Thursday, March 15, 2007

Martin Collinson's Ivorybill article

Recently published here.

In short: Same fatal mistakes as made by Sibley et al.

-- He interprets artifacts as plumage. As I have pointed out several times here, the Luneau bird shows black fringes on the leading edge of the underwing as often as it shows black fringes on the trailing edge of the underwing. Ergo, apparent black fringes must not be taken to indicate anything about the bird's actual coloration. There is nothing resolved in the underwing pattern but a lot of white with a black area at the tip, which both species share.

ADDENDUM: I have just frame-by-framed through the Luneau video one more time. In almost every single case, every black fringe seen on the birds wings can be seen to align with a darker area in the background. When this happens the black fringes often occur on both edges of the wings, and in subsequent frames if they reappear they do so fixed relative to the background, not relative to the moving wings. The only exceptions to this are the small black wingtip area, which appears most consistently at full downward extension. The supposed broad curving black area in frame 366.7 that is said to represent the outer portion of the black trailing edge of a PIWO is exactly aligned with the background shadow, and it disappears on the next frame when the wing has moved past this shadow. The black arcs a the ends of the wing in frame 300 are (on the right wing) aligned with a background shadow and (on the left wing) clearly a streak left by a rapidly moving wingtip. There is no consistent occurence of a black trailing edge or curved black wingtip; these things appear for one frame at a time and then disappear in the next frame because they are not real. Collinson has stated on his blog that "Sibley et al. have shown why the Luneau bird wasn't an Ivory-bill." This is only true if you accept video artifacts as reality.

-- He totally misses the boat on flight mechanics (as has virtually everyone else) by focusing on wingbeat rate. There is a much bigger story, in that the shape of the Luneau bird's wing strokes are markedly different from that shown by the Nolin video or any other video of a Pileated I have seen. The Luneau bird does not fully extend its wings at the top of the upstroke, and holds them in a strong downwardly-bowed arc on every downstroke. ALL pileated videos I have seen show wings that are fully extended on completion of the upstroke and held essentially straight out downstroke on every wingbeat other than the first one or two after launch (while the bird is still rolling to attain proper flight attitude). This is very evident in Collinson's side-by-side frame comparisons, and Sibley's sketches showed these dynamics as well. Yet neither of these authors have made any mention of it nor offered any explanation. Is this flight Ivorybill-ish? No one can say, we have no comparison material. But it is definitely not at all Pileated-like. The wingbeat rate issue is minor in comparison.

My longer discussions of these isues are here and here, for starters.

8 Comments:

At 5:54 AM, Anonymous MM in NY said...

Thanks for pointing this out. I did think the Collinson article was very well-written, thoughtful, and reasonable in tone, and I said so on Birdforum. After I posted, I looked at the videos one after the other (as Ilya suggested), and my impression was that the motion of the wings in the Luneau video was distinctly different from the motion of the wings in the Nolin videos. The Luneau bird has more of what I'd call a rowing quality (which is perhaps another way of saying what you've described in more detail in your post). To be honest, I'm not sure if I've seen that description elsewhere, but it's the best term I can think of.

In addition, I still think your original point about the skills needed to analyze this kind of video is very well-taken, and it would be great if some experts in forensic video analysis tackled the Luneau footage, Mike Collins's and the Choctawhatchee material (if it's ever made public). Birders and/or ornithologists may not be the most qualified people to examine this material.

 
At 9:45 AM, Anonymous IBWO atheist said...

Why not expand the discussion of woodpecker flap rates to include the extant Campephilus species? These, like all Campephilus, are noisy, conspicuous and unwary, and therefore easy to locate, study at length, and photograph, so it can't be too hard to gather relevant data. We need to know more about variability both within and between species and there is no reason to limit the discussion to Pileated vs. nonexistent IBWO.

In my opinion the Collison article contributes little new to the discussion, as the Luneau video has already been sufficiently analyzed and debunked, and his casual endorsement of the sight records as plausible demonstrates a lack of (online) scholarship and critical thinking.

Regarding artifacts in the video, if apparent field marks of Pileated are dismissed artifacts, then apparent artifacts of IBWO should be suspect too. Instead, believers interpret the lack of IBWO field marks, e.g., the underwing pattern, as the artifact. Ad hoc dismissal and fabrication of evidence should not be tolerated in science and Science.

 
At 10:07 AM, Blogger Bill Pulliam said...

Atheist --

I agree with much of what you say. I have searched online for posted videos of other Campephilus species but have not yet found any that show the birds in flight. I'd expect Ivorybill fight to be more simlar to that of other large Campephilus that to the more distantly related Dryocopus. I'd be very curious to see any useful videos that are out there.

Obviously I disagree with the assertion that the Luneau video has been thoroughly debunked, no need to repeat all that again. However, I have maintained that no diagnostic information can be gotten from the underwing pattern of the bird in the main flight sequence, a position that seems to be unpopular with both the Cornell et al. and Sibley et al. groups. I agree, if you say that the black medial bar would be obscured, you should not also say that the black trailing edge should be visible. Nor should you dismiss a suspect pixel here yet embrace another possibly suspect pixel there.

 
At 11:03 AM, Anonymous Mike Collins said...

The basic gyst of the article is: (a) the fleeing bird in the Luneau video has rapid wingbeats, (b) fleeing pileateds can also have rapid wingbeats; (c) the bird in the Luneau video is probably a pileated. The author conveniently ignores the fact that the bird in my video, which is in level flight and is not fleeing (it was perched 128 m distant just before flying) also has rapid wingbeats (about 7.5 Hz). If you're going to publish a paper with a title containing the phrase, "Does the Ivory-billed Woodpecker Campephilus principalis persist in continental North America?" and wingbeat frequency is the basis for that paper, then all relevant data should be considered. As a regular poster on birdforum, the author was clearly aware of this data.

 
At 7:48 AM, Anonymous Sidewinder said...

I will be disappointed with Cornell if they fail to include additional data on PIWO wingbeats in the rebuttal to Collinson that they are apparently preparing; surely they have accumulated a number of new flight sequences by now. Adding data from other Campephilus, if available, would be an excellent idea.

 
At 8:17 AM, Blogger Bill Pulliam said...

We are well past the point in time that some of this data should have been compiled, published, and examined, including more extensive Pileated flight info and comparisons with known Campephilus species, and especially including examination of flight mechanics more estensively than just wingbeat rate.

 
At 5:39 PM, Anonymous ibwo_agnostic said...

Should we really be this hung up on the video at this point? Shouldn't we have fantastic sightings and photos and videos? It's a BIRD, it is not a phantom. Come on!

 
At 5:50 PM, Blogger Bill Pulliam said...

Agnostic --

My answers to all your questions are found in the series of posts that immediately preceed this one. I won't repeat it all here.

 

Post a Comment

<< Home

Site Meter