Like many people, I first saw the published evidence of the Arkansas reports on 28 April 2005, when a paper by Fitzpatrick et al. was published online by Sciencexpress. That online version was replaced by a print version published 3 June 2005 in Science magazine (available here; some key differences in wording exist between these versions, which I will describe elsewhere). After a week of careful review and a few email exchanges, a group including Michael Patten, myself, and three others wrote to John Fitzpatrick, Ken Rosenberg, and Van Remsen on 6 May 2005. We outlined several questions that we felt had not been answered beyond a reasonable doubt. At that time we were still encouraged by the possible audio evidence but deeply concerned that the bird in the “Luneau video” had been misidentified. This remained a private dialogue, with helpful exchanges on both sides, and remains to this day a collaborative, constructive effort. Our aim was to seek a more accurate public portrayal of how secure the identification was but to allow those involved to decide how to handle this.
On 3 June 2005, I learned that David Sibley had independently questioned the identification along similar lines following his searches in Arkansas and subsequent review of the evidence. At the same time, I learned that Rick Prum, Mark Robbins, and Jerry Jackson had the same concerns independent of David Sibley’s. As with each of my colleagues in the initial critique, each group had arrived at these conclusions independently.
In late June 2005, I visited the Cornell Laboratory of Ornithology and spent a full day reviewing the evidence with John Fitzpatrick, Kim Bostwick, Russ Charif, Marc Dantzker, Martjan Lammertink, Curtis Marantz, Michael Patten, Ron Rohrbaugh, Ken Rosenberg, and Brian Sullivan. We studied the Luneau video in detail, listened to and discussed all the available audio evidence, including past recordings such as John Dennis’s recording from the Big Thicket in Texas, and other materials in the vast collection in the Macaulay Library. The discussion was productive, and I am still deeply appreciative to everyone for this opportunity.
Meanwhile, a paper by Rick Prum, Mark Robbins, Jerry Jackson, and Brett Benz addressing the video evidence and sightings was prepared and submitted to the Public Library of Science, an online journal. This paper was withdrawn after Prum and Robbins listened to some newly analyzed recordings from Arkansas. That they believed these recordings probably were Ivory-bills was understandable, even if not correct or not possible to determine from the audio evidence. Their withdrawal meant that going into the meeting of the American Ornithologists Union in August 2005, no formal critique of the evidence was to be published. Following this meeting and the apparent dismissal of our views, David Sibley and I along with Michael Patten, Chris Elphick, and others undertook a detailed analysis to examine whether the bird in the Luneau video was correctly identified. We sought to verify independently the identification by analyzing the same data. Images from the Luneau video were processed following the same protocols listed by Fitzpatrick et al. 2005. Use of these images is granted under special permission from the Macaulay Library at the Cornell Laboratory of Ornithology.
Video comparisons
Videotape of three juvenile Pileated Woodpeckers was obtained both in a flight cage and when released in late summer and early fall of 2006. These birds were healthy and capable of normal flight. They were cared for by Marc Payne and Diane Winn at their outstanding rehabilitation center, Avian Haven, in Freedom, Maine. Following release, these birds were observed in the neighborhood for several weeks, where they continued to appear normal and healthy. Additional footage of Pileated Woodpeckers was obtained elsewhere in central Maine.
Video was recorded to mini-DV tape with a Canon XL2. Analysis of raw video was performed with iMovie HD, Adobe Photoshop (deinterlacing, interpolation, and resampling for zooming images), and Quicktime Pro. For critical comparison, camera settings were set to replicate those derived from the raw DV data found the digital copy that David Luneau deposited at the Macaulay Library, Cornell Lab of Ornithology. The video copy comprises only 42 seconds and does not constitute a complete copy of the mini-DV tape. At the time I inquired, neither the original mini-DV tape nor a direct clone had been deposited for archival and research purposes.
Settings: shutter 1/60 second, f-stop 4.8 (published here for the first time; it was retrieved from the DV information when played back on a Canon video camera), interlaced NTSC at 29.97 frames per second (480i format). Lens set to widest angle, 42.3 mm on XL2—the GL2 used by David Luneau has a slightly wider angle zoom, i.e. slightly less telephoto, of 39.5 mm. Both the GL2 and the XL2 have 3 CCD elements. The XL2 has a slightly different light sensitivity, having a lux value of 5.5 versus 6.0 for the GL2.
Flight cage: Three individual Pileated Woodpeckers raised from young birds were kept in a 40’ L x 16’ W x 12’ H flight cage. The birds were in juvenal plumage and fully capable of normal flight. Several launch sequences were recorded confirming precise wing and tail movements that match the Luneau video. Deinterlaced video fields match precisely launch sequence from Arkansas in terms of timing of wing movement and reduced or blurred out black trailing edge to underwing). These launch mechanics were identical to free-flying Pileated Woodpeckers videotaped elsewhere in Maine.
Releases: Birds were released 20 meters from plane of camera. Camera height and height of bird at release were set to replicate the situation as reported in the Arkansas recording: camera 0.67 m above water and bird 1.3 m above water (angle +1.8°). Focus was set and measured with grid at 2 meters, as was the case for the Luneau recording.
The camera was on a stable tripod on the ground rather than a moving canoe on water. The birds released were over ground that sloped away, a constraint of my available release sites. In this way, the nature of flight, positions, and relative motion of bird and camera differed from the conditions in Arkansas. Additionally, lighting was sunnier with partly cloudy skies and birds released in the open. Unfortunately, both lighting and direction of flight—more or less directly away—are critical to replicating patterns on video. Why is that important? Because the pattern of black secondaries is mostly obscured or vanishes in directly caudal views of Pileateds. My experiments reveal that plumage patterns on known Pileateds replicate those on the bird in the Luneau video.
This might have meant birds did not have to flap as hard to maintain or gain elevation. However, because birds were released by hand, they did not have the benefit of their own thrust at launch. Comparison with video of natural woodpecker launches shows they are variable, but when birds need to maintain or gain elevation, thrust with the feet and tail (invariably the last support to leave the substrate) are typical. Woodpeckers that descend from a higher perch may use their height, and hence potential energy, to launch in a way different from that seen in the Luneau video.