also tape-recorded what they said were Ivory-billed Woodpeckers, and personnel at the Cornell Laboratory of Ornithology identified the birds on those tapes as pileated woodpeckers.” Given that one of the recordings does include kent-like calls, this criticism is not entirely warranted, and the misidentification seems irrelevant in light of the physical evidence (and ironically Tanner seems to have failed to correctly identify the calls on the first clip). Jackson also alludes indirectly to rumors that the feathers were taken from a specimen. These have circulated for years but are unsubstantiated. Agey and Heinzmann published their findings:

Agey, H. N., and G. M. Heinzmann. 1971a. The Ivory-billed Woodpecker found in central Florida. Fla. Nat. 44 (3):46–47, 64.

Agey, H. N., and G. M. Heinzmann. 1971b. Ivory-billed Woodpeckers in Florida. Birding 3:43.

To the best of my knowledge neither publication is available online at this time.

The next event, chronologically, started with a series of encounters in east Texas in the late 1960s. John Dennis obtained a recording that Tanner initially dismissed, in part for the patently absurd reason that a Pine Warbler is also heard on the clip, but later described as “a real mystery” when confronted with the sonograms and analysis that were suggestive of ivorybill. That recording is now catalogued as IBWO.

Additional slides, taken in 1970 in east Texas by a man named Neal Wright, were turned over to the Museum of Natural History in Philadelphia (search Ivory-billed Woodpecker, registration required to view image) which also houses the Fielding Lewis photographs. These were made public after the rediscovery. They show what appears to be a female Ivory-billed Woodpecker in a nest cavity. When I first viewed the Wright photographs, I was quite skeptical, but later, I was struck by the resemblance to this image from the Singer Tract.

Wright was apparently quite a local character, and I remain somewhat suspicious, as he may have had ulterior motives; however, the similarity to one of the unpublished images from the Singer Tract is strong, and in the pre-internet era, it seems unlikely that a Texas woodsman would be familiar with any of the ivorybill images that did not appear in Tanner’s monograph.

The Fielding Lewis photographs, taken in South Louisiana in 1971, are far better known, although Lewis wished to remain anonymous. His identity was not revealed until more than three decades later, and he was identified at the time only as “the Chief” (Jackson was unaware of his identity in 2004, though it was made public soon thereafter.) George Lowery presented these images at the American Ornithologists’ Union annual meeting and was attacked by many of his colleagues who deemed them to be a hoax. It’s worth pointing out that there have been hoaxes in recent years (these were swiftly and easily debunked), so there’s legitimate reason for skepticism about any claim. In the case of the ivorybill, however, skepticism has frequently been replaced by a virtually irrebuttable presumption of fraud (or error).

In a letter to Tanner, Lowery (who, unlike most of his detractors, had actual field knowledge of ivorybills) wrote: “I know the man in question very well and I am sure he would not pull something like that. In the first place, where would he have gotten the mounted specimen? Why would he have two photographs of the birds way up on two separate trees? Both of considerable diameter and not subject to being shinnied. Also, assuming he might have had a mounted bird to photograph, why didn’t he get a better picture while he was at it?”

The most common reasons given for believing that Lewis hoaxed the photos are that the bird is similarly positioned in both images (similarly but not identically, a neat trick with a specimen) and that the bill and feet are not visible. Bear in mind that these pictures were taken with a Brownie or Instamatic camera, and the quality is very poor. Nonetheless, I believe a foot is very faintly visible in both photos, positioned in a manner that would be expected of a live Campephilus woodpecker. I worked from scans of the originals and made some modest enhancements in Photoshop. Note also that in the first picture, both a cavity and bark scaling can be seen on the tree.

Regardless of whether my analysis is correct, the foundation for claiming that these photographs were faked is shaky indeed – not only for the reasons Lowery gave but also because of Lewis’s reticence about revealing his identity, insistence on keeping the location secret, and lack of any discernible motive.

The recovery plan mentions two other pieces of physical evidence. One is a feather that was purportedly found in a nest or roost cavity in 1985, and the other is a photograph supposedly taken in Georgia in 1975 (correction, 1965).

It’s not my intention to fault the IUCN or to quarrel with the “possibly extinct” designation, although I think Bill Pulliam’s analysis of the Luneau video in light of Rhein’s Imperial Woodpecker film should be dispositive. My main purpose is to call attention to the fact that there’s considerably more evidence for survival than is popularly recognized or than appears in the scientific literature. There are multiple instances of people, often outstanding birders, hesitating to come forward with reports and hard evidence – for decades. Terres waited 30 years; in the cases involving physical evidence, only Agey and Heinzmann went fully public and identified themselves in making their claims; in Stalking the Ghost Bird, Steinberg includes detailed 2005 field notes written by a Louisiana “birder for more than forty years” who “has also worked as a contract ornithologist conducting bird surveys on rice and crawfish farms for more than ten years.” This individual too requested anonymity. To a significant extent, the shortcomings in the literature are directly related to the controversy that has surrounded this species for nearly a century (remember Mason Spencer who went so far as to obtain a permit and kill an ivorybill to prove the species persisted) and the accompanying climate of unhealthy skepticism that has, if anything, grown even more unhealthy since the rediscovery. This sets the ivorybill apart from other possibly extinct species.

I’ll be heading for Louisiana at the end of the week and will post a trip report when I return.