Introduction

While I didn’t fully apprehend it at the time, the events that took place over several days in mid-March 2017 were singular. Crucially, the set of recordings made by Matt Courtman (formerly affiliated with Project Coyote) and Phil Vanbergen is unique in the history of ivorybill searches. Project Principalis’s paper included a discussion of this audio, and I have posted extensively about these recordings in the past. Nevertheless, it seems important to revisit them in the context of this series. From what I understand, a very few people, one well-regarded ornithologist among them, consider the audio excerpts the most compelling evidence in the paper. But overall and unfortunately, this important line of acoustic evidence seems to have gotten minimal attention.

Acoustic evidence may be the biggest casualty of the ivorybill controversy. The privileging of visual material, especially video, over other forms of documentation and the demand for “proof”, can lead to the virtual erasure of important scientific data, potentially at the expense of future conservation efforts. In the current climate, it seems unlikely that even a perfect match for the Singer Tract recordings or a series of recorded double knocks consistent with other Campephilus DKs would suffice as “proof” in some quarters, absent an accompanying video or series of photographs.

The Structure of Ivorybill Vocalizations

Paul Kellogg, who co-led the 1935 expedition with his longtime colleague/coauthor Arthur A. Allen and Albert Brand, engineered the recordings. Famous photographs show him perched on an oxcart near a nest, while Tanner stood nearby holding a large parabolic microphone. This activity, along with Allen’s presence even closer to the nest and waving his hands (as he says on the recording), disturbed the birds. Though Tanner publicly insisted that no disturbance had taken place, his wife Nancy said otherwise in a 2006 National Geographic interview (now behind a paywall), indicating that the birds were too agitated to double knock. Regardless of what Tanner believed, the circumstances, including the fact that the 1935 nest failed, are consistent with disturbance.

As noted, the 1935 team recorded no double knocks. In addition, according to Tanner, they did not capture one common, two syllable call that he transliterated as “kient – kient – kient”. In addition to this two syllable call, Tanner identified a conversational call he transliterated as “yent-yent-yent”. He seems to have overlooked at least one other vocalization, a soft, flicker-like call that appears at approximately the two minute mark on the Singer recordings. Given the context and apparent agitation of the John’s Bayou pair, I question the description of the calls I transliterate as “wonka-wonka” as conversational or “social chatter”. The Macaulay Library site describes them as such, probably following Tanner. These calls, examples of which can be heard starting at 3:14, may represent a form of scolding, while the softer, flicker-like calls may be the sounds he called conversational.

In the monograph, p. 62, Tanner described the recorded and unrecorded calls and pointed to examples of kents and conversational “yent-yent-yents” on film strips. He didn’t reference the sound recordings themselves. Regardless, the structure of all the recorded calls, a simple ladder with harmonic partials is consistent across call type.

A relatively weak second partial harmonic characterizes the Singer Tract sounds, the John Dennis recordings, and the Courtman and Vanbergen recordings. Emphasis varies among the 3rd, 4th and 5th partials but is frequently on the 3rd, a tendency that is most pronounced in the Courtman recordings. This pattern can be seen in the image above of a sonagram made from the Singer Tract 1935 recordings; the brightest spots indicate the emphasis. Note that the second partial is consistently weak or even undetected and that the strongest emphasis in this selection is most frequently on the third.

There are modest differences among the Singer, Dennis, and Courtman recordings. Our paper addresses these modest differences in greater depth. Regardless of differences, these sounds are not consistent with any other known North American species – bird, mammal, reptile, or amphibian.

Paul Kellogg on the Dennis Recordings: A Corrective to Tanner from a Senior Ornithologist

John Dennis recorded 17 calls over one minute and 47 seconds. The observers believed the birds to be approximately 150 feet away. This seems considerably closer than the birds in the Courtman and Vanbergen clips. Tanner was initially dismissive of the Dennis recordings. He justified his doubts by pointing to Pine Warbler vocalizations on the tape. He was suggesting, wrongly – indeed almost ridiculously – that the habitat was unsuitable. I recently learned that Paul Kellogg had a very different opinion, something I had never seen in the literature.

Kellogg was a pioneer in the field of bioacoustics. He not only knew ivorybills, he made the Singer Tract recordings, listening to the birds through headphones atop an oxcart. In Kellogg’s view:

Agitation, Timbre, and Variability

Kellogg’s perspective is interesting for several reasons. “We really need better recordings of this species” surprised me. I only wish that Kellogg were still alive to elaborate.

Feb. 5, 2024 Edited to Add: Chuck Hunter informs me that this was a reference to the poor quality of the Dennis recording only. He also pointed out that Kellogg’s took a somewhat dimmer view of the Dennis recording than the letter suggests. Per people who were at Cornell at the time, Kellogg suspected it might have been faked because when played at slow speed, it “sounded like a pencil being rolled across a table”. He also pointed out a misstatement, now deleted, that Kellogg had seen and heard ivorybills in Florida.

But Kellogg’s perspective is important not only for its emphatic take on the Dennis recording but also for the statement that “individual variation is not great.” While not great, some variation clearly exists based on available data. Though Tanner couldn’t consistently distinguish between male and female vocalizations, Allen and Kellogg suggested otherwise in their 1937 paper. Their field notes described the female’s voice as “weaker and less harsh.”

Tanner did note variation between individuals. He wrote that Mack’s Bayou Pete, a lone individual that inhabited another home range, made distinctive and identifiable sounds in 1938. Pete’s calls had a “peculiar timbre” and a “head in a bucket” quality (p. 33). Tanner says nothing about Pete’s voice in 1939 and may not have heard it; he mentions only one brief encounter that season. In 1938, he assumed that Pete was a juvenal, based on plumage, and speculated (in his dissertation) that he “may well have been” the 1937 offspring of the John’s Bayou pair.

This speculation about Pete’s origins may be accurate, but whatever the case, Pete could not have fledged in 1938. Tanner noted that kents made by the Sonny Boy (John’s Bayou, 1938) “soon became indistinguishable” from adult calls. In context, “soon” apparently means within approximately one month. This strongly suggests that even as a yearling or two year-old, Pete would have had an adult voice, albeit a distinctive one.

The available evidence makes it clear that adult ivorybill vocalizations vary somewhat in timbre, as reflected in the harmonic emphasis, but have a consistent harmonic structure. This variability apparently exists within individuals, as the Singer Tract recordings show, and between them, as indicated by Mack’s Bayou Pete’s vocalizations and Allen and Kellogg’s description of sexual divocalism, to coin a phrase.

As to duration, Tanner wrote, “The kent note, given in monotone and slowly or infrequently, is the ordinary call note. When the bird is disturbed, the pitch of the note rises, and it is repeated more rapidly, frequently doubled.” Agitation (Singer birds most disturbed, Dennis birds less so, Courtman/Vanbergen birds most relaxed and “ordinary” and therefore longest), individual variation, recording equipment, and attenuation by distance likely suffice to account for the modest differences among the recordings.

Conclusion

I’ve just republished a couple of posts from 2017. One details the events of those few days in March and a return visit at the end of that month. The other includes many of the recorded sounds (and amplified versions). The discussion reflects my understanding at the time. Some of my views may have evolved.

The Courtman and Vanbergen recordings were made on two different days. The longer set of recordings, made on March 15, is unprecedented in several ways. It should stand as compelling, and arguably, conclusive evidence of ivorybill presence on the days in question. I say this for a few reasons.

1. The Courtman recordings are unique in ivorybill history. They include approximately 200 kents that are harmonically identical to some Singer Tract kents. They also include a few double knocks, one in apparent response or reaction to an ADK.
2. The Courtman recordings appear to involve at least two birds.
3. Multiple observers heard the calls. Recordings were made on two separate days, March 11 and 15, 2017. On March 11, we heard single and double knocks a couple of miles from the recording site. In the afternoon, we heard (and Phil Vanbergen recorded) the first kents. On March 12, multiple observers heard single and double knocks at the same location. On March 13, we returned to the first location and heard impressive single and double knocks in response to ADKs. Weather was poor on March 14, and we heard nothing.
4. While we have heard isolated possible kents at times, the events of March 2017 were unique for the area.
5. On a visit later in March 2017, two observers spent several full days at the location. We heard possible double-knocks and perhaps a single kent-like call. But we heard nothing remotely approaching what we had encountered in mid-March.

In combination, these unique qualities should exclude any alternative explanation whether natural or mechanical – playbacks by an unknown interloper, Blue Jays, White-breasted or Red-breasted Nuthatches, eastern gray squirrels, frogs, tree squeaks, duck wings, road noise, etc. I have little doubt that Paul Kellogg would have agreed.

The full import of these recordings will remain unclear unless and until someone finds a nest and a population can be studied. I have a hunch that the abundance of calls in March 2017 was due to nesting activity and possibly to fledging. Like some of the trail cam imagery and along with the unusual number of knocks during this 4-day period, it could indicate sporadic use of the core study site, or some combination of the two. There may be other explanations. But from an evidentiary standpoint, I have to rank the recordings as exceptionally strong. I thought so at the time and am even more convinced at almost seven years distance.

I’m reposting some selections from March 2017. The first is Phil Vanbergen’s capture of the first calls from March 11, 2017. I have also included a couple of brief extracts from Phil’s captures on the 15th. These overlap with Matt’s captures from the same day, an almost 16-minute extract of which follows Phil’s. More clips are at the links to the republished posts and here.

Thanks to Don Scheifler, Tommy Michot, Patricia Johnson, and Bob Ford for their contributions.