Tuesday, November 08, 2011

-- Further Imperial Analysis --

------------------------------------------------------------------

Given the time it was taking Bill Pulliam to comment on Cornell's recently-released Imperial Woodpecker footage I presumed he was doing his own in-depth analysis of the clips. And indeed he was. I think we all appreciate Bill's consistent attempts to objectively analyze such pieces of data (…and all the moreso given that, amazingly, he does so on a dial-up internet connection!). His analysis here:

http://bbill.blogspot.com/2011/11/imperial-film.html

To cut to the chase (but I definitely recommend you read Bill's entire post), he concludes that the new Imperial film unhesitatingly RE-AFFIRMS the bird in the original Luneau video as an Ivory-billed Woodpecker. I won't go into the reasoning he marshals toward that conclusion (focusing in part on wing-beat data and 'bowing' motion), nor nitpick over details of his analysis, nor am I personally convinced that this particular film evidence carries as much weight as he (or others) may place on it, but his take is always interesting. Of course, for most readers here, Bill is speaking to the choir anyway.

Until Louis Bevier, or Martin Collinson, or Richard Prum, or Sibley, or some other major skeptic come forth with their own analysis of the amazing film from Mexico the IBWO debate likely won't move any steps forward, and I'm not sure skeptics will even find it worth their time to closely study this new footage of a lone bird -- I don't blame them if they pass -- from their standpoint it's probably a further squandering of their time.

I'll say again what I've said previously… even without the Luneau video (and regardless of what the bird is in that clip) there is more than enough evidence to believe that Ivory-bills persisted into the 21st century, and may still be with us today. Convincing the masses of that though at this point will require more than an analysis of flight patterns or any other tangential evidence; it will require conclusive, indisputable film, or a fresh carcass... the Imperial film can generate a lot of words, most of them falling on the deaf ears of people who have moved on (and it can't explain why no such equivalent film of an IBWO has been attainable in the last 6 years, nor a single active nest-hole found of a bird that has to be actively breeding to still be with us).


On a completely separate note, for anyone who might be interested, once again an original copy of James Tanner's IBWO manuscript is being auctioned on eBay (auction over December 1st), with an asking price of $500+. Merry Christmas!!:

http://tinyurl.com/7czjqa4
------------------------------------------------------------------

42 comments:

Bill Pulliam said...

CY -- thanks for linking.

I think you often underrate the importance of this video from the perspective of the larger community. The video is what made it "different this time." Without the video, it is just another swarm of brief, mostly single-observer, in flight sightings, and weird noises in the woods. Without the video, there is no press conference, no Science paper, no global media frenzy, no recovery plan, no multi-year, multi-state followup searches, no reallocation of funding. It becomes simply another tantalizing and mysterious ornithological footnote, like the Pearl River expeditions. For better or worse, this whole phenomenon has always fundamentally rested on those 4 seconds of blurry bird. Everything else is viewed as subjective and unconfirmable.

You also expressed concern about overgeneralizing from one individual; but of course we have no other choice. Fortunately, flight behavior is hard-wired, not learned, and in general different individuals of a species exhibit very similar flight behavior to each other, barring injury or deformity. This is of course one of the pillars of birding, the ability to reliably identify an individual bird by how it moves as much as by how it looks. Without this, pelagic birding would be an order of magnitude more difficult! It's true there is not as much Imperial footage as would be ideal, but the bird does show a variety of behavior, and flies three times with different details each time. So she does demonstrate some of the range of behaviors within her own one individual self. This is much the same as the fact that you could learn whole lot about human biomechanics, if you had never seen a human before, by a single short film of a random healthy person walking down the street, then sprinting briefly to catch a bus. Sure different humans walk differently, but none of them are going to move like a chimpanzee (which is far more closely related to us than the Ivorybill is to the Pileated).

Mark said...

Actually, it's the first time Bill has come out and said unequivocally that the bird in the Luneau video is an IBWO. His analysis is first rate and compelling. It demolishes the central argument made by Sibley et al. who at his point have an obligation to either rebut or retract, imo.

I've also got to say that your suggestion that there was anything approaching "unseemly" behavior on Cornell's part is pretty scurrilous and purely speculative. There's no basis for the conspiracy theories that are circulating, and comments like yours in the initial post only serve to feed these unfounded claims.

Bill Pulliam said...

I would not hold my breath waiting for any comment from the mainstream skeptics (Sibley, Jackson, etc.). They have never addressed many of the rebuttals to their work; for instance they have not had anything to say about the multiple assertions that all their "Pileated field marks" are actually digital image compression artifacts, nor have they ever addressed the criticism that their interpretation of the launch requires either a Pileated with a white trailing edge to its underwing or with a magical ability to bend light so that the white underwing coverts can "bleed" around the obscuring tree trunk (white bleed happens in the camera, not in the world). There have been arguments circulating for years that undermine their assertions, and they have ignored them. I doubt that will change. Addressing their critics would only serve to bring more attention to the criticisms.

Mark said...

Needless to say, I don't expect a comment, but as a non-scientist, I'm disappointed (if no longer disillusioned) by the failure to engage, admit mistakes, or respond to what appears to be a very compelling counter-argument.

cyberthrush said...

I'll play devil's advocate a bit here and guess the view"skeptics" have:

I'd guess Jackson will likely have a serious look-see at the Imperial footage, but in the end find it inconclusive (as far as relating it to the Luneau video) and therefore not worth prolonged discussion.

The other major critics feel the IBWO debate has already drawn too much attention away from more pressing conservation issues, and won't want to be dragged into unresolvable debate over yet another piece of brief tape (though of course they'll view it with interest).
I think Louis Bevier has worked directly with the Cornell folks on occasions and might be the most likely, of anyone, to say something about the film (though I doubt he'd agree with Bill's analysis). I'm guessing any response from skeptics will be via comments to The Auk.

This is just speculation on my part, and these folks can speak up for themselves if necessary (and of course I hope they do!).

Mark said...

Just to be clear: they obviously haven't had time to respond to this particular and particularly compelling counter-argument, but history suggests they won't. I hope they'll prove me wrong.

Bill Pulliam said...

I think from a conservation perspective many people feel (with justification) that if you can't find the bird there is not much you can do for it even if it does still exist, other than broad-scale ecosystem-level habitat stuff, which is being done anyway. So the extinction question is kind of moot from that point of view.

They are there or they are not, regardless of what any of us believe.

David Sibley said...

I find the just-released Imperial Woodpecker movie very sad, especially as it comes with all the baggage of the recent Ivory-billed furor, but I will comment on Bill Pulliam's new analysis. First, Bill is apparently unaware of Louis Bevier's website http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html which has a detailed discussion of wingbeat rates showing that Pileateds can flap very quickly on takeoff, more quickly than the Imperial, and can keep it up just as long.

The fundamental flaw in Bill's approach is that he can't rule out the possibility of Pileated. Instead he selectively emphasizes a few very specific and subjective similarities between the Imperial footage and the Arkansas video. Flapping rates are variable in all birds. All woodpeckers are capable of launching in multiple ways.

The Arkansas bird could be launching the way Fitzpatrick et al proposed, showing white on the folded wing, but it could also be launching the way we proposed, showing white under the wing. There's nothing about the Arkansas bird's behavior that is outside the norm for PIleated (or for Ivory-billed, apparently). We suggested a bunch of plumage features that point towards Pileated, but Bill has simply dismissed all of those as artifacts. I disagree, but OK, if they're all artifacts then all of the white and black patterns shown in the video are unreliable. In that case there is nothing diagnostic in the plumage, and nothing diagnostic in the behavior, and the Arkansas video is unidentifiable.

Mark said...

At the risk of speaking for Bill, it's pretty clear he does reference Bevier's website with regard to flap rate:

Its wingbeat frequency was higher than that shown in any of the Pileated Woodpecker comparison material available; in spite of claims of unpublished videos showing otherwise, there has still not been a video of a Pileated Woodpecker presented to the public in which the bird achieves and sustains such a rapid wingbeat rate.

Bevier makes assertions about flap-rate, but as far as I'm aware his claims are mere assertions, unsupported by any available hard data. In fact, the available hard data suggest he was mistaken.

David Sibley said...

Oh, I see. But by selectively accepting only the evidence that supports your conclusion, you're falling into the same trap that got us into this mess in the first place.

Mark said...

How is saying that there's no hard data, only one person's entirely unverifiable assertions, (unless he's got video that he's prepared to release for independent analysis) "selectively accepting" evidence?

Bill Pulliam said...

My comment appears to not be going through; on the off chance that it is because it is too long I will repost it in pieces. If multiple copies appear I will delete the extras (so don't read anything nefarious into any "deletedby author" notes you might see).

Part 1:


I am extremely familiar with the Bevier website, having posted my own response to it shortly after it was put up. This response is easily found by clicking my Luneau Video link from my sidebar in my blog. The problem I have with his site is that there is more to this that just the flap rate. One of the most prominent features is that even when flying fast, all the Pileated videos I have seen show the birds inserting brief closed-winged pauses between wingbeats within the first 3-7 flaps after launch, which the Luneau bird does not do even at the 60 fps deinterlaced rate (nor does the Imperial demonstrate this). Without actually seeing any of Bevier's source material I cannot judge anything more about it other than the one (selective) datum he chooses to present. As I have maintained all along, flap rate is only one part of the apparent anomalies in the flight of that bird that I have been unable to reconcile with the Pileated. And my initial approach to this video was in fact an attempt to reconcile it with a Pileated, because I didn't want to waste time, effort, and emotions on the consequences of a mistaken identification any more than anyone else did.

Bill Pulliam said...

(apparently it was a size issue)

Part 2:


Of course the bird could be launching as you propose. Until a couple of weeks ago I strongly dissented with Cornell on their hypothesized launch sequence, and have stated this loudly and clearly many times in my writings. My own examinations of videos of Pileateds and Magellanics launching supported your model. However, as I have pointed out before, your model for launch requires a bird that has a white tailing edge to its underwing. In the last frame before the wing vanishes, the remaining trace is white. This cannot be explained as white bleed, because this phenomenon occurs in the camera, not in the world. If the white is not in line of sight to the camera, it cannot bleed. Yet the last vanishing trace of the underwing is still white. So, no, that launch sequence does not reconcile with a Pileated either way it is sliced. And yes, in fact, I can rule out a Pileated.

Bill Pulliam said...

Part 3 (final):


I must bristle a bit at the assertion that I am just selectively choosing the data I like to support my own ideas. I have written extensively about this video, I have gone through every frame of it and discussed them one-by-one, I have discussed and answered just about everything that has been written about it online or in the peer-reviewed literature. I did not simply "dismiss" those supposed plumage features as artifacts, I went through in great detail and demonstrated why they clearly are artifacts. I analyzed them and foudn them to be artifacts, I did not just "dismiss" them. All of this analysis is readily available on my blog, just a couple of clearly marked clicks from the opening page. If you have not gone through it in detail, please do so before you "dismiss" my conclusions as "subjective impressions."

My comments on the Imperial built on this extensive background I had arleady presented about the Luneau video, taking features I had already identified and discussed at length and comparing them to the behaviors shown by the Imperial. I am in no way just cherry picking a frame here, a subjective impression there, to fit an a priori conclusion.

David Sibley said...

Frankly Bill, I view your entire discussion of video processing artifacts as questionable, because (as your friend Mark said above) "there's no hard data, only one person's entirely unverifiable assertions". I don't have anything against unverifiable assertions, but you should acknowledge yours and not be so hard on others. And we are at an impasse.

Also, you should reread Don Crockett's comment about "white bleed" (on your website) relative to your comment part 2 here.

But I think I'm preaching to the wrong choir, so I'm going back to my real work now.

Bill Pulliam said...

Well there's nothing particularly "unverifiable" about the phenomenon of edge artifacts in digital imaging -- it is universal, ubiquitous, predictable, and extremely well known, and should be taken in to account by anyone who attempts to infer information from this or any other highly magnified digital video or still image. Show that sequence to an experienced digital photographer who knows nothing about birds (I have done this) and they will likely agree that the black blobs look and behave like artifacts.

Re: White bleed, I agreed with Don Crockett at the time and still do. But this misses the point that on that last frame in question (f50) it would appear that there should not be any white in that particular frame in the first place to bleed, according to your own interpretive sketches where only a small sliver of wing remains unobscured by the tree. A judgement call, surely.

As for preaching to the choir, there are generally many more people reading a blog than commenting on it, and I expect our little dialogue has been seen by a far more diverse and open-minded audience than you might suspect. My own blog gets read by many people who politely suspect that I am cracked in the head about my Ivorybill opinions but continue reading for the rest of the content; however they almost never comment.

And finally, my mantra: The bird is either extinct or it is not, our own individual beliefs don't change this.

cyberthrush said...

For now, I'll just sincerely thank David for taking time out of his busy schedule to address some of the issues here. I'm sure he's not really comfortable revisiting this whole debate given the sometimes unproductive nature of exchanges… most of us are simply at unresolvable impasses on many matters.

And I do hope some of the other serious skeptics will find a venue to comment on technical aspects of Cornell's latest paper (or Bill's blog post for that matter) -- "The Auk" is only a quarterly publication, and it would be well to see some further commentary on the Imperial footage (and Cornell's conclusions) without waiting that long.

Mark said...

It merits noting that there has been a significant, if not explicitly acknowledged, change in David Sibley's stance:

Sibley et al. in 2006: With our new understanding of the bird's movements, all observed features are consistent with a typical pileated woodpecker and some are inconsistent with a normal ivory-billed woodpecker. We conclude that one cannot reject the hypothesis that the bird is a normal pileated woodpecker (i.e., the null expectation); moreover, the evidence firmly supports this hypothesis.

Sibley now: The Arkansas bird could be launching the way Fitzpatrick et al proposed, showing white on the folded wing, but it could also be launching the way we proposed, showing white under the wing. There's nothing about the Arkansas bird's behavior that is outside the norm for PIleated (or for Ivory-billed, apparently). We suggested a bunch of plumage features that point towards Pileated, but Bill has simply dismissed all of those as artifacts. I disagree, but OK, if they're all artifacts then all of the white and black patterns shown in the video are unreliable. In that case there is nothing diagnostic in the plumage, and nothing diagnostic in the behavior, and the Arkansas video is unidentifiable.

I don't think his other sometimes flippant comments come close to calling Bill Pulliam's detailed analysis into question. He certainly has not addressed the issues of flight mechanics, and Bill's argument that multiple aspects thereof are in fact inconsistent with a PIWO. Beyond that no one has produced any video to support Bevier's claims about PIWO flap rates (claims which, as has already been noted, are at odds with the numerous published videos of PIWOs in escape flight), and even if one accepts Bevier's interpretation of Tobalske, the documented flap rate of the IMWO demolishes his argument about predicted rates.

For the record: I've never met Bill Pulliam. We have corresponded from time to time over the past few years. We have agreed about some things and disagreed, sometimes vehemently, about others. I like and respect him, but we are not friends, not even in the loose internet sense. The word seems to have been used to insinuate that personal loyalty had something to do with my pointing out that he was in fact referencing (as opposed to being unaware of) Bevier, a distinction that is not at all inconsequential. If the word was being used to insinuate something, it's a really cheap shot.

David Sibley said...

Bill and Mark,

I don't see any significant change between my published statement from 5 years ago and my quick comments yesterday, I have always said the bird was consistent with a Pileated Woodpecker. That doesn't necessarily mean it has to be a Pileated, and I can agree that other explanations coexist, but the possibility of Pileated has not been excluded.

Stepping back from the minutiae, my main point is that everything seen in the Arkansas video can be plausibly explained by a Pileated Woodpecker. Given that, it doesn't matter if ALL of it also fits Ivory-billed. The only way to identify that bird as an Ivory-billed is to find something that makes it IMPOSSIBLE for it to be a Pileated, and Bill's analysis and all of the Cornell responses have still failed to do that.

The very limited published info on flap rates shows that Pileated can be extremely close to the Arkansas bird. Louis Bevier has videotape, and has posted sonograms and analysis, extending the known range of Pileated flap rates to include the Arkansas bird. Choosing to dismiss that evidence as mere "assertions"… sorry, but that appears biased. Furthermore, given what's known about flap rates, I don't think Louis's results should be a surprise to anyone.

I mentioned Don Crockett's comment because he emphasizes the issue of motion blur, combined with white bleed. A black and white wing, moving quickly, will blur to mostly white, and a black band, moving quickly, will blur into the background and disappear. I think this explains your apparent white trailing edge of the wing disappearing behind the tree.

You don't have to agree that this is likely, but I think you have to admit that it is a POSSIBLE explanation. Even if you argue that Ivory-billed is the more likely explanation, as long as Pileated remains a possibility the video cannot be held up as evidence of Ivory-billed.

Bill Pulliam said...

As I have said before, my flight dynamics issues extend beyond flap rate and I have not asserted that this is a solid piece of evidence on its own. It could have excluded an Ivorybill, and if the Imperial had shown a consistently lower flap rate than the Pileated I think this would have been strong pro-Pileated evidence; however this did not happen. What I have looked for, and not found, are two other features: the apparent bowed wings (which could be considered subjective and dependent on interpretation), and the complete lack of closed-wing pauses between wingbeats in the Luneau bird through at least 10 or 11 cycles (not evident at any point in the video, but there are later segments where the bird is obscured or where it has become too small to tell). I have not seen these features in any Pileated video, mine or others, and I cannot tell if Bevier's birds displayed them or not as the videos are not available for independent examination. I can say that the manner in which the flap rate in his fastest Pileated drops (8.8. Hz cycles 0-5, 7.4 Hz cycles 6-9, 6 Hz cycles 10-12) would strongly suggest to me that this bird also showed this distinctive, discrete, clearly identifiable (from a video) feature of its flight. But I cannot tell this without seeing the video.

Again, in the material I have had access to, this is a distinctive, universal feature of Pileated flight appearing within 3-7 flaps from launch, not shown by the Luneau bird, not shown by the Imperial. But I feel like I have already made this point many times here and elsewhere, and it has been repeatedly answered with only a reiteration of the flap rate data.

As for impossibility, this is not the standard for any scientific claim. Impossibility can never be proven.

cyberthrush said...

well, I'll chime in here so David doesn't feel totally ganged up on, because I have problems with Bill's analysis too:

I'm quite sure I've seen PIWOs in full flap mode for extended periods; while the flap-bound style is "typical" it's simply NOT their only mode of travel -- moreover the Luneau bird is flying through woodland probably maneuvering and evading tree trunks along the way -- this is not cruising flight, it is maneuvering flight, likely requiring constant flapping or adjustment of wings.
Again, I have trouble with all the "flight dynamics" evidence because it is based on such tiny sample sizes, and not under identical conditions -- it is simply not possible to say with certitude what the overlap is, if any,between IBWO and PIWO; for me, it remains speculative.

Having said all that, from a sheer "gizz" or impression standpoint, I've never comprehended how the skeptics match the frame-by-frame white patterns of Luneau to a normal PIWO (I just don't get it, though I don't get some of Cornell's analysis either) -- and I would concede the Luneau bird could be a leucistic PIWO, which for me has never been adequately ruled out.

Both the wing-bowing Bill refers to (I've sometimes called it "oaring motion") and overall proportions of the Luneau bird are what have most made me favor IBWO as best fit -- but I'll concede these may be subjective impressions that appear one way on video (which may be distorted), and might appear very differently seeing the bird in the field.
So while I personally think the Luneau bird most likely IBWO, I understand the skeptics' view that the "null hypothesis" of PIWO is not ruled out (though Bill thinks it can now be -- and I will try to spend more time with Bill's videos this weekend to see if he can convince me, but I'm doubtful). And I think that is David's bottom-line point: that it doesn't even matter how well the Luneau bird 'fits' an IBWO; it only matters can the null hyp. be discarded -- that's the ultimate impasse here -- Bill may think it can be bridged scientifically; I believe there remain too many imprecise and subjective elements.

Mark said...

David: thank you for taking the time to write a thoughtful response. Bill has covered most of what I would say in rebuttal and more effectively than I could hope to have done. I have only a little to add.

First, I think it is a significant shift to go from asserting, almost categorically, that the bird is a normal PIWO to saying that this is a "possible explanation."

Second, regarding Louis Bevier's data, and in addition to Bill's point that flap rate is only one consideration among several, his assumptions about flap rate and mass have been shown to be mistaken based on the actual IMWO footage. These assumptions are central to several aspects of his argument. In addition, it seems to me that releasing the videos so others can analyze them would be appropriate and transparent and should have been done a long time ago. Given all the attacks on Cornell for failing to be transparent, it's more than a little ironic that Bevier has provided a sonogram and his interpretation of the data but has not released the footage itself.

Bill Pulliam said...

Cy -- these very brief (shorter than 1/10th second) closed-wing pauses between flaps are generally not obvious to the naked eye. They manifest in real-time as a slightly slower apparent wingbeat. This is different from the slow bounding flight where the wings are closed for extended periods with isolated flaps in ones and twos between the bounds. The transition between "pausing" flight and "bounding" flight is smooth, with the "pauses" growing longer until they become noticeable to the naked eye. The transition between "steady" flight and "pausing" flight is not so obvious without a film or video that can be watched in slow motion. In EVERY Pileated example I have seen it has happened within <<10 wingbeats after launch.

cyberthrush said...

Thanks for the clarification Bill; as I say I'll try to find time this weekend for a slower, more studious perusal of your videos and post, than I've had the chance to do thus far.

FAV said...

Those calling to see these Sibley/Bevier PIWO clips have evidently forgotten that they cleverly refused to present PIWO data out to 27 cycles. No such tapes of 27 cycles exist that match the AR IBWO. Neither the deniers or likely anyone can have PIWO tapes showing > 8.3 Hz 27 cycles post launch since a PIWO probably reaches the optimal speed, considering flight mechanics, after 8 cycles.

They avoided the pertinent comparison by erroneously claiming that the AR video's late cycles are unclear so they only examined ~ 8 cycles of PIWO and claimed victory. They unilaterally ignored the latest cycles even though other reviewers routinely gathered 27 wing beat cycle data from the subject video.


See here how this misleading presentation was worded in order to indirectly support the wrong conclusion in the Science Note:

>> (5) Although some claim that the bird in the Luneau video sustains a wingbeat rate of 8.6 Hz for over four seconds, the measurable duration is barely 2 seconds in that video. Beyond that point, the image is so blurred that one cannot accurately measure wingbeats. Moreover, my own analysis indicates that the bird in the Luneau video begins slowing within this first 2 seconds (measuring angle of wing in successive wingbeats).<<<<



>>>In one case, I measured 8.8 Hz for the first 5 cycles, 8.2 Hz over the first 9 cycles combined, and 7.5 Hz over 12 cycles with a duration of about 3 seconds. That is essentially the same as my own measurements for the bird in the Luneau video because that bird clearly slows after the first 5 cycles and, according to my measurements, flies at the same frequency as Pileated Woodpeckers that I recorded. My samples of wingbeats closely match, beat for beat, those of the Luneau bird.<<<<



tks Fred V

Mark said...

Point taken Fred V., all the more reason that the complete, unedited videos should be made available for others to examine and analyze.

David Sibley said...

Just to clarify, it is still my very strong feeling that the bird in the Arkansas video is a normal Pileated Woodpecker. Most of that comes from the fact that nothing in the video excludes PIleated. If there is more than one possible explanation, the more likely explanation is... most likely. In this case it's a common species versus a species that hasn't been confirmed in 60 years. If images of a randomly-selected large woodpecker from Arkansas are so blurry and small that they fit both Pileated and Ivory-billed, then it's a PIleated beyond any reasonable doubt. In the same way a 1-pixel speck in a photo could be anything, so there's no basis to claim it's an Ivory-billed Woodpecker.

On top of that, I don't think video artifacts would show up around the wings in a consistent pattern through multiple wingbeat cycles, against changing backgrounds, so I believe most or all of the dark marks are real plumage patterns. That is a more direct line of evidence that helps me to be confident that it's not an Ivory-billed, but mostly it's the lack of an answer to the basic question "Why isn't it a Pileated?"

cyberthrush said...

I suspect David may have opened the floor to another flurry of arguments of why it "CAN'T" be a PIWO! -- so if you wish to state (re-state, re-re-state??) those arguments I'll just ask everyone to keep it civil, please.

Mark said...

@cyberthrush As I see it, Bill Pulliam, who did not make a firm statement about the identification (although he pointed to a number of elements that he deemed inconsistent with PIWO) until the release of the IMWO video, has made a compelling case that it can't be a PIWO. I still don't think that David (or anyone else) has said anything to refute the central points, continuing instead to focus on the dispute about what is an artifact and on the flap rate for the first eight cycles, which is (to repeat) only one element among several. . .and, as others have pointed out, is something of a distraction.

There are new elements to Bill's analysis, and I was hoping to read a critique of those new elements, not just a reiteration of arguments that have been made in the past. And while David may stand by his conclusion that the bird in the Luneau video is a normal PIWO, I still maintain that he's changed his stance substantially from consistent with PIWO/inconsistent with IBWO to consistent with either one and therefore a PIWO because IBWOs have not been documented in 60 years.

This is important because if we think of the Sibley et al. argument as a stool or a chair, even from the perspective of the lead author, one of the legs has been cut off. As I see it, Bill's analysis, knocks the the whole thing down. I'm sure David, Louis Bevier and others do not agree, but I'd love it if they'd show their work.

David Sibley said...

Well, you can believe what you want, but I haven't changed my stance at all. And technically, if we're using metaphors, the PIleated argument would be a floor, not a stool. It is the default and your challenge is to build a stool that will support the Ivory-billed argument.

And I'm curious, has anyone ever asked Louis if they could analyze his videos? That would be the polite way to do it, right?

cyberthrush said...

David, maybe I misunderstood, but you first seemed to be arguing that our challenge was to refute the possibility of the Luneau bird being a PIWO, because until that was done, any evidence of it being an IBWO was rather inconsequential (from your standpoint). But now you seem to be stressing the need for us to put forth a case for it being IBWO, because otherwise it merely becomes a PIWO by default (rather than by evidence). I think you're trying to have it 2 ways...

Moreover, at least 2 separate lines of evidence have already been extensively used (over and over) to build the IBWO "stool" you ask for: 1) flight dynamics and 2) the black/white frame-by-frame pattern. Obviously, you disagree with those lines of evidence (or at least think them open to counter-interpretations) but it's not fair to imply they haven't been mightily put forth, and many people stand by them. (if someone wishes to try and succinctly state them again here, fine with me)
I truly am sorry that these impasses seem so unresolvable.

FAV said...

I'll answer succinctly on only wing beat Hz; others can quote on some of the important points they have studied/noted.

I do however object to the inferred standard that the video was the only evidence of scientific value in AR. It wasn't. The acoustical evidence was strong as it is in FL and LA.

Also object to anyone casually calling any species extinct. Which is what the authors themselves relate when they are careless in getting the heuristic rate of large Campephilus completely wrong (see Beviers site). Extinction declarations based on rushed work is inexcusable.

A sad but correct, valid and timely extinction determination requires a lack of any credible evidence of presence.

There is plenty of credible evidence of presence not the least being written and LENGTHY sightings by competent observers and counter-calling DKs and kents.

The AR RBC (approved the sighting notes, sketches, acoustical evidence, etc) and the USFWS (consisting of several scientists) rejected the Science Note as unconvincing and flawed. The BODY of evidence proved an IBWO was present in AR per unbiased review.

tks

FAV said...

AR video isn't a PIWO because behavioral, efficiency and flight dynamic constraints limit the flap rate during level, escape flight of PIWO's soon after take off. Between ~ 8 and 12 cycles post take off, PIWO flap rate begins to drop off. Note that the AR bird has leveled off in late cycles which is not apparent due to the tilted camera and canoe (personal conv. Luneau).

Supporting evidence: No video of a large NA woodpecker video exists of a bird doing ~ 8.3 Hz, 27 cycles post takeoff except the subject video and other putative IBWO videos. I have observed and checked hundreds of field PIWOS and videos and have observed/found none like either the AR or LA IBWO videos. All skeptics and Bevier have been asked to produce the videos years ago.

In addition the % bounding is higher in all (or possibly all) similar PIWO flights.

The historical descriptions support the qualitative if not quantitative fact that IBWOs have a noticeably faster flap rate than PIWO in escape flight.

The field tested models by Nudds, Pennycuick and Tobalske (references below) show that PI and IB should have significantly different flap rates.

The recent Imperial W. tapes also show that the largest congeneric within the Campephilus has a high flap rate plotted against only body weight. This contradicts heuristic assertions by the Science Note authors via website of the IBWO's flap rate. The authors had the completely wrong slope direction for the plot line and a Hz of ~ 2.5 for IBWO and an even lower inferred Hz for IMWO. The authors errors in not including wing surface area or aspect ratio was pointed out years ago (see Virrazzi et al. posts) and no correction was made.

Pennycuick CJ (1996) Wingbeat frequency of birds in steady cruising flight: New data
and improved predictions. J Exp Biol 199: 1613–1618.

Nudds RL, Taylor GK, Thomas ALR (2004) Tuning of Strouhal number for high propulsive efficiency accurately predicts how wingbeat frequency and stroke amplitude
relate and scale with size and flight speed in birds. Proc Roy Soc Lond B 271:
2071–2076.

Pennycuick CJ (1990) Predicting wingbeat frequency and wavelength of birds. J Exp
Biol 150: 171–185.

Tobalske BW (1996) Scaling of muscle composition, wing morphology, and
intermittent flight behavior in woodpeckers. Auk 113: 151–177.

tks Fred Virrazzi

Mark said...

I may get into this in more depth later, but for the present, I'm sorry, but David, with all due respect, your statements and stance in 2006 are significantly different from what you're saying now. Just look at your own words. It's disingenuous to claim otherwise.

Regarding Louis's footage, just to echo Fred V., there have been many requests for him to release his video, and even if that were not the case, his posting stuff on the web and claiming it's somehow dispositive without providing the raw data is not at all in keeping with the basic tenets of science and the and the review process, something that's obvious, even to a layman like me. Until this week, I hadn't looked at Louis's page in a few years and may have something more to say about the substance in future (I supsect there may be some very significant issues that Louis has missed but am not sure yet.) For now, the failure to make the raw data available is a stark contrast to Cornell insofar as transparency is concerned. So too is his rather misleading use of averages in discussing flap rate, something that can easily confuse the mathematically challenged because it obscures the rapid decrease in flap rate after the first 5-8 wingbeats. In any case, no one should have had to ask Louis to release the footage. . .Ever.

cyberthrush said...

Since there are some newer readers to the blog lately who may not be familiar with the skeptical views being referenced, I've added direct links to 3 of the major skeptical offerings (Sibley et.al., Bevier, and Collinson) to the left-hand column, under "Some Skeptics' Takes" to better acquaint yourselves (they all pertain chiefly to the Luneau video).

Bill Pulliam said...

David Sibley -- I won't rehash here (refer to my blog, posts easily found from sidebar link), but just summarie that I found the things I identified as artifacts were not consistent between wingbeat cycles or adjacent frames and were highly dependent on the changing background. I realie my original expositions of this are lengthy and can be tedious, but it is spelled out in detail. None of that is new information (something like 4 years old), but it has still not been addressed in detail or specifics by those who originally proposed that these features were real and diagnostic.

And absolutely yes, losing those marks only makes the bird indeterminate, not an Ivorybill. And of course indeterminate defaults to Pileeated. But (again won't rehash) other anomalies that appear to exclude Pileated have been identified, and not adequately rebutted (or really even addressed) by critics, who seem to have decided the matter was settled years ago and have not seemed to feel the need for any further discussion, even when new ideas or information are presented.

Re: Asking Louis about making his videos available, of course, multiple people, multiple times.

Buck said...

Without the video, there is no press conference, no Science paper, no global media frenzy, no recovery plan, no multi-year, multi-state followup searches, no reallocation of funding. It becomes simply another tantalizing and mysterious ornithological footnote, like the Pearl River expeditions. For better or worse, this whole phenomenon has always fundamentally rested on those 4 seconds of blurry bird. Everything else is viewed as subjective and unconfirmable...(Bill)

In this case it's a common species versus a species that hasn't been confirmed in 60 years. If images of a randomly-selected large woodpecker from Arkansas are so blurry and small that they fit both Pileated and Ivory-billed, then it's a PIleated beyond any reasonable doubt...(David)

Figure S5A is likely a branch stub (J. Fitzpatrick pers. comm., 29 July 2005), rather than a perched Ivory-billed Woodpecker as suggested by Fitzpatrick et al. (2005b)... (Jackson)

Having said that though, it is again disconcerting that a single amateur on the back of a mule 55 years ago was able to attain film, at quite some distance and obviously with a 1950's camera, of an Imperial Woodpecker that clearly shows the white 'saddle' back, while 6 years of more recent effort with far better equipment, by far more individuals, with more leads, searching far more locales, for the Ivory-bill, has failed to produce such, either by a person or an automatic camera...(Cyber)

Most of the story is in the above quotes.

Mark said...

This will have to be a three-part post. I've decided to go ahead and post some further observations about Louis Bevier's material on wingbeats, which can be found here:

http://web.mac.com/lrbevier/ivorybilled/Wingbeat.html.

I'm not a scientist, and I don't consider myself at all skilled at reading sonograms. Nor do I have any specific expertise in avian anatomy or flight mechanics. Everything I know, I've learned on the fly, so to speak, so some of what I'm suggesting below may be completely wrong, but I think most of it is significant.

Prior to the start of this particular discussion, I hadn't looked at Louis Bevier's material in a couple of years. Of course, the IMWO video proves that Bevier's interpretation of Tobalske was mistaken. Although this was, no doubt, an error made in good faith, significant problems with the interpretation have been pointed out by others in the past, and it's pretty clear that a more painstaking and in depth analysis of the literature would have led to a different projection, or at the very least a far less confident assertion about the expected rate. In any event, a retraction of this aspect of his analysis is clearly in order.

There are other issues that have not, to the best of my recollection, been pointed out in a public forum before. For the most part, they pertain to the validity of certain comparisons, inferences, and unsupported assumptions.

Mark said...

Part 2:

1) Bevier compares the flap rate of an IBWO recorded in the Singer Tract with the flap rate of a flushed Pileated Woodpecker recorded in Maine. Although he implicitly acknowledges that this comparison may be inapt, the acknowledgement tends to obfuscate what I think is the most significant difference. The PIWO is in escape mode, In contrast, Bevier hypothesizes (rightly, I think) that the IBWO was recorded in flight between branches. However agitated the Singer Tract bird may have been, it was not fleeing. Indeed, it was making a brief flight and apparently landed after 8 flaps. Therefore, it seems logical to conclude that the bird was flapping at a rate below what would be expected from a flushed bird.

2) Bevier suggests that the differences between his PIWO sonogram and the sonogram for the Singer Tract birds are a function of the microphones that were used (parabolic in the Singer Tract, omnidirectional for the PIWO.) This is speculation on his part, and based on the literature, which describes IBWOs as having very loud wingbeats, it seems safe to assume that there would be significant differences between the sonograms, something that at appears to be the case, at least to this layman. The differences are not only in amplitude. I'd invite people more skilled than I to look closely at these differences and would be interested in their interpretations.

3) Although the video has not been made available, Bevier's own sonogram may support what Bill Pulliam has said about wing tucking. The final beat in the cycle presented in the sonogram has a noticeably lower amplitude than the earlier ones, and the duration between the last two flaps appears to have increased. So I'm left to wonder whether this is the point at which the PIWO does its first wing tuck of the kind Bill has described.

4) The use of hand-released birds for comparison strikes me as being inapt. I would think that this only demonstrates what the uppermost limits of the initial flap rate would be, roughly the equivalent of a bird escaping after being temporarily caught by a predator, something that is of little or no relevance to the conditions under which the Luneau video was obtained. It seems to me that this distinction is of added significance when it comes to a woodpecker, since the released bird would not only be highly stressed but would also be unable to use its legs and tail to assist in the launch. It therefore seems that such a bird would be flapping quite frantically, especially at first. This is purely intuitive, not based on any data, but it makes sense. In addition, since Bevier only provides averages at 5,9, and 12 cycles (and has not produced the video), there's no way to ascertain how and when the flap rate drops off and whether, for example, the first three flaps were far more rapid than the subsequent ones.

In a personal communication, David Martin, who has analyzed the Luneau video, Louis Bevier's page, and David Nolin's many clips of Pileated Woodpeckers in escape mode, had the following to say:

Mark said...

Part 3:

"When the Luneau bird becomes visible from behind the tree, it has already virtually completed a wingbeat. Since the wingbeat rate for the Luneau bird is decreasing over time, and would be expected to do so for most any bird, it is important to choose the starting point carefully when making comparisons. In my analyses I have generally used the middle of the first downstroke as a starting point. Of couse this is not visible in the Luneau video, but it is easy to estimate within a frame or so. With this beginning, it is possible to measure the wingbeat rate of the bird fairly accurately out to 1.17 seconds. At 0.92 seconds the value obtained is 9.0 Hz, as 1.17 seconds 8.8 Hz. Similar values on the Nolin pileateds are 6.5-7.3 Hz. From the Singer Tract sonagram I estimate the rate at 0.92 seconds to be 8.7 Hz.

Since Bevier uses a very different method for estimating rates, a comparison is somewhat difficult. However, estimating a rate from his sonagram gives a value of 8.3 Hz at 0.92 seconds. His hand-released bird manages 8.2 Hz at 1.1 seconds. If this represents an extreme value for a stressed bird desperate to gain altitude, it is not encouraging as any kind of match for the Luneau bird."

When I wrote to Martin, I inquired about the Singer Tract recording, voicing my suspicion that the flap rate should be lower than it would be for a bird in escape flight. I had no prior knowledge of his analysis, but according to him: "The Singer Tract recording wingbeat rate is somewhat below that of the Luneau bird, with a value of 8.7 hz at 0.92 seconds, compared to 9.0 hz for the Luneau bird at this point." So the actual data tend to support my speculation in this regard.

Martin's unpublished paper includes a statistical analysis of multiple recordings of PIWOs in escape flight (in contrast to Bevier's purely anecdotal presentation of data.) According to Martin, the flap rate of the Luneau bird is 9.5 standard deviations above the average flap rate of a PIWO at 1.17 seconds.

In response to a follow-up query, Martin elaborated:

"The standard deviation referred to in my paper is based solely on the Nolin flights, and does not include Bevier's data. In fairness, his first pileated sonogram indicates rates far above those of any of the Nolin flights, although still well below those of the Luneau bird. "

In my view, this makes it even more essential for Bevier to release the footage for independent analysis. Martin continued:

"In my studies of flushed black-bellied whistling ducks, I found that without exception their wingbeat rates dropped off in the first second of flight. The mean rates (sample size = 30) are as follows:

10 frames (0.17 seconds) - 7.30 Hz

30 frames (0.50 seconds) - 6.81 Hz

60 frames (1.00 seconds) - 6.43 Hz

It is worth noting that the variation among individuals also declines somewhat, with a standard deviation of 0.68 Hz at 35 frames and 0.63 Hz at 60 frames. This is similar to the pattern seen in the Nolin videos; variation in wingbeat rate declines as the flights proceed. This might seem surprising given that the pileateds are starting to exhibit folded-wing pauses as the flights proceed, but keep in mind that the rate is cumulative. Yet, as Bill Pulliam long ago pointed out, the Luneau bird seems to be DIVERGING from the overall trend of the Nolin videos."

To put it simply, Louis Bevier's material on flap rates is deeply flawed in multiple ways. The evidence that the bird in the Luneau video is inconsistent with a PIWO, normal or otherwise is abundant and very strong. Now that we have actual footage of the IBWO's nearest congener that supports the IBWO hypothesis, the case should be closed.

Bill Pulliam said...

I've summarized my thoughts on the wingbeat comparisons here:

http://bbill.blogspot.com/2011/11/woodpecker-wingbeats-revisited.html

Comments on my stuff can always be left on my own blog.

Steve said...

This is real car crash stuff. You have to view it from between your fingers!