Archived Comments for:
Video analysis of the escape flight of Pileated Woodpecker Dryocopus pileatus: does the Ivory-billed Woodpecker Campephilus principalis persist in continental North America?
agreements and disagreements with Dr. Collinson's assertions
David Martin, Gladys Porter Zoo
21 March 2007
Some months ago, Dr. Collinson kindly sent me a copy of his paper for my comments, after it had been accepted. I told him I felt that it deserved to be published, although I had a lot of problems with it. Although the final paper has apparently been edited to address some of my concerns, many remain, and I feel I should present those here, as well as point out what I feel are some critical flaws in his analysis. I still feel that the paper deserved to be published, but there should have also been provided an opportunity for rebuttal from Fitzpatrick et al. Perhaps that will yet occur.
general remarks:
1) In their analysis of the Luneau video, Fitzpatrick et al. were careful to try to duplicate the conditions as closely as possible, using the same camcorder, focused in the same manner, at the same location, and in similar lighting conditions. It would not be too surprising to find that a different camcorder, focused differently in different lighting conditions would produce very different results even on the same subject. That being said, I think it is possible to make some valid comparisons between the Luneau and Nolin videos, even on plumage characteristics.
2) The Nolin video frames he has displayed are interlaced. The Luneau video fields have been de-interlaced. In some cases the interlaced frames produce very misleading plumage features. It would have been much preferable to use de-interlaced frames. Again, this in itself does not negate all of the analysis, but it makes comparisons in some of the figures quite suspect.
specific remarks.
The statement that Fitzpatrick et al. did not discuss Sibley’s “3 points,” is partially incorrect. They did in fact address 2 out of 3. Point 1 was that the Luneau bird appears to have black secondaries. Fitzpatrick et al. stated that, “The left wing is white to its base as it emerges from behind the tree...and in additional fields that show dorsal views during the latter stages of upstrokes.” Point 2 was that the Luneau bird has “particularly bright white primary bases.” This point was indeed unaddressed by Fitzpatrick et al. and I think is one of Sibley’s strongest.. Point 3 was that the Luneau bird appears to have black curving around the edge of the wing. Fitzpatrick et al. stated that, “The claim by Sibley et al. that in certain fields the black wingtips of the woodpecker show a curved shape suggestive of pileated fails to acknowledge that movement blur is most pronounced at the wingtips and that in other fields of the Luneau video the shape is more suggestive of ivory-billed.”
The statement that aberrant pileateds have been reported from the area is misleading, in that it implies that such birds might explain the Luneau video. The only aberrant pileated yet documented from the Big Woods is the “white” bird photographed in White River NWR. This bird bears no resemblance to the Luneau bird and no other aberrant pileateds whatsoever have been reported from the area, let alone documented.
Critical to Dr. Collinson’s interpretation of the Nolin flights is his suggestion that the first 4 wingbeats of the Nolin birds constitute “escape flight,” while the remainder reflects the birds slowing their flight as they approach the trees. This is quite arbitrary and not supported by anything in the videos. At one second post-launch all of the birds are far from the trees and show no indication of preparing to land. In every case the cumulative wingbeat rate at one second post-launch is quite elevated compared to the typical pileated rate of 5.2 Hz. Bill Pulliam and I have looked at the wingbeat rates without any presumption of when “escape flight” ends. Pulliam’s graph (http://static.flickr.com/70/203089268_5f62aa0750.jpg) and mine (http://home.att.net/~fangsheath/wingbeatrates3.jpg), which includes Nolin flight videos recently posted, both show that while a reduction in wingbeat rates is not perfectly smooth and steady in all individuals, the basic pattern is similar among pileated flights. Fitzpatrick et al. cited a takeoff extracted from the old Singer Tract recordings, and I have independently analyzed the sonagram of this recording. The Luneau bird shows a pattern strikingly similar to that of this recording. In analyzing videos of flushed black-bellied whistling ducks (http://home.att.net/~fangsheath/treeduckrates.jpg), I have found no evidence whatever of a change from “escape flight” to some other kind of flight within the first second post-launch. Some individuals landed a short distance away. Others left the area. Regardless, the basic pattern is similar among flights. While wingbeat rates sometimes exceed 8 Hz momentarily post-launch, among 30 flights the wingbeat rate at one second varies from 5.1 to 7.7 Hz. I find no support in either the pileated or whistling-duck data for this arbitrary 4-wingbeat cutoff. Indeed, looking at Pulliam’s graph, which includes other flights besides the Nolin flushed birds, I am forced to wonder, what exactly is “escape flight” in terms of wingbeat rate, as opposed to “normal takeoff”? Is there any difference? For the first 4 or 5 wingbeats the Nolin flights seem to be well within the range of other pileated takeoffs.
The statement that “there is no data to suggest whether Pileated Woodpeckers can maintain a wingbeat frequency approaching 8.6 s-1 for 8 or more wingbeats” ignores the data right in front of him. The Nolin videos constitute such data. Three independent analyses, those of Collinson, Pulliam, and myself, are in general agreement that none of the Nolin birds achieves a cumulative wingbeat rate even close 8.6 Hz over a period of one second as is unequivocally exhibited by the Luneau bird. In fact according to my calculations the highest one-second wingbeat rate amongst the 5 Nolin flights is a mere 7.4 Hz, with a range of 6.7-7.4 Hz. On their web site, the CLO acknowledges that pileateds can produce very high wingbeat rates momentarily after launch. The question is whether the Luneau bird is exhibiting a pattern that can reasonably be expected of a pileated.
The statement that the expected appearance of the upperwing of a pileated, mostly black with a small white patch at the base of the primaries, is obscure and only clearly resolvable as the Nolin birds are flying near vertical, is incorrect. The upperwing pattern is quite apparent in some earlier frames in which the wings are oriented downward. It is also apparent in some frames in which the long axis of the bird is more or less aligned with the camera line of sight, which occurs late in the flights before the birds swoop upward. In fact a thorough comparison of any video with the Luneau video should take into account the fact that the view of the Luneau bird is clearly slightly dorsal as it rises relative to the background. A close examination of comparable portions of the Nolin videos, even interlaced, is in my mind quite revealing.
The statement that the black trailing edge of the pileated is often inconspicuous and may disappear completely is quite correct. The problem is that the reverse is also true. The black trailing edge in every one of the Nolin flights is quite conspicuous in some frames and impossible to overlook, right down to the base of the wing. Numerous frames demonstrate this even with the interlacing. With de-interlaced frames I believe it would be even more apparent. The assertion that the black trailing edge is in general more conspicuous near the wingtip is not borne out by the Nolin videos. In many frames it is the black edge near the base of the wing that is most conspicuous. In others the entire edge is quite noticeable.
The statement that a black trailing edge is discernable on the Luneau bird is not supported by the example given, field 366.7. On that field I see a black wingtip smeared by motion blur. The comparable frame from the Nolin video looks nothing like it and in fact shows a clear indication of a black trailing edge near the base of the wing. If the frame were de-interlaced I suspect that a black trailing edge would be quite apparent on both the near and far wings.
The allusion to the models used in the re-enactment as “flat-winged” is incorrect. There seems to be a widespread misconception that the models simply flapped up and down with no flexion or twisting motion. The models in fact flexed and were designed to reproduce the twisting motion that is typical of bird wings in flight.
The relevance of his statement that the assertion that ivory-bills flap their wings more rapidly than pileateds is “counterintuitive” escapes me, since he seems to agree that the sound in the old Cornell recording is indeed an ivory-bill flapping its wings, and the rate is very high. If we agree that the Luneau bird is a large woodpecker videotaped in the wilds of Arkansas, the question is whether the wingbeat rate is too high to be reasonably attributed to a pileated. If so then we have no reasonable alternative, even if we had zero data on wingbeat rates in ivory-bills.
He states that in all Luneau frames which appear to show white on the dorsum, the bird is distant, and it is difficult to distinguish dorsum from wing. In fact, in Luneau video fields in which white is most apparent on the back of the bird, the wings are below horizontal and cannot possibly contribute to the effect. At least one field shows white all the way across the bird. In the next field the wings are lower and the white on the wings clearly separates from that on the back. Even interlaced, not a single frame in the Nolin videos produces this effect. The argument that the white on the body only appears when the bird is distant and therefore should be suspect is exactly the opposite of what we find to be true in the Nolin videos. As the distance between the birds and the camera increases, there is less motion blur, the birds are facing more directly away from the camera, the view is less ventral, and the wings are more clearly distinguished from the body. The result is that the absence of white on the body of the pileated is more apparent, not less so.
Some final observations:
In many of the side-by-side frame/field comparisons in Dr. Collinson’s paper, I am at a lost to determine what is supposed to be similar. Since he didn’t even bother to reverse the early Nolin frames so that the birds would be in roughly the same orientation, the trailing edges of the wings on the Luneau bird often appear to be in a similar relative position to the leading edges on the Nolin bird. A reader who has not seen the videos might understandably be quite befuddled. So what have we learned from this exercise? Some of the points made are entirely valid in my view. In my opinion, what the Nolin videos do indeed tell us is that with some camcorders and lighting conditions, more white appears on a normal pileated than we would naively expect. Looking at individual frames is clearly fraught with problems. It is true that some frames of the Nolin videos show nothing that would enable us to identify the species. In fact some frames show no black on the wings whatsoever. In some Luneau fields I can barely distinguish if there is even a bird there. If you were to show me those fields without the rest of the video I would be at a loss to identify the bird in any way. Videos are clearly fraught with artifacts that do not necessarily reflect the actual plumage characteristics of the bird. Trying to make measurements of the wings from the Luneau video is a very questionable proposition. I agree with Dr. Collinson that there is enough evidence concerning pileated wing dynamics in these videos to bring into serious question some of the interpretations of Luneau fields by Fitzpatrick et al. They also tell us that flushed pileateds can produce very high wingbeat rates for at least very short periods. The question remains, however, whether any existing video of a normal pileated even comes close to matching critical elements of the Luneau video. My answer is a resounding no. Dave Nolin in fact presented his videos as a contrast to the Luneau video. How good a match can they be? Bill Pulliam has additionally pointed out some very odd features of the wing dynamics of the Luneau bird, features that are in no way matched by any pileated video I am aware of. Again I am forced to wonder, why should this one pileated with a strong indication of white on its dorsum and other unusual features be so distinct in its behavior?
In my opinion every flight in the Nolin videos, when looked at in its entirety, is readily distinguished from the Luneau video and unequivocally shows a pileated. None of the flights exhibits the high wingbeat rate of the Luneau bird over a period of one second. The CLO has already acknowledged that pileated wingbeat rates may range as high as 7.5 Hz. The highest one-second cumulative wingbeat rate among the Nolin flights is only 7.4 Hz, not even close to the 8.6 Hz exhibited by the Luneau bird. I find Dr. Collinson’s comment that “there is no data to suggest” whether pileateds can produce a wingbeat rate comparable to the Luneau bird for 8 wingbeats or more to be quite revealing of his approach. The Nolin flights, now numbering 5, constitute such data in my view. But since none of the flights even approach the required 8.6 Hz for one second, Dr. Collinson considers that we have “no data.” With each new pileated video failing to even come close to matching critical characteristics of the Luneau video, I am wondering how many will be necessary before he considers that we “have data.” Could a pileated achieve such a high wingbeat rate? It’s certainly possible. If we assume that wingbeat rates are normally distributed, we might use the mean one-second rate of the 5 Nolin flights (7.2 Hz) and the standard deviation (0.29 Hz), to estimate the likelihood that a pileated flight will achieve the one-second rate of the Luneau bird (8.6 Hz). According to my calculation we would expect that about one out of 100 flushed pileateds will achieve such a rate. If someone produced a video of a pileated doing this, we would at least be able to answer the question of pileated capabilities in this regard. I would love to see it. The other issues of course would remain. How many pileated videos will be required before we see the odd wing dynamics of the Luneau bird pointed out by Bill Pulliam? How many will be required before we see white on the dorsum? And how many before one incorporates all of the very unusual features together? At some point we must acknowledge that our hypothesis is a bad fit. In my view the pileated hypothesis is a very bad fit for the Luneau bird. Perhaps someday a video will appear that is indeed a decent match. The Nolin flights do not come close in my opinion.
David L. Martin
Competing interests
None declared
Imaging artifacts and flight mechanics in the Luneau video
William M. Pulliam, .
21 March 2007
Dr. Collinson clearly and thoughtfully addresses several of the issues involved in attempts to identify the bird in the Luneau video. However, there are some additional matters of concern that I would like to bring up. First, the interpretation of the underwing pattern of the flying bird is tenuous. One must bear in mind that the black features around the edges of the bird's wings are generally only about one pixel in width. Close inspection of the images from the Luneau video reveals that these black features occur on the leading edges of the wings as well as at the wingtips and trailing edges; neither the Ivory-billed nor the Pileated Woodpecker should display a black leading edge to the underwing. In a number of cases, these black features also appear to align with darker areas in the background behind the moving wings. The black areas near the tips of the birds wings are the only black underwing features that are consistent and repeated between adjacent frames; even these are scarcely more than one pixel in width and strongly motion-blurred. Both species display a black area a the tips of the underwing, and distinguishing between the two wingtip patterns under these conditions seems difficult at best. I feel it is unwise to attempt to infer any detail about the black/white pattern of the underwing from these frames. As Dr. Collinson demonstrates, the black trailing edge of a Pileated Woodpecker's underwing is easily obscured in low-resolution video of a fleeing bird. The combination of motion blur, low resolution, spurious black fringe artifacts on all edges of the wing images, and interactions with the background likely render these images incapable of being used to distinguish between the two species in question.
Dr. Collinson's analysis of the wingbeat frequency of the flying bird addresses the pertinent issues well. There are additional features of the bird's flight mechanics evident in the video, however. In particular, I am struck by the rather sharp contrast between the Luneau bird and known Pileated Woodpeckers in the apparent postures of the wings during the downstroke of each wingbeat. The Luneau bird appears to hold its wings with a strong downward bend at the wrist through most of the downstroke. In contrast, the Pileated Woodpecker images displayed in the paper show wings that are held fully extended during the downstroke, giving a "flat" impression. All other comparable videos of Pileated Woodpeckers in escape flight that I have seen show a similar, "flat-winged" geometry, including the Pileated comparison videos posted by Fitzpatrick et al (reference 7 in the original paper). There is also an impression that the wings of the Luneau bird may not achieve full upward extension at the beginning of each downstroke, in contrast to what is shown in Pileated Woodpecker comparison videos. Of course, in the absence of any comparison films of known Ivory-billed Woodpeckers in flight, it is not possible to determine if these flight mechanics are in fact characteristic of an Ivory-billed Woodpecker. However, they are a prominent and consistent character of the bird in the Luneau video that is sharply inconsistent with the documented flight characteristics of the Pileated Woodpecker.
Competing interests
None declared
JM Collinson response to comments
Jon Martin Collinson, University of Aberdeen
6 May 2008
I welcome David Martin’s and Bill Pulliam’s comments and apologise that it has taken so long to respond to them.
In Fitzpatrick et al. (2005), the Luneau video bird was offered as ‘conclusive proof’ (Fitzpatrick, quoted in Stokstad, 2007) ‘confirming the existence of at least one male’ Ivory-billed Woodpecker. The technical comment and subsequent correspondence by Sibley and colleagues (Sibley et al., 2006; 2007), and to some extent my paper (Collinson 2007) showed formally through reanalysis of the video evidence that there was significant doubt. I concluded in my paper that further research was required.
David Martin suggests that my conclusions and discussion about the data speak volumes about my approach, and I agree – although I think it is the right approach. Given that Pileated Woodpecker is very common and Ivory-billed Woodpecker is very rare or extinct, I took the view that a bird with a wingbeat frequency that is physiologically possible, if possibly unusual, for a Pileated Woodpecker was statistically more likely to be a Pileated Woodpecker than an ivorybill. The relevant research issue in relation to the Luneau video bird is not ‘find a Pileated Woodpecker that flies like this’, but ‘show me why this bird cannot be a Pileated Woodpecker’, and more pertinently, ‘show me why this has to be an Ivory-billed Woodpecker’. My analysis, both in the paper and subsequently has, I think, shown that there is no reason to think that the Luneau bird could not have been a Pileated Woodpecker, a conclusion now widely accepted.
David Martin and Bill Pulliam have argued that the Luneau video shows features not expected of Pileated Woodpecker. So it is important to realise how far the argument has moved. In 2005 it was perhaps justifiable to present the Luneau video as evidence for the persistence of ivorybills, because the plumage patterns shown were strikingly different from our expectations of what a Pileated Woodpecker would look like in flight, and parameters of wingbeat frequency and apparent size of the bird were outside our experience of Pileated Woodpecker. Further research in the papers cited above has chipped away at most of the evidence, showing that either it was unreliable (size estimates) or that the plumage and behavioural features of the Luneau video bird were in fact strikingly similar to Pileated Woodpecker. There are now fewer lines of evidence left from the Luneau video that are still accepted as showing support for identification as an ivorybill, and the current comments defend the remaining suggestive evidence.
The wingbeat frequency analysis presented by Bill Pulliam and David Martin repeats the analysis put forward in this paper showing that Pileated Woodpeckers can flap their wings at frequencies similar to that of the Luneau video bird on initial take-off. The scientific advance demonstrated was rejection of the argument that the Luneau bird may have been an ivorybill because it flapped faster than any Pileated. I had at that time no data to suggest whether they could maintain that frequency for longer periods, although David Martin implies, perhaps correctly, that if many Pileated Woodpeckers were measured and none of them approached the sustained wingbeat frequency of the Luneau video bird, that may be regarded as evidence that the Luneau bird was not a Pileated Woodpecker. However this possibility has now been resolved: comparable data obtained by Louis Bevier of escaping Pileated Woodpeckers is presented at http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html, and shows that Pileated Woodpeckers can maintain 8.4 kHz for 8 wingbeats during the initial escape flight. Plotting the new data on the graph presented by Bill Pulliam at http://static.flickr.com/70/203089268_5f62aa0750.jpg (I have done this at http://www.abdn.ac.uk/ims/staff/details.php?id=m.collinson) makes the argument that the Luneau video bird could not have been an ivorybill untenable. At no time did I intent to give the impression that there was some qualitative difference between the first and second four wingbeats of the Pileated Woodpecker videos.
David Martin asks why I consider it counterintuitive that Ivory-billed Woodpecker should flap its wings in flight at around 8.6 beats per second, when the archival audio (recorded by Arthur Allen and Peter Paul Kellogg in 1935 at an Ivory-billed nest site and presented at http://www.birds.cornell.edu/ivory/evidence/wingbeats.wav) suggests that it does. He also wonders why I question the archival recorded frequency. Basic flight dynamics predicts that on average heavier, longer-winged birds will flap their wings more slowly than smaller birds, and this has been shown to be the case for woodpeckers (Tobalske, 1996). This is presented graphically by Louis Bevier at http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html, and extrapolation of the data suggests that Ivory-billed Woodpecker, in level flight at least, would be expected to flap more slowly that Pileated Woodpecker. This is why I considered the figure of 8.6 Hz cited for Ivory-billed Woodpecker on the basis of the archival recording to be couinterintuitive. In my paper I questioned whether the bird in the archival audio of Ivory-billed Woodpecker was in flight. In fact I should have questioned whether it was an Ivory-billed Woodpecker. The recording was made without accompanying film of the bird making the apparent flight and no accompanying notes by the recordists has been made available that specifically ties this sound to an ivorybill; it should be noted that other species, e.g. Red-bellied Woodpecker Melanerpes carolinus, were present and loudly recorded in the same line the parabolic recording was made. So is the bird in that recording in flight? The wingbeats do not fade away to nothing, as would be expected of a bird in flight, but stop suddenly after 8 cycles. The amplitude trend is downward, but stop suddenly with a final thud which may be suggestive of a bird landing. The bird is probably in flight, and it is convenient to assume it is probably an Ivory-billed Woodpecker, but the data are consistent with a short movement between branches. The flap-rate of the Luneau video bird may well be achievable by Ivory-billed Woodpecker in escape flight. It is now clear, however, that it is definitely achievable by Pileated Woodpecker. It does not, therefore, constitute evidence either way on the identity of the mystery bird, except that identification as Pileated Woodpecker cannot be rejected on this character.
Bill Pulliam argues that the flexion of the wings during the flight of the Luneau video bird is unusual or unprecedented for Pileated Woodpecker and that by default this argues that the bird was an ivorybill. I agree that during most of the flight of the Luneau bird, the wings appear to form a downward arc when passing through the horizontal and fold before vertical downward extension. I believe it represents a rapid flight stroke generating power from the wrist and primaries, and, with the possible exception of the initial take-off, differs from the Pileated Woodpeckers in David Nolin’s video. What I do not understand is why a Pileated Woodpecker would never fly like this. Flexion of the wings during downstroke is widely used by many large species of birds to facilitate wing acceleration and airflow over the wing in the initial stages of flight, as this Pileated Woodpecker appears to be doing here: http://www.flickr.com/photos/birdchick/439102229/. I agree that those who would identify the Luneau video bird as a Pileated Woodpecker need to explain what is going on. I would suggest that this unusual flight pattern may be maintained by Pileated Woodpecker if it were attempting to accelerate powerfully and climb in urgent, rapid flight, as in the Luneau video. While the flight pattern of the Luneau video bird has not been explained in terms of what is known about either species, I feel it will be explicable in terms of escape flight of a panicking Pileated Woodpecker. In any case, there is no way of knowing if it matches Ivory-billed Woodpecker. If the Luneau bird were a Pileated Woodpecker, it would be predicted that it should return to a more usual ‘flappy’ flight in later stages of its escape after gaining height and speed. In fact this would appear to be the case. In later frames of the Luneau video the bird appears to extend its wings fully during the downstroke – the point is illustrated on the Cornell Lab website at http://www.birds.cornell.edu/ivory/evidence/segments/whiteback which, while being used to illustrate a white patch on the mantle, neck or head of the escaping bird, has put a later wingbeat on a loop and demonstrates the distinctive deep flapping flight of a Pileated Woodpecker.
It could be noted that the arguments about the identity of a claimed ‘Bigfoot’ filmed by Patterson and Gimlin in 1967 have been kept going for 40 years on the basis of apparent features of the gait, limb length and flexion that are supposed to be inconsistent with the anthropoid in the film being a man in a suit. Nevertheless it very likely was a man in a suit, and the arguments only have validity if one ignores that fact. Similarly, arguments about the flight style, wing length and flexion of the Luneau video bird only have validity if one discounts the fact that it was very likely to be a Pileated Woodpecker.
It is implied that parts, or all, of my frame-by-frame video comparison of the Luneau and Nolin were invalid because the images were processed differently, in particular that I was unable to deinterlace the digitised version supplied of the original analogue recording. This was acknowledged in my original paper with the caveat that images presented by me from the Nolin video would represent two overlaid frames and that this must be borne in mind. A mature interpretation is required. Bill Pulliam also cites the possibility that edge-effect artefacts induced by video processing make both sets of videos unreliable for determining plumage patterns. Both potential problems are addressed by comparable video of Pileated Woodpeckers in escape flight, not only deinterlaced, but shot on the same type of camera, on the same settings as used by David Luneau, and from the same distance and at the same angle. These important controls have been performed by Louis Bevier and are presented at http://web.mac.com/lrbevier/ivorybilled/Overview.html where it is shown unequivocally that the critical frames used by Fitzpatrick et al. to identify the Luneau video bird as an ivorybill can be produced by Pileated Woodpeckers. Even though the interlaced ‘four-winged’ birds in my paper represented an additional level of complexity in attempts to compare the two videos, it was always obvious to me that there was sufficient similarity between the Pileated Woodpeckers and the Luneau video bird to cast doubt on the identity of the Luneau video bird as an ivorybill. All subsequent work has confirmed that the comparisons had value in showing that the apparent plumage patterns shown by the Luneau video bird could be produced in a low-resolution video of a Pileated Woodpecker. I would welcome an explanation from anyone as to what the substantive differences are between Nolin frame 775 and Luneau frame 366.7, between Nolin frame 457 and Luneau frame 416.7 and between Nolin frame 758 and Luneau frame 283.3 that mean the Luneau bird cannot be a Pileated Woodpecker. Why would an Ivory-billed Woodpecker, with plumage that is very different from Pileated Woodpecker, produce video images that are so very similar to Pileated Woodpecker? I find it much more compelling to accept the null hypothesis that Pileated Woodpeckers produced very similar images in both videos.
I am very receptive to the idea that the American ornithological community gave up on the Ivory-billed Woodpecker too soon, that areas of primary growth in areas such as the Choctawhatchee, Florida, were never fully logged out and that the Ivory-billed Woodpecker could have held on. At the same time, I feel that the failure to relocate this conspicuous and noisy bird in the 2006-7 search season (and at the time of writing, the 2007-8 season also appears to have failed) despite intensive organised and ad hoc search efforts by many ornithologists and birders speaks volumes for its current status, probably more than any sort of argument over the identity of the Luneau video bird.
References:
Collinson, JM: Video analysis of the escape flight of Pileated Woodpecker Dryocopus pileatus: Does the Ivory-billed Woodpecker Campephilus principalis persist in continental North America? BMC Biology 2007 5:8.
Cornell Lab of Ornithology 2006. The search for the Ivory-billed Woodpecker: Video analysis [http://www.birds.cornell.edu/ivory/evidence/segments/whiteback]
Fitzpatrick JW, Lammertink M, Luneau MD Jr, Gallagher TW, Harrison BR, Sparling GM, Rosenberg KV, Rohrbaugh RW, Swarthout ECH, Wrege PH, Swarthout SB, Dantzker MS, Charif RA, Barksdale TR, Remsen JV Jr, Simon SD, Zollner D: Ivory-billed Woodpecker (Campephilus principalis) persists in continental North America. Science 2005, 308:1460-1462 .
Sibley DA, Bevier LR, Patten MA, Elphick CS: Comment on “Ivory-billed Woodpecker (Campephilus principalis) persists in continental North America” Science Online 2006, 311:1555a (dol. 10.1126/science1122778).
Sibley DA, Bevier LR, Patten MA, Elphick CS: Ivory-billed or Pileated Woodpecker? Science 2007 315:1495-1496 DOI: 10.1126/science.315.5818.1495.
Stokstad, E: Gambling on a ghost bird. Science 2007 317:888-892.
Tanner JT: The Ivory-billed Woodpecker, Research Report No. 1. National Audubon Society; 1942.
Tobalske BW: Scaling of muscle composition, wing morphology and intermittent flight behavior in woodpeckers. Auk 1996 113: 151-177.
agreements and disagreements with Dr. Collinson's assertions
21 March 2007
Some months ago, Dr. Collinson kindly sent me a copy of his paper for my comments, after it had been accepted. I told him I felt that it deserved to be published, although I had a lot of problems with it. Although the final paper has apparently been edited to address some of my concerns, many remain, and I feel I should present those here, as well as point out what I feel are some critical flaws in his analysis. I still feel that the paper deserved to be published, but there should have also been provided an opportunity for rebuttal from Fitzpatrick et al. Perhaps that will yet occur.
general remarks:
1) In their analysis of the Luneau video, Fitzpatrick et al. were careful to try to duplicate the conditions as closely as possible, using the same camcorder, focused in the same manner, at the same location, and in similar lighting conditions. It would not be too surprising to find that a different camcorder, focused differently in different lighting conditions would produce very different results even on the same subject. That being said, I think it is possible to make some valid comparisons between the Luneau and Nolin videos, even on plumage characteristics.
2) The Nolin video frames he has displayed are interlaced. The Luneau video fields have been de-interlaced. In some cases the interlaced frames produce very misleading plumage features. It would have been much preferable to use de-interlaced frames. Again, this in itself does not negate all of the analysis, but it makes comparisons in some of the figures quite suspect.
specific remarks.
The statement that Fitzpatrick et al. did not discuss Sibley’s “3 points,” is partially incorrect. They did in fact address 2 out of 3. Point 1 was that the Luneau bird appears to have black secondaries. Fitzpatrick et al. stated that, “The left wing is white to its base as it emerges from behind the tree...and in additional fields that show dorsal views during the latter stages of upstrokes.” Point 2 was that the Luneau bird has “particularly bright white primary bases.” This point was indeed unaddressed by Fitzpatrick et al. and I think is one of Sibley’s strongest.. Point 3 was that the Luneau bird appears to have black curving around the edge of the wing. Fitzpatrick et al. stated that, “The claim by Sibley et al. that in certain fields the black wingtips of the woodpecker show a curved shape suggestive of pileated fails to acknowledge that movement blur is most pronounced at the wingtips and that in other fields of the Luneau video the shape is more suggestive of ivory-billed.”
The statement that aberrant pileateds have been reported from the area is misleading, in that it implies that such birds might explain the Luneau video. The only aberrant pileated yet documented from the Big Woods is the “white” bird photographed in White River NWR. This bird bears no resemblance to the Luneau bird and no other aberrant pileateds whatsoever have been reported from the area, let alone documented.
Critical to Dr. Collinson’s interpretation of the Nolin flights is his suggestion that the first 4 wingbeats of the Nolin birds constitute “escape flight,” while the remainder reflects the birds slowing their flight as they approach the trees. This is quite arbitrary and not supported by anything in the videos. At one second post-launch all of the birds are far from the trees and show no indication of preparing to land. In every case the cumulative wingbeat rate at one second post-launch is quite elevated compared to the typical pileated rate of 5.2 Hz. Bill Pulliam and I have looked at the wingbeat rates without any presumption of when “escape flight” ends. Pulliam’s graph (http://static.flickr.com/70/203089268_5f62aa0750.jpg) and mine (http://home.att.net/~fangsheath/wingbeatrates3.jpg), which includes Nolin flight videos recently posted, both show that while a reduction in wingbeat rates is not perfectly smooth and steady in all individuals, the basic pattern is similar among pileated flights. Fitzpatrick et al. cited a takeoff extracted from the old Singer Tract recordings, and I have independently analyzed the sonagram of this recording. The Luneau bird shows a pattern strikingly similar to that of this recording. In analyzing videos of flushed black-bellied whistling ducks (http://home.att.net/~fangsheath/treeduckrates.jpg), I have found no evidence whatever of a change from “escape flight” to some other kind of flight within the first second post-launch. Some individuals landed a short distance away. Others left the area. Regardless, the basic pattern is similar among flights. While wingbeat rates sometimes exceed 8 Hz momentarily post-launch, among 30 flights the wingbeat rate at one second varies from 5.1 to 7.7 Hz. I find no support in either the pileated or whistling-duck data for this arbitrary 4-wingbeat cutoff. Indeed, looking at Pulliam’s graph, which includes other flights besides the Nolin flushed birds, I am forced to wonder, what exactly is “escape flight” in terms of wingbeat rate, as opposed to “normal takeoff”? Is there any difference? For the first 4 or 5 wingbeats the Nolin flights seem to be well within the range of other pileated takeoffs.
The statement that “there is no data to suggest whether Pileated Woodpeckers can maintain a wingbeat frequency approaching 8.6 s-1 for 8 or more wingbeats” ignores the data right in front of him. The Nolin videos constitute such data. Three independent analyses, those of Collinson, Pulliam, and myself, are in general agreement that none of the Nolin birds achieves a cumulative wingbeat rate even close 8.6 Hz over a period of one second as is unequivocally exhibited by the Luneau bird. In fact according to my calculations the highest one-second wingbeat rate amongst the 5 Nolin flights is a mere 7.4 Hz, with a range of 6.7-7.4 Hz. On their web site, the CLO acknowledges that pileateds can produce very high wingbeat rates momentarily after launch. The question is whether the Luneau bird is exhibiting a pattern that can reasonably be expected of a pileated.
The statement that the expected appearance of the upperwing of a pileated, mostly black with a small white patch at the base of the primaries, is obscure and only clearly resolvable as the Nolin birds are flying near vertical, is incorrect. The upperwing pattern is quite apparent in some earlier frames in which the wings are oriented downward. It is also apparent in some frames in which the long axis of the bird is more or less aligned with the camera line of sight, which occurs late in the flights before the birds swoop upward. In fact a thorough comparison of any video with the Luneau video should take into account the fact that the view of the Luneau bird is clearly slightly dorsal as it rises relative to the background. A close examination of comparable portions of the Nolin videos, even interlaced, is in my mind quite revealing.
The statement that the black trailing edge of the pileated is often inconspicuous and may disappear completely is quite correct. The problem is that the reverse is also true. The black trailing edge in every one of the Nolin flights is quite conspicuous in some frames and impossible to overlook, right down to the base of the wing. Numerous frames demonstrate this even with the interlacing. With de-interlaced frames I believe it would be even more apparent. The assertion that the black trailing edge is in general more conspicuous near the wingtip is not borne out by the Nolin videos. In many frames it is the black edge near the base of the wing that is most conspicuous. In others the entire edge is quite noticeable.
The statement that a black trailing edge is discernable on the Luneau bird is not supported by the example given, field 366.7. On that field I see a black wingtip smeared by motion blur. The comparable frame from the Nolin video looks nothing like it and in fact shows a clear indication of a black trailing edge near the base of the wing. If the frame were de-interlaced I suspect that a black trailing edge would be quite apparent on both the near and far wings.
The allusion to the models used in the re-enactment as “flat-winged” is incorrect. There seems to be a widespread misconception that the models simply flapped up and down with no flexion or twisting motion. The models in fact flexed and were designed to reproduce the twisting motion that is typical of bird wings in flight.
The relevance of his statement that the assertion that ivory-bills flap their wings more rapidly than pileateds is “counterintuitive” escapes me, since he seems to agree that the sound in the old Cornell recording is indeed an ivory-bill flapping its wings, and the rate is very high. If we agree that the Luneau bird is a large woodpecker videotaped in the wilds of Arkansas, the question is whether the wingbeat rate is too high to be reasonably attributed to a pileated. If so then we have no reasonable alternative, even if we had zero data on wingbeat rates in ivory-bills.
He states that in all Luneau frames which appear to show white on the dorsum, the bird is distant, and it is difficult to distinguish dorsum from wing. In fact, in Luneau video fields in which white is most apparent on the back of the bird, the wings are below horizontal and cannot possibly contribute to the effect. At least one field shows white all the way across the bird. In the next field the wings are lower and the white on the wings clearly separates from that on the back. Even interlaced, not a single frame in the Nolin videos produces this effect. The argument that the white on the body only appears when the bird is distant and therefore should be suspect is exactly the opposite of what we find to be true in the Nolin videos. As the distance between the birds and the camera increases, there is less motion blur, the birds are facing more directly away from the camera, the view is less ventral, and the wings are more clearly distinguished from the body. The result is that the absence of white on the body of the pileated is more apparent, not less so.
Some final observations:
In many of the side-by-side frame/field comparisons in Dr. Collinson’s paper, I am at a lost to determine what is supposed to be similar. Since he didn’t even bother to reverse the early Nolin frames so that the birds would be in roughly the same orientation, the trailing edges of the wings on the Luneau bird often appear to be in a similar relative position to the leading edges on the Nolin bird. A reader who has not seen the videos might understandably be quite befuddled. So what have we learned from this exercise? Some of the points made are entirely valid in my view. In my opinion, what the Nolin videos do indeed tell us is that with some camcorders and lighting conditions, more white appears on a normal pileated than we would naively expect. Looking at individual frames is clearly fraught with problems. It is true that some frames of the Nolin videos show nothing that would enable us to identify the species. In fact some frames show no black on the wings whatsoever. In some Luneau fields I can barely distinguish if there is even a bird there. If you were to show me those fields without the rest of the video I would be at a loss to identify the bird in any way. Videos are clearly fraught with artifacts that do not necessarily reflect the actual plumage characteristics of the bird. Trying to make measurements of the wings from the Luneau video is a very questionable proposition. I agree with Dr. Collinson that there is enough evidence concerning pileated wing dynamics in these videos to bring into serious question some of the interpretations of Luneau fields by Fitzpatrick et al. They also tell us that flushed pileateds can produce very high wingbeat rates for at least very short periods. The question remains, however, whether any existing video of a normal pileated even comes close to matching critical elements of the Luneau video. My answer is a resounding no. Dave Nolin in fact presented his videos as a contrast to the Luneau video. How good a match can they be? Bill Pulliam has additionally pointed out some very odd features of the wing dynamics of the Luneau bird, features that are in no way matched by any pileated video I am aware of. Again I am forced to wonder, why should this one pileated with a strong indication of white on its dorsum and other unusual features be so distinct in its behavior?
In my opinion every flight in the Nolin videos, when looked at in its entirety, is readily distinguished from the Luneau video and unequivocally shows a pileated. None of the flights exhibits the high wingbeat rate of the Luneau bird over a period of one second. The CLO has already acknowledged that pileated wingbeat rates may range as high as 7.5 Hz. The highest one-second cumulative wingbeat rate among the Nolin flights is only 7.4 Hz, not even close to the 8.6 Hz exhibited by the Luneau bird. I find Dr. Collinson’s comment that “there is no data to suggest” whether pileateds can produce a wingbeat rate comparable to the Luneau bird for 8 wingbeats or more to be quite revealing of his approach. The Nolin flights, now numbering 5, constitute such data in my view. But since none of the flights even approach the required 8.6 Hz for one second, Dr. Collinson considers that we have “no data.” With each new pileated video failing to even come close to matching critical characteristics of the Luneau video, I am wondering how many will be necessary before he considers that we “have data.” Could a pileated achieve such a high wingbeat rate? It’s certainly possible. If we assume that wingbeat rates are normally distributed, we might use the mean one-second rate of the 5 Nolin flights (7.2 Hz) and the standard deviation (0.29 Hz), to estimate the likelihood that a pileated flight will achieve the one-second rate of the Luneau bird (8.6 Hz). According to my calculation we would expect that about one out of 100 flushed pileateds will achieve such a rate. If someone produced a video of a pileated doing this, we would at least be able to answer the question of pileated capabilities in this regard. I would love to see it. The other issues of course would remain. How many pileated videos will be required before we see the odd wing dynamics of the Luneau bird pointed out by Bill Pulliam? How many will be required before we see white on the dorsum? And how many before one incorporates all of the very unusual features together? At some point we must acknowledge that our hypothesis is a bad fit. In my view the pileated hypothesis is a very bad fit for the Luneau bird. Perhaps someday a video will appear that is indeed a decent match. The Nolin flights do not come close in my opinion.
David L. Martin
Competing interests
None declared
Imaging artifacts and flight mechanics in the Luneau video
21 March 2007
Dr. Collinson clearly and thoughtfully addresses several of the issues involved in attempts to identify the bird in the Luneau video. However, there are some additional matters of concern that I would like to bring up. First, the interpretation of the underwing pattern of the flying bird is tenuous. One must bear in mind that the black features around the edges of the bird's wings are generally only about one pixel in width. Close inspection of the images from the Luneau video reveals that these black features occur on the leading edges of the wings as well as at the wingtips and trailing edges; neither the Ivory-billed nor the Pileated Woodpecker should display a black leading edge to the underwing. In a number of cases, these black features also appear to align with darker areas in the background behind the moving wings. The black areas near the tips of the birds wings are the only black underwing features that are consistent and repeated between adjacent frames; even these are scarcely more than one pixel in width and strongly motion-blurred. Both species display a black area a the tips of the underwing, and distinguishing between the two wingtip patterns under these conditions seems difficult at best. I feel it is unwise to attempt to infer any detail about the black/white pattern of the underwing from these frames. As Dr. Collinson demonstrates, the black trailing edge of a Pileated Woodpecker's underwing is easily obscured in low-resolution video of a fleeing bird. The combination of motion blur, low resolution, spurious black fringe artifacts on all edges of the wing images, and interactions with the background likely render these images incapable of being used to distinguish between the two species in question.
Dr. Collinson's analysis of the wingbeat frequency of the flying bird addresses the pertinent issues well. There are additional features of the bird's flight mechanics evident in the video, however. In particular, I am struck by the rather sharp contrast between the Luneau bird and known Pileated Woodpeckers in the apparent postures of the wings during the downstroke of each wingbeat. The Luneau bird appears to hold its wings with a strong downward bend at the wrist through most of the downstroke. In contrast, the Pileated Woodpecker images displayed in the paper show wings that are held fully extended during the downstroke, giving a "flat" impression. All other comparable videos of Pileated Woodpeckers in escape flight that I have seen show a similar, "flat-winged" geometry, including the Pileated comparison videos posted by Fitzpatrick et al (reference 7 in the original paper). There is also an impression that the wings of the Luneau bird may not achieve full upward extension at the beginning of each downstroke, in contrast to what is shown in Pileated Woodpecker comparison videos. Of course, in the absence of any comparison films of known Ivory-billed Woodpeckers in flight, it is not possible to determine if these flight mechanics are in fact characteristic of an Ivory-billed Woodpecker. However, they are a prominent and consistent character of the bird in the Luneau video that is sharply inconsistent with the documented flight characteristics of the Pileated Woodpecker.
Competing interests
None declared
JM Collinson response to comments
6 May 2008
I welcome David Martin’s and Bill Pulliam’s comments and apologise that it has taken so long to respond to them.
In Fitzpatrick et al. (2005), the Luneau video bird was offered as ‘conclusive proof’ (Fitzpatrick, quoted in Stokstad, 2007) ‘confirming the existence of at least one male’ Ivory-billed Woodpecker. The technical comment and subsequent correspondence by Sibley and colleagues (Sibley et al., 2006; 2007), and to some extent my paper (Collinson 2007) showed formally through reanalysis of the video evidence that there was significant doubt. I concluded in my paper that further research was required.
David Martin suggests that my conclusions and discussion about the data speak volumes about my approach, and I agree – although I think it is the right approach. Given that Pileated Woodpecker is very common and Ivory-billed Woodpecker is very rare or extinct, I took the view that a bird with a wingbeat frequency that is physiologically possible, if possibly unusual, for a Pileated Woodpecker was statistically more likely to be a Pileated Woodpecker than an ivorybill. The relevant research issue in relation to the Luneau video bird is not ‘find a Pileated Woodpecker that flies like this’, but ‘show me why this bird cannot be a Pileated Woodpecker’, and more pertinently, ‘show me why this has to be an Ivory-billed Woodpecker’. My analysis, both in the paper and subsequently has, I think, shown that there is no reason to think that the Luneau bird could not have been a Pileated Woodpecker, a conclusion now widely accepted.
David Martin and Bill Pulliam have argued that the Luneau video shows features not expected of Pileated Woodpecker. So it is important to realise how far the argument has moved. In 2005 it was perhaps justifiable to present the Luneau video as evidence for the persistence of ivorybills, because the plumage patterns shown were strikingly different from our expectations of what a Pileated Woodpecker would look like in flight, and parameters of wingbeat frequency and apparent size of the bird were outside our experience of Pileated Woodpecker. Further research in the papers cited above has chipped away at most of the evidence, showing that either it was unreliable (size estimates) or that the plumage and behavioural features of the Luneau video bird were in fact strikingly similar to Pileated Woodpecker. There are now fewer lines of evidence left from the Luneau video that are still accepted as showing support for identification as an ivorybill, and the current comments defend the remaining suggestive evidence.
The wingbeat frequency analysis presented by Bill Pulliam and David Martin repeats the analysis put forward in this paper showing that Pileated Woodpeckers can flap their wings at frequencies similar to that of the Luneau video bird on initial take-off. The scientific advance demonstrated was rejection of the argument that the Luneau bird may have been an ivorybill because it flapped faster than any Pileated. I had at that time no data to suggest whether they could maintain that frequency for longer periods, although David Martin implies, perhaps correctly, that if many Pileated Woodpeckers were measured and none of them approached the sustained wingbeat frequency of the Luneau video bird, that may be regarded as evidence that the Luneau bird was not a Pileated Woodpecker. However this possibility has now been resolved: comparable data obtained by Louis Bevier of escaping Pileated Woodpeckers is presented at http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html, and shows that Pileated Woodpeckers can maintain 8.4 kHz for 8 wingbeats during the initial escape flight. Plotting the new data on the graph presented by Bill Pulliam at http://static.flickr.com/70/203089268_5f62aa0750.jpg (I have done this at http://www.abdn.ac.uk/ims/staff/details.php?id=m.collinson) makes the argument that the Luneau video bird could not have been an ivorybill untenable. At no time did I intent to give the impression that there was some qualitative difference between the first and second four wingbeats of the Pileated Woodpecker videos.
David Martin asks why I consider it counterintuitive that Ivory-billed Woodpecker should flap its wings in flight at around 8.6 beats per second, when the archival audio (recorded by Arthur Allen and Peter Paul Kellogg in 1935 at an Ivory-billed nest site and presented at http://www.birds.cornell.edu/ivory/evidence/wingbeats.wav) suggests that it does. He also wonders why I question the archival recorded frequency. Basic flight dynamics predicts that on average heavier, longer-winged birds will flap their wings more slowly than smaller birds, and this has been shown to be the case for woodpeckers (Tobalske, 1996). This is presented graphically by Louis Bevier at http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html, and extrapolation of the data suggests that Ivory-billed Woodpecker, in level flight at least, would be expected to flap more slowly that Pileated Woodpecker. This is why I considered the figure of 8.6 Hz cited for Ivory-billed Woodpecker on the basis of the archival recording to be couinterintuitive. In my paper I questioned whether the bird in the archival audio of Ivory-billed Woodpecker was in flight. In fact I should have questioned whether it was an Ivory-billed Woodpecker. The recording was made without accompanying film of the bird making the apparent flight and no accompanying notes by the recordists has been made available that specifically ties this sound to an ivorybill; it should be noted that other species, e.g. Red-bellied Woodpecker Melanerpes carolinus, were present and loudly recorded in the same line the parabolic recording was made. So is the bird in that recording in flight? The wingbeats do not fade away to nothing, as would be expected of a bird in flight, but stop suddenly after 8 cycles. The amplitude trend is downward, but stop suddenly with a final thud which may be suggestive of a bird landing. The bird is probably in flight, and it is convenient to assume it is probably an Ivory-billed Woodpecker, but the data are consistent with a short movement between branches. The flap-rate of the Luneau video bird may well be achievable by Ivory-billed Woodpecker in escape flight. It is now clear, however, that it is definitely achievable by Pileated Woodpecker. It does not, therefore, constitute evidence either way on the identity of the mystery bird, except that identification as Pileated Woodpecker cannot be rejected on this character.
Bill Pulliam argues that the flexion of the wings during the flight of the Luneau video bird is unusual or unprecedented for Pileated Woodpecker and that by default this argues that the bird was an ivorybill. I agree that during most of the flight of the Luneau bird, the wings appear to form a downward arc when passing through the horizontal and fold before vertical downward extension. I believe it represents a rapid flight stroke generating power from the wrist and primaries, and, with the possible exception of the initial take-off, differs from the Pileated Woodpeckers in David Nolin’s video. What I do not understand is why a Pileated Woodpecker would never fly like this. Flexion of the wings during downstroke is widely used by many large species of birds to facilitate wing acceleration and airflow over the wing in the initial stages of flight, as this Pileated Woodpecker appears to be doing here: http://www.flickr.com/photos/birdchick/439102229/. I agree that those who would identify the Luneau video bird as a Pileated Woodpecker need to explain what is going on. I would suggest that this unusual flight pattern may be maintained by Pileated Woodpecker if it were attempting to accelerate powerfully and climb in urgent, rapid flight, as in the Luneau video. While the flight pattern of the Luneau video bird has not been explained in terms of what is known about either species, I feel it will be explicable in terms of escape flight of a panicking Pileated Woodpecker. In any case, there is no way of knowing if it matches Ivory-billed Woodpecker. If the Luneau bird were a Pileated Woodpecker, it would be predicted that it should return to a more usual ‘flappy’ flight in later stages of its escape after gaining height and speed. In fact this would appear to be the case. In later frames of the Luneau video the bird appears to extend its wings fully during the downstroke – the point is illustrated on the Cornell Lab website at http://www.birds.cornell.edu/ivory/evidence/segments/whiteback which, while being used to illustrate a white patch on the mantle, neck or head of the escaping bird, has put a later wingbeat on a loop and demonstrates the distinctive deep flapping flight of a Pileated Woodpecker.
It could be noted that the arguments about the identity of a claimed ‘Bigfoot’ filmed by Patterson and Gimlin in 1967 have been kept going for 40 years on the basis of apparent features of the gait, limb length and flexion that are supposed to be inconsistent with the anthropoid in the film being a man in a suit. Nevertheless it very likely was a man in a suit, and the arguments only have validity if one ignores that fact. Similarly, arguments about the flight style, wing length and flexion of the Luneau video bird only have validity if one discounts the fact that it was very likely to be a Pileated Woodpecker.
It is implied that parts, or all, of my frame-by-frame video comparison of the Luneau and Nolin were invalid because the images were processed differently, in particular that I was unable to deinterlace the digitised version supplied of the original analogue recording. This was acknowledged in my original paper with the caveat that images presented by me from the Nolin video would represent two overlaid frames and that this must be borne in mind. A mature interpretation is required. Bill Pulliam also cites the possibility that edge-effect artefacts induced by video processing make both sets of videos unreliable for determining plumage patterns. Both potential problems are addressed by comparable video of Pileated Woodpeckers in escape flight, not only deinterlaced, but shot on the same type of camera, on the same settings as used by David Luneau, and from the same distance and at the same angle. These important controls have been performed by Louis Bevier and are presented at http://web.mac.com/lrbevier/ivorybilled/Overview.html where it is shown unequivocally that the critical frames used by Fitzpatrick et al. to identify the Luneau video bird as an ivorybill can be produced by Pileated Woodpeckers. Even though the interlaced ‘four-winged’ birds in my paper represented an additional level of complexity in attempts to compare the two videos, it was always obvious to me that there was sufficient similarity between the Pileated Woodpeckers and the Luneau video bird to cast doubt on the identity of the Luneau video bird as an ivorybill. All subsequent work has confirmed that the comparisons had value in showing that the apparent plumage patterns shown by the Luneau video bird could be produced in a low-resolution video of a Pileated Woodpecker. I would welcome an explanation from anyone as to what the substantive differences are between Nolin frame 775 and Luneau frame 366.7, between Nolin frame 457 and Luneau frame 416.7 and between Nolin frame 758 and Luneau frame 283.3 that mean the Luneau bird cannot be a Pileated Woodpecker. Why would an Ivory-billed Woodpecker, with plumage that is very different from Pileated Woodpecker, produce video images that are so very similar to Pileated Woodpecker? I find it much more compelling to accept the null hypothesis that Pileated Woodpeckers produced very similar images in both videos.
I am very receptive to the idea that the American ornithological community gave up on the Ivory-billed Woodpecker too soon, that areas of primary growth in areas such as the Choctawhatchee, Florida, were never fully logged out and that the Ivory-billed Woodpecker could have held on. At the same time, I feel that the failure to relocate this conspicuous and noisy bird in the 2006-7 search season (and at the time of writing, the 2007-8 season also appears to have failed) despite intensive organised and ad hoc search efforts by many ornithologists and birders speaks volumes for its current status, probably more than any sort of argument over the identity of the Luneau video bird.
References:
Collinson, JM: Video analysis of the escape flight of Pileated Woodpecker Dryocopus pileatus: Does the Ivory-billed Woodpecker Campephilus principalis persist in continental North America? BMC Biology 2007 5:8.
Cornell Lab of Ornithology 2006. The search for the Ivory-billed Woodpecker: Video analysis [http://www.birds.cornell.edu/ivory/evidence/segments/whiteback]
Fitzpatrick JW, Lammertink M, Luneau MD Jr, Gallagher TW, Harrison BR, Sparling GM, Rosenberg KV, Rohrbaugh RW, Swarthout ECH, Wrege PH, Swarthout SB, Dantzker MS, Charif RA, Barksdale TR, Remsen JV Jr, Simon SD, Zollner D: Ivory-billed Woodpecker (Campephilus principalis) persists in continental North America. Science 2005, 308:1460-1462 .
Sibley DA, Bevier LR, Patten MA, Elphick CS: Comment on “Ivory-billed Woodpecker (Campephilus principalis) persists in continental North America” Science Online 2006, 311:1555a (dol. 10.1126/science1122778).
Sibley DA, Bevier LR, Patten MA, Elphick CS: Ivory-billed or Pileated Woodpecker? Science 2007 315:1495-1496 DOI: 10.1126/science.315.5818.1495.
Stokstad, E: Gambling on a ghost bird. Science 2007 317:888-892.
Tanner JT: The Ivory-billed Woodpecker, Research Report No. 1. National Audubon Society; 1942.
Tobalske BW: Scaling of muscle composition, wing morphology and intermittent flight behavior in woodpeckers. Auk 1996 113: 151-177.
Competing interests
None