Historic Range Hypothesis: A Bits ‘n’ Pieces Epilogue

In retrospect, I realize that I could have stated my hypothesis about historic range more explicitly in the Bits ‘n’ Pieces series, here, here, here, here, here, and here. My treatment of this subject draws on and expands upon Tanner, Jerome Jackson’s extensive review of the historical record, the work of Benjamin Leese, and Appendix E in the U.S. Fish and Wildlife Service’s Ivory-billed Woodpecker Recovery Plan, supplemented by my own research and review of the source material.

I got somewhat too focused on specimen records and even more so on habitat types. The version of the map I created for the series reflects that focus, by including barrier islands, for example. The subjects of range and habitat are of course intertwined, but I’ll state the hypothesis with respect to range upfront and follow it with a more detailed discussion:

The Ivory-billed Woodpecker’s range was considerably more extensive and than is recognized by the general public and than has been represented in much of the literature, both popular and scientific. All published range maps of which I’m aware, including Hasbrouck’s (1891), Tanner’s (1942), Jackson’s revised and redrawn version of Tanner (2004), and online, poorly reflect the historic range, especially east of the Mississippi.

This has implications in terms of habitat requirements and adaptability as well.

I suggest that the northern limit of the ivorybill’s range was just above 40 degrees North and that the western limit was approximately 98 degrees West. There are no archaeological or other records from over 2000′, so I’d exclude higher elevations in the Appalachians and Ozarks. I think the previous maps are accurate to the extent that they show the range as extending farther west in the Red and Arkansas River basins than in the Missouri.

To express it somewhat differently, I think ivorybills could be found as far north as the lower Delaware River on the Eastern Seaboard and that they could be found in riparian corridors into the lower reaches of the Appalachians from there south.

In Florida, I think it’s conceivable that the range extended to the Upper Keys. In the southwest, I’d draw the line near Port Lavaca, Texas to west of Tulsa, Oklahoma. Except for an archaeological find in northeastern Nebraska that almost certainly involved trade items, the northernmost specimen record from west of the Mississippi is from Forest Park, MO (more on that below).

The story seems to have been somewhat different in the Ohio River watershed, where both archaeological evidence and sight records suggest that birds were present far upstream from the Mississippi confluence and along tributaries well north of the river itself.

I’ve created another map showing both the limits of my hypothesized range and selected reports from what could be considered the edges of that range. A screen cap is below. Click on this link for annotations. Although some of the records are pre-Columbian (the earliest estimated as dating to 0-200 CE/AD), I suspect this was the approximate range until the mid to late-18th century.

Screen Shot 2018-09-23 at 8.41.28 AM

Hypothesized limits of Ivory-billed Woodpecker range and selected records from areas outside those shown on most other maps.

Thus my hypothesized range is considerably more extensive than the conventional one (shown immediately below). On the Eastern Seaboard, it extends much farther north and also farther inland. It is far greater within the Mississippi watershed, extending to north of Columbus, Ohio, and encompassing all but the higher elevations in the Appalachians (based on the archaeological site at 2000′ near Lebanon, VA), reaching farther north and west in Arkansas and Missouri.

It was also somewhat more extensive in Texas and Oklahoma, reaching farther into the plains and a little farther south than the maps suggest. Outside of coastal areas, the presence was probably limited to riparian corridors, with those in the western part of the range (at least) being narrow. The version below is from the IUCN Red List entry for the ivorybill.

Screen Shot 2017-12-22 at 3.42.31 PM

IUCN Map derived from Jackson’s revision of Tanner’s 1942 publication. Similar maps are widely reproduced.

Screen Shot 2017-10-08 at 4.05.13 PM

Hasbrouck’s Ivory-billed Woodpecker Range Map (1890). Shaded areas reflect what Hasbrouck believed to be the range at the time of writing.

While what I’m suggesting may seem heretical to some, it’s well-supported by the archaeological record, accounts from early explorers and naturalists, and collection records or reports thereof, albeit to a lesser extent. I covered many of the records that led to my formulating this hypothesis in previous posts; I won’t recapitulate all of them here, but one passage is worth revisiting. Two notes on technical terminology: middens are most easily defined as “dumps for old domestic waste“, and tarsometatarsi are the lower leg bones found in birds (and some dinosuar fossils).

If our knowledge of the Ivory-billed Woodpecker were based on the archaeological record alone, we’d think of it as an upland species. Further, we might very well assume that it ranged from the hills of Georgia, to the Alleghenies in Virginia, to central Ohio and west-central Illinois. 

The remains found in Native American middens were unlikely to have been trade goods; ivorybill parts seem to have been a valuable commodity for ceremonial use west of the Mississippi but not east of it, and in several cases, the remains found were tarsometatarsi, which would be consistent with use as food:

There is strong physical evidence of  ritual value for woodpecker scalps and bills from the upper Midwest and Plains . . .  Remains of the Ivory-billed Woodpecker can be found in sacred bundles, on pipe stems, on amulets, and with burials among the Native Americans of the region. The evidence comes from the western Great Lakes and the Plains; no evidence of a particular use of Ivory-billed Woodpeckers has yet been uncovered from the eastern area of the Great Lakes (Ohio, Indiana, and Michigan).

(Leese, 2006.) Leese also points out (in several of his publications) that there’s no evidence that ivorybill parts other than scalps and bills had any trade value.

In an unpublished 1989 update to his monograph (housed in the archives at Cornell), Tanner accepted additional reports, including archaeological finds from Scioto County, Ohio (15th or 16th-century) and Madison County, Illinois (Cahokia Mounds, approximately 15th-century), as well as a collection record from Forest Park, Missouri and an 18th century record from Lincoln County, Kentucky. Since that time, additional archaeological records from Georgia, Virginia, West Virginia, and Ohio have turned up.

Thus, the case for ivorybill presence both in the lower reaches of the Appalachians and well up the Ohio River (and its tributaries) is compelling, and any suggestion that items found in Native American middens might be trade goods is pure speculation with no evidence to support it.

In a paper titled, “Putative Records of the Ivory Billed Woodpecker in Pennsylvania and the Mid-Atlantic” (2016 Pennsylvania Birds, 20(2):71-72), Leese suggested that eastern Pennsylvania and Southern New Jersey might be the northern limit on the Atlantic Coast. I agree and might even go a step further and argue that the available evidence strongly supports that view. The discussion below relies heavily on material Leese discovered, though I disagree with Leese’s conclusion that a more extensive range and more varied habitats support Snyder’s argument that hunting was the “main cause of the species’ extinction”.

This idea that ivorybills ranged so far north on the East Coast will probably be the hardest for many to swallow. Bear in mind, however, that Audubon described the ivorybill as an accidental in Maryland. Audubon’s discussion of range is odd; he didn’t mention Florida, and there’s nothing to support the idea that the western part of the range extended to the foothills of the Rockies. But he would have been far better informed about the Eastern Seaboard.

In any event, his description dates to the 1820s, by which time the range was likely shrinking. As discussed previously in the series, I suspect that the extirpation of the beaver played a major role in the ivorybill’s decline outside of Florida. Beavers were probably extirpated in New Jersey by 1820 (Van Gelder 1984). It’s reasonable to infer a similar date for eastern Pennsylvania.

To expand on this beyond the previous posts in the series, Peter Kalm, a prominent European naturalist and student of Linnaeus, listed the ivorybill as present in the Delaware Valley and distinguished it from the Pileated, which he described as “a Black Woodpecker with a red head” and “frequent in the Pennsylvanian forests”.

It’s not clear whether Tanner was aware of Kalm’s 1749 record, which was described in The Auk in 1903. The author of that paper suggested that this record should be taken “cum grano salis,” primarily because there have been no other records from the state. Or have there?

While it is ambiguous, Leese discovered a letter from Alexander Wilson, written during the early years of his career, when he was living on William Bartram’s property near Philadelphia; it is strongly suggestive of ivorybill, describing a large and “most extraordinary Blackheaded Woodpecker”. This was at a time when Wilson seemed to be in the process of learning his birds, and he did not mention it in later correspondence, which leaves room for doubt.

Screen Shot 2018-08-19 at 7.31.28 AM

While Wilson’s description is not very detailed, I find it suggestive. And as noted previously, I think the only basis for rejecting Kalm’s account (and he seems to have otherwise been meticulous) is what we think we know about the ivorybill, a knowledge base derived from observations dating to between 60 and 190 years later. While there’s no way to prove it, I think Kalm’s listing should be treated as credible; it would be if a more common species were involved. Leese mentioned a couple of additional 18th century reports from Pennsylvania (included on the map) but found them questionable.

I’m not aware of anyone else having made this observation, but I think it’s a very important one. In the 1740s, the ivorybill had not yet acquired the mystique that would accompany it from the mid-1800s to the present, a mystique that was popularized if not invented by Audubon. This lends greater credibility to earlier reports such as Kalm’s. I think the same principle supports Jefferson’s listing of the ivorybill as resident in Virginia (which may have been based on his own observations) a couple of decades after Kalm and a few decades before Audubon.

To return briefly to Wilson and another report from an unexpected location, Jackson (2004) references a specimen collected by Wilson from somewhere between Winchester, Virginia and Martinsburg in what is now West Virginia (presumably in the Potomac watershed). I have not located the primary source for this reference, which may be in the archives of the Philadelphia Academy of Natural Sciences, but here’s Jackson’s:

Screen Shot 2018-08-18 at 10.43.04 AM

The Putnam County, WV find is also interesting, as it suggests a bird that may have been collected for trade rather than being acquired through it.

I want to touch briefly on two additional records, both involving specimens. The facts related to these records point to how even specimen evidence is not free from ambiguity. For one thing, specimen tags often reflect shipment rather than collection locations. In addition, labels could have been changed and replaced at any point in the chain of custody between hunter and museum.

The first specimen is housed at Cornell (and I have seen it myself). It was collected in 1898 from the “Florida keys”. In Appendix E of the Ivory-billed Woodpecker Recovery Plan, it is suggested that this might actually refer to one of the forested “keys” in the Everglades rather than to one of the offshore islands. However, given the ivorybill’s use of barrier islands on the Atlantic Coast and the 10,000 Islands in Florida, it does not seem farfetched that the specimen could have been collected somewhere in the Upper Keys proper.

From my current perspective, based on the Recovery Plan’s information about habitat on the southwest Florida coast, both possibilities seem plausible and neither is particularly surprising. However, neither location would ever have had many characteristics in common with stereotypical ivorybill habitat –  vast, contiguous bottomland hardwood forests. (Although he sometimes seemed to embrace this stereotype later in life, Tanner recognized that ivorybills lived in varied habitats, as have others who have studied the subject in depth.)

The other specimen, labeled  as being from Forest Park, Saint Louis, Missouri, escaped Tanner’s notice initially but was included in his 1989 update. It is housed at the Colorado Museum of Natural History in Denver and was collected in 1886. Matt Courtman pointed out that by 1886, Forest Park was hardly an isolated spot; it attracted “hundreds of thousands” of visitors a year,was located on the outskirts of a city with a population of 400,000, had been at least partially landscaped, and was accessible by streetcar.

So not only did this collection come from well north of what’s commonly accepted as the historic range; it may have come from a managed parkland on the outskirts of a major city, hardly stereotypical ivorybill habitat. The location and date are surprising, so it’s possible that the specimen was collected somewhere else (though not in a different region) and that Forest Park was the shipping location.

This may be so, but an 1886 collection from anywhere near Saint Louis is a paradigm-buster, regardless of the exact location, as this 1884 map suggests. Forest Park, to the southeast of Florisant and Ferguson, which are indicated by the red circle on the map, is not shown.

Screen Shot 2018-09-21 at 1.32.16 PM

Detail from 1884 map of Missouri, showing Saint Louis and environs. Note both the network of railroad lines and the apparent population density.

Tanner’s original map showed the ivorybill as never having ranged farther north than the Missouri bootheel (the southeastern corner of the state) and as having been extirpated from all but the southernmost tip of that region by 1880. The last specimen collected in Missouri dates to 1895, from just north of the bootheel “eight miles southwest of Morley.” (Widdman, 1908). Tanner seems to have been unaware of this record, which is mentioned in Jackson, who could not locate the specimen itself.

Tanner received at least one report (1937) from Dallas County, Missouri, which is northwest of Springfield, in the upper (southern) reaches of the Missouri River basin and well outside the generally accepted historic range. The source was a local Audubon Society officer, and the letter is archived at Cornell. There appears to have been no follow-up. According to the Recovery Plan, reports from this general area continued until 1949.

Though his map of the range ca. 1890 suggests a northernmost limit in Mississippi and Louisiana, Hasbrouck (citing Cooke, Bird Migration in the Mississippi Valley) mentioned 1884 records from near Kansas City and undated claims of former breeding near Fayette, along the river in central Missouri; Tanner considered these questionable or accidental. I agree that they’re somewhat questionable but not that they could have been accidental, since both purportedly involved breeding.

Reading Cooke reveals that both were second-hand accounts. The Kansas City report includes no details except that birds were observed “during the past few winters”, and “it probably still breeds in that vicinity”. For Fayette, the source was identified as a local farmer and egg collector by the name of Lientz, but the former breeding claim is devoid of any details, including the informant’s name.

Thus, the possibility of mistake or miscommunication exists in both instances, as in the seemingly credible report from eastern Nebraska discussed in Part 2; however, Jackson also references Harris’s Birds of the Kansas City Region (1919), which lists a “Judge Guinot” and others” for the Kansas City area records, in the “deep woods of the Missouri bottoms”. No date is given, but it seems possible if not likely that Guinot (1855-1935) was Cooke’s source

While I think the historic range probably extended as far as Kansas City, the case seems a little weaker to me than it does for the eastern seaboard and Ohio Valley –  given the paucity or absence of archaeological evidence or early reports from the Missouri River watershed, a key route for explorers, traders, and early settlers. For example, Featherstonhaugh, who crossed the Missouri River basin near the confluence with the Mississippi in the 1830s, described the northern limit as being considerably farther south, in the Arkansas-White River watershed. But perhaps there’s more to uncover about the ivorybill’s history along the Missouri.

As with anything ivorybill related, it’s probably best to be comfortable with not knowing and even to revel in the uncertainty.

How The Ivory-billed Woodpecker Might Have Survived

I have been corresponding with Christopher Carlisle, a Mississippi birder who has recently started searching for ivorybills in the Pascagoula watershed. These exchanges, along with a post on IBWO.net about the Choctawhatchee, led me to start thinking, in a somewhat more methodical fashion, about habitat conditions and how the ivorybill might have survived and to consider the significant differences between the old Project Coyote search area and the new one.

I suspect there may have been three behaviors, probably operating separately and together, that made survival possible and detection and documentation difficult. These behaviors are consistent with the historical record, which suggests the species was considerably more adaptable than the simplified reading of Tanner that treats “virgin” or old growth timber as essential. Tanner’s own views on this issue became considerably more dogmatic over time, but even his early surveys of possible habitat in the southeast were influenced by this belief, which I’d suggest was more cultural than scientific, since Tanner was a product of an era in which frontier mythology profoundly influenced American thinking. Tanner’s description of the Singer Tract and especially his ivorybills are reminiscent of Edward S. Curtis’s Native Americans, magnificent, worthy of honor, but ultimately doomed and incapable of adapting.

The reality about Native Americans was considerably more complex of course, and the same goes for the Singer Tract. It was by no means a vast tract of virgin forest in the 1930s; much of it had been under cultivation prior to the Civil War, less than a century prior to Tanner’s seasons there. In pre-Columbian times, the southeast had a significant indigenous population, with perhaps as many as 5.2 million people in the Mississippi Valley alone. Thus, much of what was deemed to be “virgin” forest in the 20th-century was quite likely agricultural land in the pre-Columbian era. As I see it, Tanner’s preconceptions about what constituted “suitable habitat” led him to be somewhat cavalier, even in the early days – dismissing areas based on very brief visits, most often due to lack of “virgin” forest. As a result of this prejudice, Tanner may have underestimated the IBWO population.

This is not to negate the argument that IBWOs are more specialized than PIWOs; their morphology clearly places them in a different ecological niche, and their considerably more limited historic range also points to a higher degree of specialization. The fact that IBWOs do not regurgitate may place limits on behavior and home range during breeding season. The specialization, however, has far more to do with foraging behavior than with habitat requirements.

The IBWO.net post suggested (and implied some direct knowledge) that Geoff Hill and the Auburn team focused too heavily on the main channel of the Choctawhatchee. A person calling himself “Ranger Mike” wrote:

“During the last expedition by Auburn here on the Choctawhatchee River Basin I noticed two things that in my amateur opinion could have been better-First every time I saw researchers, or others who hunted and fished the area saw them, they were far to close to the main river. I am well aware of how hard it is access the flood plain areas, but the birds will likely be closer to the banks in the floodplains where there are very large old growth long leaf interspersed with the large cypress and gum that has been missed by logging due to its inaccessibility. This transition zone will be good habitat, and in my very significant time in these in both work and hunting/fishing has produced some interesting sightings and sign, with nothing of interest in the more frequently traveled areas. Second, I don’t think much time was spent on East Island. Its a very remote and promising area, but I think it was avoided because it would have required traversing by foot instead of floating around in kayaks or canoes.”

While I have not verified this account with anyone on the Auburn team, the general observation makes sense and is consistent in some respects with what Tanner observed in the Singer Tract. Nesting habitat was primarily located along tributary streams, not the Tensas River itself.

Edited to add: I have never been to the Choctawhatchee, and so have no first-hand knowledge with which to assess Ranger Mike’s comment. My intention was not to criticize Dr. Hill or the search efforts but to highlight a possible survival scenario. Since posting, I’ve heard from a couple of people with more knowledge of the area; one emailer stated that East Island is “way overlooked”, so if anyone’s interested in revisiting the Choctawhatchee, that might be a place to start.

The two Project Coyote search areas are quite different in character and forest composition. The current search area is secluded, very mature and surrounded by a good deal of contiguous lower quality habitat. It is part of a major watershed but is somewhat distant from the main river system, in the floodplain of a tributary. The old search area was on a small parcel of private land, about 3/5 of a mile distant from a medium-sized WMA, with bean fields in between. There are a number of larger WMAs in the surrounding area. The habitat in this area is mostly if not all of lower quality than in the new area and is discontinuous. Frank Wiley described it as “pearls on a string”. Assuming that ivorybills are or were present in both areas, it seems likely that the differences in habitat require different behaviors, and this may in turn point to how the species managed to persist post-Singer Tract. The following may be the behaviors that made survival possible.

Survival Strategy 1. Seek seclusion in remnant stands of mature forest and areas that were selectively logged, with substantial tracts of contiguous lower quality habitat associated; core habitats provide sufficient food supply during breeding season, and surrounding areas, pine plantations, younger upland hardwood forests and the like, might provide additional food sources the rest of the year. Examples of this type of area may include the Choctawhatchee, possibly other Florida rivers, the Carolinas, and the Pearl.

Survival Strategy 2. Expand home range substantially to forage in degraded habitat, using forested corridors to move around whenever possible but traversing open fields when necessary. This would entail having roosts and possibly nest sites near field edges in some instances, as appeared to be the case in the old Project Coyote search area, near Patterson, LA , a few other Louisiana locations, and possibly Wattensaw WMA in Arkansas. This idea is in part inspired by the November ’48 report to Tanner from Gus Willett, Singer Tract warden, saying that he saw a pair of ivorybills at “North Lake #1” and that the “birds are moving over a much larger area than formerly.” This is the last letter to Tanner pertaining to the Singer Tract, and it points to the likelihood that the widely circulated story of Don Eckelberry’s encounter with the last female IBWO is just another facet of the mythology that surrounds the species.


Suspected recent nest cavity found by Frank Wiley near the edge of a bean field, East-Central Louisiana, 2009.


Black and white image of tree in which suspected nest cavity was found. Compare with the appearance of this nest tree from the Singer Tract.

cavity 2 report

Close up of suspected roost cavity in a willow at the edge of a bean field. This cavity was in the tree where we obtained a Reconyx image of a possible ivorybill, discussed here.

Survival Strategy 3. The IBWO can function as a “disaster species”, as Dennis and others have argued, having a high degree of mobility when necessary. “Disaster” in this context could be a slow motion sort of disaster as in the Singer Tract, where sweet gums were undergoing a die-off, so this could include elements of either 1 or 2.

Strategy 1 would help explain some of the difficulty finding, relocating, and documenting the species, since these are remote and inaccessible locations and sightings would most often take place outside of core habitats. Strategy 2 might also help explain why detection is so difficult. These areas are sparsely populated, birds would only be in the open briefly, and would not stay in one place for long, except during nesting season. And as Frank Wiley pointed out, the main human activity during nesting season is turkey hunting, and David Kullivan notwithstanding, turkey hunters are unlikely to see ivorybills, in part because of where they’re focused and in part because their methods would be likely to be disruptive and cause any birds in the vicinity to flee.