Ivory-bill Absent from Sounds of the Bayous
More than 4,000 hours of recordings analyzed


Technician Chris Tessaglia-Hymes identified the recorded sounds by listening to them and inspecting the spectrographic images.
Diane L. Tessaglia-Hymes
In February 2002, a six-member search team funded by Zeiss Sports Optics emerged from the Louisiana bayous, unrewarded but hopeful. They had not found an Ivory-billed Woodpecker during a month of searching, but they had heard and tape-recorded several loud double raps that were reminiscent of the species long feared to be extinct. Meanwhile, we hoped that 12 autonomous recording units (ARUs) deployed by the Lab's team would capture the sounds of the magnificent woodpecker if it still existed there.

We recovered 4,146 hours of data recorded at 12 dispersed locations in the bottomland forests of the lower Pearl River, Louisiana. Sadly, after scrutinizing more than 130,000 sounds deemed most likely to have been produced by an ivory-bill, we can state with some confidence that no Ivory-billed Woodpeckers were present in the area between January 25 and March 17, 2002.

Although we are disappointed to have found no evidence of ivory-bills in what we believed to be the most promising terrain, we are pleased that the ARUs functioned as intended. Each one recorded sounds unobtrusively and continuously during two four-hour sessions each day. Our analysis also allowed us to pinpoint and examine the mysterious double rap recorded by the Zeiss team.

a. Ivory-billed Woodpecker
b. Pileated Woodpecker
c. Red-shouldered Hawk
d. Carolina Wren
Each sound produced by a bird leaves a recognizable signature in a sound spectrogram. Above, (a) Ivory-billed Woodpecker vocalizations and knocks from the 1935 recording, (b) Pileated Woodpecker wek wek wek contact notes, (c) Red-shouldered Hawk keeyuur calls, (d) Carolina Wren song.
With detection software developed here at the Lab, our technician, Chris Tessaglia-Hymes, could analyze the recordings in a matter of months, compared with the two years that would have been required otherwise. Our program detected and marked all sounds that could have been made by an ivory-bill, based on the only Ivory-billed Woodpecker recordings in existence, made by Lab founder Arthur Allen and codirector Peter Paul Kellogg.

For each of the 12 ARUs, automatic detection analysis identified thousands of sounds that had some Ivory-billed Woodpecker characteristics. This level of "false positive" was desirable because we preferred to search through many such erroneous detections rather than set the parameters too tight, risking that a true positive might not be identified.

Chris inspected the spectrographic image of each sound detection on a computer screen and listened to audio playback through headphones. The most common detections included songs and calls of Carolina Wren (Thryothorus ludovicianus), Tufted Titmouse (Parus bicolor), Red-shouldered Hawk (Buteo lineatus), Pileated Woodpecker (Dryocopus pileatus), Red-bellied Woodpecker (Melanerpes carolinus), spring peeper (Hyla crucifer), and gunshots.

We paid special attention to the acoustic events of Sunday afternoon, January 27, a day when the Zeiss team and our team independently heard a double-rap sound reminiscent of the display drumming of the Ivory-billed Woodpecker. We carefully inspected and listened to the acoustic signals detected by ARU #10, the closest unit, and easily identified the time period coinciding with the sounds we had heard.

The signals in question had an explosive onset equally represented by a wide band of frequencies, followed by a tapering "smudge." This signature is clearly that of a gunshot, with the smudge representing decaying reverberations of the sound through the forest. We detected the same sound, although faint, at ARU #9, a distance we deemed too far for a woodpecker drum-rap, of any species, to travel through a full-canopy bottomland forest and still be detected by our ARUs. These acoustic signatures exactly matched those of hundreds of other presumed gunshots recorded by our ARUs and dashed all hopes that the mystery sounds would prove to be the long-lost woodpecker.

Although we concluded that no Ivory-billed Woodpeckers were present during the period monitored, we still view it as possible (though the chances are slim) that one to several pairs of ivory-bills could be using portions of the Pearl River forests outside the scope of our search. Ivory-billed Woodpeckers were highly mobile, sometimes moving great distances as their principal resources (large, recently dead trees) became available or aged beyond use. The bottomland hardwood forests of the lower Pearl River are extensive, and they are in better condition to support large woodpecker populations today than they have been for a century or more.

There have been no confirmed sightings of Ivory-billed Woodpeckers in the United States for more than 50 years, but there have been persistent accounts bearing a reasonable level of credibility. The history and profile of this great bird, the tragedy of its loss, and the importance of any potential rediscovery all demand a truly concerted effort by scientists willing to conduct a careful and objective search. We hope to contribute to that effort by conducting at least one more search using ARU technology at a new location in the future.

Suggested citation: Fitzpatrick, John W., Ivory-bill Absent from Sounds of the Bayous. Birdscope, newsletter of the Cornell Lab of Ornithology, Summer 2002. <>

For permission to reprint all or part of this article, please contact Miyoko Chu, Editor, Cornell Lab of Ornithology, 159 Sapsucker Woods Rd., Ithaca, New York. Phone (607) 254-2451. Email