Evidence collected by Project Coyote in two parts of Louisiana from 2009-2018 should, on its own, suffice to justify maintaining the Ivory-billed Woodpecker’s current listing as an extant, endangered species in the United States. Others may differ, but I think the totality of the evidence does even more, making a compelling case that the species persists in more than one Louisiana location.
This will be a two part post. The initial impetus for writing it was a conversation I had with Erik Hendrickson about how the scientific, birding, and legal approaches to evidence seem to differ. That will be the main focus of Part 2. As for Part 1, after talking to Erik, I realized that I’d never done a single post aggregating the evidence we’ve obtained over the years, so when it was announced that the U.S. Fish and Wildlife Service would be conducting a status review for the Ivory-billed Woodpecker, the need for writing this up became more urgent. That’s the primary focus of Part 1, but some of what I’m planning for Part 2 will be foreshadowed in this lengthy treatment.
There is a strong tendency to privilege visual evidence over other forms – audio and circumstantial. I’m inclined to think that our audio evidence is the most compelling material overall; as regular readers know, I also think circumstantial evidence, feeding sign in particular, is very important. But the structure of this post will track conventional attitudes, starting with visual material followed by a discussion of sound recordings and auditory encounters, and concluding with more circumstantial forms of evidence.
Project Coyote has been focused on two areas in Central and East-central Louisiana. I found what I call the “old” search area after reading a newspaper account of a local landowner’s efforts to get his reports taken seriously. This is the landowner who sketched corrections to the Louisiana Hunting Guide’s ivorybill illustrations and who showed Frank Wiley (my late collaborator) bark scaling when he visited the area.
To be clear about what he was trying to show in the corrections: the gentleman died before I had a chance to meet him, but as Frank told me after his August 2009 interview, all of the corrections are in red ink, so the intent was not to put a male crest on a female bird; it was to show the female’s crest as more erect and recurved than the drawing and indicate that the male’s crest also is more erect and prominent (less Pileated-like) than the game guide representation. Also note the drawings at top right, in which he compared IBWO and PIWO wings, implying that the wings in the game guide image are a little too rounded.
What’s most salient about these observations is not whether they are perfectly accurate, although they do seem to reflect some little-known nuances, it’s that the landowner had enough claimed observations of male and female ivorybills to recognize subtle differences and to distinguish them from Pileateds. In addition, he was offended by the treatment he received when he tried to alert the authorities. I have personal knowledge about this last aspect based on conversations with family members who wanted to see him vindicated.
He was certainly convincing enough for Frank to start visiting the property on a weekly basis during the summer of 2009. Thus, Project Coyote was born.
The landowner’s reports go back to the 1990s. (The linked post includes audio from 2009). There are several medium to large state WMAs and National Wildlife Refuges in fairly close proximity to the site. We heard additional claims from residents of the area while we were focused there.
Putative ivorybill activity on the property seems to have diminished or ended altogether after an adjoining parcel was logged in late fall 2010. As a result, we gradually shifted our attention to the new search area.
The old search area received minimal attention during the post-Arkansas period; some peripheral Ivory-billed Woodpecker-related research was conducted in the general vicinity, but there was no formal, funded search effort. Nor was it visited by Cornell’s Mobile Search Team, although Martjan Lammertink did spend several days on site, after the logging.
Similarly, our “new” search area was entirely off the radar. It doesn’t show up on anyone’s list of promising locations, notwithstanding the fact that it is in one of the most sparsely populated, heavily forested parts of the state. There’s also a long history of local reports. Most of the claims seem to have been fairly recent, but some go back to the late 1990s.
And these are only the claims I know about . . .
For reasons that I hope will become clear, I have not kept track of all our possible sightings since 2009 and have not always mentioned them in blog posts. Nevertheless, I’ll begin with sightings. Over the years, Frank had more possibles and reported having better views than anyone else. I’m probably second, with approximately six since 2009, but unfortunately no good looks. Steve Pagans has had several possibles; a few visitors have also made claims.
While this is largely a subject for the next post, it’s well-known that eyewitness testimony is unreliable, at least when the source is an untrained observer. Nevertheless, eyewitness testimony is central to many a criminal trial, and it has a strong impact on jurors. There seems to be a parallel with respect to the ivorybill. Accounts of possible sightings tend to attract more interest than many more substantive and important posts.
Thus my Sunset Sighting post, which is not quite two months old, has had approximately 800 pages views, but my March 18, 2017 post entitled, Numerous Kent-like Calls Recorded on March 11 and 15, 2017 received a total of 744 views last year. Similarly, the post on Joseph Saucier’s October 2017 sighting and our follow-up visit to that location, Change of Pace Change of Place, posted in late November, received 804 page views, the second highest total for the year, (well behind the announcement of Frank’s passing).
According to a survey conducted in the aftermath of the Arkansas controversy, “[R]espondents were most confident in the sightings, less confident in the Luneau video and recordings of double-knocks and kent calls, and least confident in the FishCrow video.” The discrepancy was strongest for the “Definitely of IBWO” category, where 22 people listed sightings as conclusive, compared to 10 for the Luneau video, 9 for kent calls, 8 for double knocks, and 4 for the first Fishcrow video. (PDF) The Great Ivory-billed Woodpecker Debate: Perceptions of the Evidence. Available from: https://www.researchgate.net/publication/259337980_The_Great_Ivory-billed_Woodpecker_Debate_Perceptions_of_the_Evidence[accessed Jun 28 2018].
So sightings have an almost irresistible appeal; in some cases, there’s good reason to credit them, but it is very difficult to separate the wheat from the chaff, to determine what’s trustworthy and what isn’t. And there are many kinds of chaff – wishful thinking, self-delusion, ignorance, mistake, and deception to name the most obvious.
I’ve retrospectively rejected a couple of my possible sightings (and almost did so with the recent “Wooden Wings” incident), partly because I’m disposed to look for reasons to doubt myself and partly because I realized that I was in some way primed in a couple of those instances; for example, in one case, I’d been looking at specimens the day before. In the absence of a good look, I think it’s better to err on the side of self-skepticism.
When it comes to claims from others, it’s impossible to know with certainty what someone else saw, even when reports are accompanied by detailed sketches and field notes, as is the case with several of Frank’s sightings. I was with him for one, on April 3, 2015, when he made notes (available at the link) and a sketch (at my urging). I have no doubt that he was convinced he saw an ivorybill, as he was visibly shaken immediately afterwards and was very slow to regain his composure.
But I really like corroboration, and to the best of my recollection, Frank and I had two possible two-person sightings in all our time together – the first was in the old search area, the day after an auditory encounter and a week before a suggestive trail cam capture was obtained, effectively from the same spot. I consider this my strongest possible sighting.
The second was much weaker, a flyover with two birds seen in silhouette within 15-20 minutes of an ADK series. (Three of my possible sightings, two of which Frank also saw, have been associated with ADKs.) We did not think much of this possible at the time but could not explain it away entirely. I should add that Frank had several other two or more person sightings, but I was not present for those.
With regard to the local reports, I’m always mindful of the possibility that I’m being played, especially as an obvious northerner in rural Louisiana, but I don’t think that has ever happened. It’s still necessary to listen closely and ask the right questions. Pileated, Red-headed, and even Red-bellied Woodpeckers can be confusing for non-birders. I’m more likely to credit accounts from people who make it clear they know Pileateds and know the difference (like the landowner in the old search area) or from people who say they thought the ‘ones with the white backs are the males’ and the ‘ones with the black backs are the females’. I’ve heard this only a couple of times. But I like it best when we get ivorybill claims from different, unconnected people referencing the same location; this has happened in the vicinity of the current search area.
To conclude, members of our group have had multiple sightings of varying quality, in two different Louisiana locations since 2009. In addition, there have been multiple claims from locals in both areas.
One final note, relevant to the USFWS review: I am aware of a number of additional post-2009 claims from Louisiana (and a few from other states). In a couple of cases, Frank interviewed the people involved and found them to be credible. A cellphone video was obtained in one instance; the observers’ excitement over apparently seeing a pair of ivorybills is evident, but the quality of the video is extremely poor. In the other case, two people saw the bird and one provided a highly detailed description and sketch. There’s also Joseph Saucier’s report from last fall; our follow-up suggested that much of the habitat in this large expanse of forest was likely suitable, and Matt Courtman, who is an excellent birder, heard what he thought was a very good double knock on one of his follow-up visits.
I recently posted a couple of Frank’s 2009 sightings and accompanying photographs that hadn’t been made public previously. These are the only photographs associated with possible sightings that we have obtained. Full details are available at the link; rather than recapitulate them here, I’ll focus on trail cam photos, including several that have been discussed previously plus a few intriguing ones from the old search area that have not been posted previously.
These are the images I think are most strongly suggestive of Ivory-billed Woodpecker and for which I am unaware of a strong counterargument.
The first two have been discussed repeatedly at length in a number of other posts. Both are from the old search area; the first was taken in August 2009 and the second, which appears to show a female Ivory-billed Woodpecker, was taken in late November, a week after the sighting discussed in the previous section.
To reiterate a couple of points about the August 2009 image: in the original photograph, there appears to be some red in the crest (though it’s not positioned right for either PIWO or IBWO). Careful review and image enhancement made it clear to me that the red is an artifact and the crest is black. In addition, as shown in the enhanced detail, there appears to be a large, light-colored bill protruding from a cavity in the background. This apparent bill is absent from captures obtained on other days. It changes position in the two frames where it is visible, suggesting it is attached to something within the cavity. (Unfortunately, the other frames from that morning have long since been lost.)
The image of a bird in flight, from the new search area, is more problematic, as it is not demonstrably a woodpecker. The still above includes an inserted Imperial Woodpecker for comparison, with motion blur reduction applied to the bird captured by the trail cam. One prominent birder stated that it is a Blue Jay, but as discussed in the original post, the bird is behind the snag and appears to be as large as or larger than a Pileated Woodpecker captured on the trunk.
Although we’ve determined that one of the initially intriguing images from that deployment shows a Red-headed Woodpecker, a few others remain interesting.
There’s an additional image from the spring of 2010 (old search area) that has always intrigued me, though I never had a chance to examine the original. I think it shows a large woodpecker, with a black crest, high on a honey locust. If it were a Pileated, male or female, I’d expect some red to be visible in the crest, even under these conditions.
Finally, there are a couple of additional trail cam captures from the old search area that I have not posted previously. By now, the original sequences are lost, and I’ve long since forgotten any details or discussion we might have had when Frank sent me the images. While they are far from conclusive, I think they are interesting enough to post in this context. At minimum, they further illustrate the challenges associated with obtaining definitive trail cam imagery.
Note: If you listen to the recordings below with headphones (recommended) be aware that some have loud sounds, clarinet toots and ADKs, in the foreground. This applies to the clip from March 2013 that I recorded and to many of Matt and Phil’s recordings from March 2017.
Possible auditory encounters far outnumber possible sightings, and over the years we have made numerous recordings of putative kents and double knocks (more than once in combination, which should add to their evidentiary value). Some visitors have been rewarded by hearing something within a day or two of arriving, and this has been true in both search areas, although others have been less fortunate, and I heard nothing suggestive during the most recent field season (though Matt Courtman recorded three possible calls in April of this year).
I will not be including all the audio evidence we’ve obtained, just some of the material I think is most significant. Some of the putative kent calls that have been recorded over the years may be Blue Jays; it has been suggested that Matt’s calls from this spring are too low pitched for ivorybill and might be from a heron, though I think they come from the same source. Regardless, many of the calls are simply inconsistent with any known bird species, are fairly close to the Singer Tract recordings harmonically, and are perfectly consistent with written descriptions of Ivory-billed Woodpecker sounds.
One of the most dramatic encounters involving both apparent kents and double knocks was from the old search area. It seemed to have been triggered by banging from the tin roof of an old deer stand. Six people were present, and much of the incident was recorded. There were apparently two birds involved. More details are in this post, and here is Frank’s 58 minute recording:
Extracts from the encounter and other material recorded in person and on remote units in the old search area and environs, as well as sonograms, can be found here, via the Wayback Machine. Some of the Wayback Machine links are not functioning, and the material posted there includes a couple of double knocks; here are several of the recorded kent-like calls:
Note that the fundamental frequencies on many, but not all, of these calls are in the 900-925 hz range, higher than the Singer Tract recordings and possibly consistent with Blue Jay. But in the second clip above, the fundamentals of both the lower calls at the beginning and the higher calls at the end appear to be a little over 800 hz., and during the extended encounter mentioned above, the calls went on for a prolonged period without the source revealing itself as a Blue Jay, which would be typical if the animal were indeed a jay.
The same applies to calls I heard and recorded in the new search area on the morning and afternoon of March 2, 2013. The morning calls went on for ~20 minutes. I did not think they were Blue Jays in the field.
We have recorded apparent double knocks in apparent reaction to ADKs and to Barred Owl calls and sometimes in contexts where the trigger is unclear, as in this example from an October 2015 post written by Frank:
“Bob and I continued upstream for another half mile, located a nice spot with a good view, and I performed an ADK series, followed about ten minutes later by a series of electronic playbacks of Singer Tract ivorybill calls. Shortly thereafter, Bob heard a double rap drum, that was captured on my digital recorder. I personally don’t believe that the drum was a direct response to my ADKs as there was at least a fifteen minute interval after the last of the ADK/playback series.
The double rap is not “perfect” in that the “intra-knock interval” is about .05 seconds longer than the “ideal” – based on averages of the intervals of other Campephilus drums – but it sounds very good.”
The 15 minute interval between ADK/playback and apparent response is similar to the lag time that has been observed in other Campephilus woodpeckers responding to ADKs; it may be worth distinguishing between a slower ‘response’ and a more immediate ‘reaction’ in this context. (M. Lammertink, pers. comm.)
In my view, the audio recorded by Phil Vanbergen and Matt Courtman in March 2017 is compelling, perhaps the most compelling evidence we’ve obtained. On the morning of March 15, they recorded over 200 calls, along with a number of possible double knocks, over an approximately three hour period. The calls have lower fundamental frequencies than many of those in earlier recordings; these frequencies are considerably lower than any kent-like Blue Jay calls I’ve been able to locate online. Other suggestions have included, Red-breasted Nuthatch (too high) and Wild Turkey (no typical Turkey vocalizations were recorded that morning).
Overall, the calls appear to be inconsistent with any other known species of bird, mammal, or amphibian. Their association with Campephilus-like double knocks strengthens the argument for Ivory-billed Woodpecker as the source. As far as I know, the sheer number of recorded calls is unprecedented for a single incident in the post-Singer Tract era.
I was lucky enough to hear a couple of the calls on March 11, the day Phil first captured them, along with a couple more and some knocks at the same location on the following day. I’ve blogged about these recordings at length in multiple posts; I won’t repeat all that material here, but the discussions are worth reading. I will repost some of the original recordings and enhanced versions.
Here is the first recording Phil made, one clip with the calls isolated and one with our talk and reactions. It may be difficult to hear the calls without headphones.
Here are some brief extracts from Phil’s March 15 recordings:
Here are some from Matt’s:
In the interest of brevity, I encourage readers to visit this post for in depth analysis, various extracts, and amplified versions that highlight the calls and some of the knocks. As always, it helps to listen through headphones.
Indirect Evidence: Bark Scaling, Bark Chips, and Cavities
I’ve discussed my bark scaling hypothesis – that a certain type or types of scaling may be diagnostic for Ivory-billed Woodpecker – multiple times over the years. As explored in a couple of recent posts, we are in the process of testing it. I won’t recapitulate the hypothesis here except to say that I remain convinced that neither Pileated Woodpeckers nor squirrels are physically capable of removing hickory bark in the manner we’ve found in the new search area. There may be other diagnostic characteristics, but my main focus is now on the hickories, since the work on that species is the most extreme outlier. Below are some dramatic examples of bark scaling (plus one image of interesting feeding sign on a sweet gum sapling that had been stripped of bark) on hickories, oaks, and sweet gums from both search areas.
For Tanner, bark scaling was one of the strongest indicators of ivorybill presence, even though he accepted some reports from trusted sources in areas where he found no scaling. The absence of scaling (along with poor habitat “quality”) was one of his main reasons for dismissing reports during his 1930s survey and rejecting the ones from the Big Thicket in the 1960s. He also doubted reports from the Chipola and Apalachicola area in 1950 based on finding no cavities or scaling during a 4 day visit. He concluded his notes on the trip by stating:
Conclusion: No I-b now in Chipola R. and neighboring Apalach R. swamps. Many of Kelso’s reports are mistaken, – not deliberately false, but due to ignorance and wishful thinking. There appear to be contradictions in some of his stories. I could not get any clear statement of what Ivory-bills sound like from him. He said that the local name of I-b was “Saddleback”, – which appears good but in many ways odd. Also “Van Dyke”! There certainly is no fresh sign in any area we visited indicating that the birds are present. The only possibility is that of scaling on pine; this may be solved by watching Pileateds in the pine woods.
(F. Bryntesson, pers. comm.)
While I think it’s possible, indeed likely, that some of the work Tanner ascribed to ivorybills was done by Pileated Woodpeckers and squirrels, there’s no question that an abundance of bark scaling is relevant in assessing possible ivorybill presence in a given area.
Based on visits, some lasting a week or more, to suspected ivorybill haunts in South Carolina, Florida, and Louisiana and on information provided by other searchers and people who have spent years in southeastern forests, I feel safe in saying that the quantity and quality of bark scaling found in both search areas is unusual. We also found an abundance of scaling in the area where Mr. Saucier had his sighting last year, something that I think lends added credibility to his report.
In the past, many in the searcher community speculated that lateral groves or bill marks in the cambium or sapwood might be suggestive, but this view has fallen out of favor; much of what was thought to be “scaling” in these instances was in fact shallow excavation with associated bark removal. We found one example of superficial lateral marks in 2013 that continues to intrigue me, although the snag in question is a young sweet gum that could easily have been stripped by a squirrel or Pileated Woodpecker.
The edges of the scaled area may tell more of a story, especially in conjunction with any bark chips found on the ground. On the trees shown above, damage to the still-adhering bark suggests that lateral blows were integral to the scaling. This would be expected if the source were a Campephilus woodpecker.
An abundance of large, hard, thick bark chips (or chunks) around the base of a tree is also a likely indicator. Pileated Woodpeckers and squirrels seem to remove bark in smaller pieces and in strips. Additionally, when squirrels are involved and the bark has been recently removed, it’s likely that there will be a lot of smaller debris in the mix, since squirrels have to gnaw their way through the bark. Some of the chips we found in the old search area seem to have bill marks consistent with Ivory-billed Woodpecker.
The bills of specimens, at least, seem to fit perfectly into indentations in the chips, but Pileated Woodpecker bills seem to be too small. The larger chips shown below are suspected Ivory-billed Woodpecker from a variety of species – oak, hickory, sweet gum, sugarberry, and honey locust. The specimen shots are with a honey locust chip from the old search area. The first few images show squirrel and Pileated woodpecker leavings for comparison.
Cavities are vexing. I’ve become convinced that there’s no bright line for distinguishing Pileated from Ivory-billed Woodpecker, notwithstanding the dimensions Cornell gave for “large” PIWO (3.5″ x 3.7″) versus “small” IBWO (4.0″ x 5.0″). The small ivorybill sample was limited to nest holes, and I’ve certainly seen plenty of oddly shaped and large holes being used by Pileated Woodpeckers. Thus, while I keep my eyes open for cavities, I don’t think there’s a reliable way to determine which species is responsible for creating or using large, irregularly shaped cavities, although extensive bark scaling on the tree may be indicator.
It has been a challenge to find cavities of any kind in our current search area, probably due in part to the high canopy. We did find a cluster of interesting cavities in what I call the northern sector in the 2014-2015 season. This cluster was in the vicinity where we’ve found an abundance of scaling over the years and where Phil and Matt made the recordings last year. Below are some of the most interesting cavities we’ve found in the two search areas. Cavities were much easier to find when we followed up on the Saucier sighting, but I’ve omitted those from this document.
In conclusion, there is an abundance of evidence suggesting that Ivory-billed Woodpeckers persist in Louisiana. It is self-evident that none of this evidence rises to the level of proof required for establishing beyond all doubt that the species has survived, but I think the case for persistence is a compelling one, without regard to evidence obtained in other areas since 2005.
Stay tuned for a more theoretical discussion of evidence, proof, and professional approaches thereto.
Thanks to Wylie Barrow, Fredrik Bryntesson, Patricia Johnson, Tommy Michot, Steve Pagans, and Phil Vanbergen for their input. Some photos by Steve Pagans and Erik Hendrickson.
While I have not been keeping close track of numbers, I’m going to give some guesstimates on trail cam results. Our current settings (~12 hours daily at 20 second intervals) result in ~2000 captures per day with cards filling and beginning to overwrite at somewhere between 80-90 days. PlotWatcher’s proprietary software makes the process a lot easier than it was with the old Reconyx images, which had to be stepped through manually. This is the first time I’ve really immersed myself in reviewing imagery, and even with the ability to review the images in a video format, it’s a tedious process involving a lot of stops and starts.
I’d estimate that I’ve reviewed about 500,000 images over the past 8 weeks, and I’m struck by the relatively small number of hits, especially avian ones. Squirrels make up a significant percentage, probably more than half, the total captures, and I suspect that Northern Cardinals are second. Other common species show up infrequently. I don’t recall seeing any identifiable Blue Jays (although there may be one among the images below). Nor were any Wild Turkeys captured (including on a camera that was mis-aimed, with the target tree at the right edge of the frame and a large open area to the left as shown in some of the captures below).
Woodpecker captures have been extremely rare, except on the targeted hickory discussed post before last, that was already extensively scaled. Among the rest of the images I’ve reviewed thus far, I’ve seen three identifiable Pileated Woodpeckers – one on a target tree (below) and the other two on an adjacent tree that had one pre-existing PIWO foraging pit in the frame – and as best as I can recall, a couple of identifiable Red-bellied Woodpeckers and a Hairy (captures not saved/posted).
I suspect it would be fairly easy to obtain Pileated imagery in quantity by targeting trees with obvious PIWO foraging sign, but otherwise, it seems even this abundant species is hard to capture in a camera trap. This should help illustrate another reason I don’t think the so-called failure to obtain conclusive ivorybill imagery using trail cams is very meaningful.
Assuming the ivorybill persists, it is a very rare species with incompletely understood behavior, and no one alive today, myself included, knows exactly what makes a good target tree. The best anyone can do is make educated guesses and hope to get lucky, against extremely long odds.
Now some housekeeping: the next couple of posts will be about evidence. I’ve never assembled all of Project Coyote’s evidence in a single post and discussed it as I plan to in the next post, which I expect to take live sometime next week. Look for a more theoretical piece on evidence to follow that. Both of those posts will be password protected for review, though I will try and make them public as soon as possible. After that (probably in August), I’m planning a postscript to the habitat and historic range series.
I won’t be doing regular updates on the trail cam imagery unless there’s something significant to report or I have insights similar to those expressed here.
And now some imagery. It should be illustrative of some of the challenges. See what you can find, including in the featured image . . .
I suggest reading Part 1 for background and context, if you haven’t already.
The target of this deployment (5/3-6/3/2018) was the sweet gum stub discussed here. The tree was killed when its top was blown off in spring 2015. A patch of recent scaling was found this season. I suspect the initial scaling is woodpecker work, but squirrel is also possible. The extent is modest in terms of what I hypothesize is diagnostic for Ivory-billed Woodpecker:
A particular and distinctive looking type of extensive scaling (large contiguous areas with bark removed) with associated insect tunnels on bitternut and pignut hickory boles – live trees, snags, and stubs – may be diagnostic for ivorybill. For recent work, the presence of large bark chips at the base of such trees is a related potential diagnostic.
Insect tunnels are present on this stub. Species is/are unknown, and tunnels are small compared to those found in the hickories.
In contrast to the hickory discussed in the previous post, there were no woodpecker captures over the course of this deployment and squirrels were very active on the scaled area, appearing on May 4, 5, 8, 10, 12, 13, 15, 17, and 22. There were multiple visits on some days, and the total time spent on the scaled area was significant, upwards of an hour, with at least one visit lasting nearly 25 minutes. It was surprising that squirrel activity ended on the scaled area ended on May 22nd, and there was none over the next 11 days.
Over the course of this deployment, squirrels removed a modest quantity of bark, apparently in strips, from part of the scaled area. They did this inefficiently – with some difficulty and with the grain. The bark, already softer and weaker than hickory, has weakened in death and is at best moderately tight (relatively easy to peel off by hand). Captures from the first and last full days (note the Hooded Warbler on the branch to the left) of the deployment reveal how little bark was removed, all or almost all from the right side of the scaled area. (Click on the images to enlarge them.)
This suggests that squirrels are unlikely candidates for removing bark from hickory boles in quantity, leaving large chips behind, or initiating extensive scaling on hickories. In my view, it’s probably impossible for them to do so. The results for Pileated Woodpecker from the hickory deployment and squirrel from this one support my hypothesis that Ivory-billed Woodpeckers are the source of the initial hickory scaling. But more data are needed.
Before turning to the trail cam captures and accompanying images of the scaled surfaces, I’ll provide some background information on the impetus for this post and on squirrel behavior.
An email discussion of squirrels and bark scaling was ongoing prior to my starting to review the images from this deployment. Wylie Barrow suggested an alternative explanation: that squirrels might be the source of much of the scaling (including the work on hickories) that’s taking place in the search area. He pointed out that . . . “Squirrels have removed bark from 1/4 to 1/2 of the trunk and several large branches from large oaks in my yard… and they work with great speed. They often leave large bark chips on the ground beneath the trees. Trees are living and bark is tight and fairly thick.” (W. Barrow, pers. comm.)
At first, I took some umbrage at this suggestion, thinking that I had thoroughly examined and considered what squirrels might be doing on the hardwoods in our search area and what the upper limits of their capacities might be. While my basic views on this are unchanged, and the trail cam images tend to support those views, I’m grateful to Wylie for keeping me on my toes.
It’s certainly true that in the past I have failed to consider squirrels and the role they might play in bark scaling, and this has led me down some blind alleys, as was discussed in a series of posts in early 2016. I have also been too confident in those conclusions, even though I think this material supports them. Wylie’s suggestion led me to conduct additional online research on squirrels (and he provided additional references).
I had a number of off-the-cuff theoretical and observation-based objections to Wylie’s suggestion.
One evolutionary objection is reflected in a comment I made early in our exchange: “the predator in question would have to have evolved to take advantage of this very narrow window of opportunity when the insects are near the surface . . .” I thought and still think this points toward a woodpecker as the source, and toward a Campephilus woodpecker in particular, since this foraging strategy is characteristic of the genus.
The hickory scaling is associated with sapwood dwelling Cerambycid infestation, and signs of woodpecker activity (targeted digging around exit tunnels) are present in all cases. The homepage tree was very recently scaled when found, and woodpecker evidence was present. Wylie replied that squirrels are opportunistic and might be feeding on larvae; he went on to suggest that woodpeckers following the squirrels and doing targeted digs around the exit tunnels was a possibility.
In one paper on a tropical species of squirrel, it was observed that they prefer palm nuts infested with beetle larvae. The authors also note that squirrels have a strong preference for obtaining food in the most efficient manner, and that Eastern gray and fox squirrels will choose nuts lacking an endocarp (the hard inner shell) over those that are harder to open. When confronted with an endocarp, the tropical squirrels would attack it at its weakest and thinnest point, as do Eastern gray and fox squirrels :
Two of these pores have dead ends (with 1-mm depth), and the third is the germinal pore, which is deeper but is closed by a soft and easily penetrable tissue, located on the side opposite the fruit’s internal gibbosity. The internal gibbosity is a projection of the endocarp that inhibits the squirrel’s access to the endosperm when the fruit is opened from the side containing the dead-end pores. The squirrel must determine the position of the internal gibbosity to avoid it and thus save energy and time in obtaining the endosperm. These rodents are known to identify the side without the internal gibbosity even before beginning to open the fruit, with >90 percent success (Bordignon et al. 1996, Mendes & Candido-Jr 2014). However, how the squirrel identifies the side without the internal gibbosity remains unknown. As the gibbosity is always on the side opposite the germinal pore (Bordignon et al. 1996), this pore is an important access point that the squirrel can use to open the fruit efficiently. It is believed that the squirrel manipulates the fruit by pressing the three pores with its upper incisors, using the pore without a dead end for support so that the lower incisors can open the endocarp (Bordignon et al. 1996).
Efficiency is one of the main factors that determine the foraging strategy of Sciuridae. A laboratory study conducted with the squirrels S. carolinensis and S. niger found that individuals preferred various species of nuts with low energetic value that lacked an endocarp or shell over high energy nuts with an endocarp (Smith & Follmer 1972). These results suggest that there is a high cost in energy expenditure for processing seeds with endocarps for these species.
(Alves et al. “Queen palm fruit selection and foraging techniques of squirrels in the Atlantic Forest,” Biotropica 50(2): 274–281 2018). Efficiency is an important consideration in this context, especially with respect to hickories.
The reasons squirrels strip bark are poorly understood. Pine (or red) squirrels attack a number of tree species, “[d]uring winter, spring, and early summer, bark stripping and tree girdling for consumption of phloem and cambial tissues is common (Hosley, 1928; Linzey and Linzey, 1971; Pike, 1934). Pine squirrels also eat the bark of rust galls (Salt and Roth, 1980) as well as sap from sugar maple trees (Acer saccharum) in the northeast (Hamilton, 1939; Hatt, 1929; Heinrich, 1992; Kilham, 1958; Klugh, 1927; Layne, 1954) and yellow birch (Betula alleghaniensis) in the Great Smoky Mountains (Linzey and Linzey, 1971). Widespread, systematic sugar tapping by pine squirrels occurs in New England (Heinrich, 1992).” (Steele, M. A. 1998. “Pine squirrel (Tamiasciurus hudsonicus),” Mammalian Species 586:1–9).
Red squirrels have also been observed feeding on spruce bark beetles. (Pretzlaw, et al. “Red Squirrels (Tamiascurius hudsonicus) Feeding on Spruce Bark Beetles (Dendroctonus Ruffipennis): Energetic and Ecological Implications”, Journal of Mammalogy, 87(5):909–914, 2006). This was a novel observation at the time, and the behavior appears to have been a sudden and opportunistic response to a climate change-related bark beetle outbreak that lowered cone production. Spruce bark is soft, flaky, and fairly loosely adhering, and the bark beetles spend approximately a year, the entirety of their larval life cycle, in the phloem and hence are a readily available food source for a prolonged period. Moreover, “[f]oraging for larval spruce bark beetles by red squirrels is an obvious and stereotyped behavior; squirrels situate themselves on the trunk of the tree near ground level and peel off the bark to reveal and ingest larvae.”
There seems to be less agreement as to why Eastern gray and fox squirrels strip bark. It has been suggested that a calcium deficiency might be primary driver. C.P. Nichols et al., “A novel causal mechanism for grey squirrel bark stripping: The Calcium Hypothesis,” Forest Ecology and Management 367 (2016) 12–20. Bark stripping by Eastern gray and fox squirrels seems to be more prevalent in areas where the species have been introduced, “[b]ark-stripping behaviour, reported so often in Europe (Shuttleworth et al. 2015), is extremely rare in their native range (Kenward 1989).” (Koprowski et al. “Gray not grey: The ecology of Sciurus carolinensis in their native range in North America”, posted on Researchgate.com, 2016).
While “extremely rare” is an overstatement, it does appear that bark stripping occurs more frequently in areas where gray and fox squirrels have been introduced. It is a major problem in the U.K and Europe but mostly an annoyance in the United States. It seems reasonable to infer that it is more common in suburban and residential areas than in mature bottomland hardwood forests, though Wylie points out that the discrepancy in the reporting may be due to demographic factors and that squirrel behavior in bottomland hardwood forests has been poorly studied.
Gray and fox squirrel bark stripping seems to occur most frequently on branches, and I found no images in which insect infestation of the scaled areas was apparent. In addition, the examples of extensive squirrel scaling found online in no way resemble what we’re finding on hickories. Thus far, we have found only two references to squirrels stripping bark from trees in the genus Carya, one from pecans in Georgia and one from limbs in West Texas pecan orchards, where fox squirrels have been introduced. It’s not clear what parts of the trees were involved in Georgia and whether this report also came from an orchard, but regardless, pecan bark is flaky and not criss-crossed, making it easier to scale.
While neither Wylie nor I conducted an exhaustive literature review, we found no records of gray or fox squirrels scaling bark from any bitternut or pignut hickories (Carya cordiformis and Carya glabra), be it on limbs or boles, in several Google searches. Given the extensive range of these species – most of the Eastern United States and into Canada – and the association between squirrels and oak-hickory habitats, if squirrel scaling of hickories occurred with any regularity within the natural ranges, one would expect references to be abundant in both the popular and scientific literature.
As mentioned in the previous post and implied above, I suspect that the criss-cross pattern that characterizes pignut and bitternut hickory bark is one factor that deters squirrels from removing it and may prevent them from removing it in large pieces. This relates more generally to the question of efficiency. The characteristics of hickory bark make it extremely difficult for any creature to remove. In addition to the pattern of the grain, it is literally the hardest, strongest, thickest bark in the forest. On mature boles it can be 3/4″ thick (compared to around 1/16″ for a hickory endocarp). It is tight (though less so when sap is flowing), and it retains these characteristics long after death. Bitternut hickory bark does not flake, and pignut does so infrequently and superficially.
Thus, both species are exceedingly poor candidates for stripping by squirrels, especially when sweet gums and an array of other much easier targets are available. In contrast to the hickories, the target tree in this deployment was a sweet gum, three years dead, with thinner, considerably softer, loosening bark
As I see it, all of this militates against squirrels as the original source of the hickory scaling. While this is inferential and we have yet to document whatever initiates the scaling, the data obtained thus far support the inference. Only recently have we been able to deploy enough trail cameras for a meaningful and sustained effort. Nevertheless, we have had many hours of captures since 2009, in both search areas. To my knowledge, the only prior unambiguous capture of squirrel scaling is the one from 2015; it involved a downed, immature sweet gum with thin bark, which was easy for squirrels to scale. A second clip may show a squirrel removing a very modest quantity of thin bark from a sweet gum limb that was already being scaled by Pileated Woodpeckers (second video clip at end of post), and Wylie observed a squirrel scaling a sweet gum branch (on a roadside just outside the main search area) in December 2015.
I no longer think scaling on sweet gum limbs (so heavily emphasized in Tanner) is a strong indicator of ivorybill presence, at least not on its own, although what we’ve found in the search area seems to be unusual. Abundance, lack of correlation with low mast years, bark chips, absence of incisor marks, and indications of woodpecker activity, especially targeted digging, may all be suggestive. Sweet gums, which are very attractive to beavers, are likely one of the most desirable targets for squirrels as well, for reasons of flavor and efficiency.
But we have documented no squirrel scaling on hickories, live or dead, on limbs or on boles, partially scaled or with bark intact.
I think the results from this deployment shed considerable light on the issue of squirrels and bark scaling, especially what they do (or can do?) on a mature bole with thick bark. So let’s go to the videotape, as a New York sportscaster used to yell.
Squirrels on a Sweet Gum Bole
As with the previous post, our Plotwatcher Pro trail cam is programmed to capture one image every twenty seconds, and these time-lapse sequences have been converted into QuickTime movie format. If you want to get a clearer sense of how the squirrels are behaving, you can step through the films frame-by-frame. If you elect to watch just one of the clips, the one from May 8 that starts at Frame 1500 (the squirrel spent 24 minutes on the scaled surface) or the one from May 12 that starts at Frame 1574 might be your best bets. Discussion and close-ups of the scaled surface follows the bonus imagery.
While we had no woodpecker detections on the stub and bird captures were few, we did catch some hogs (piglets?) and a beaver. Also captured but not shown were a Northern Cardinal and an Eastern Phoebe.
Discussion and Details
As best I can tell, the only expansion of the scaled area involved a narrow strip at the upper right, probably no more than 12″ x 2″, and a little widening at the very top, although this was an area where the squirrels spent a considerable amount of time.
Let’s look at some details from that scaled area.
While there appears to have been some woodpecker excavation at the middle left of the larger scaled patch, there’s no readily apparent sign that woodpeckers have been after the insects that are feeding in the sapwood. Nor is there any strong indication that squirrels were feeding on insects over the course of this deployment, though it’s possible they took advantage of snails and beetles, like the ones in the photo, or slugs, which I also saw on the scaled patch.
The edges of the bark shown in the close-ups, especially the one at the top, show signs of having been gnawed, although this is subtle, and sometimes impractical as an identifier, since such close examination is not always possible in the field. I presume that the abundant squiggly abrasions to the surface of the underlying wood are incisor marks, something we have not observed with other scaling we’ve found.
With regard to what was left behind, the first three photos show what I found at the base of the snag when I discovered the scaling on May 1, 2018.
The large, though narrow, strip of bark was the biggest one I found at the base and is one of the main reasons I suspect that woodpeckers initiated the scaling with squirrels following, although I would not rule squirrel out completely. In any event, the bark was so soft and weak that it broke in my hand when I picked it up on June 11. The other thin strips are more consistent with what I’d expect for squirrel, and the tiny orange pieces of cambium are a giveaway.
The situation had changed little during this most recent trip. The picture with my boot shows the larger pieces of bark I found at the base, including the one shown above after it broke. They may be consistent with woodpecker (possibly including Red-bellied or Hairy), but I suspect that both squirrels and woodpeckers were involved in the bark removal.
Edited to add: For any extensive squirrel work on mature boles, especially hickories, I would expect to find many small pieces of bark on the ground, similar to those shown above, as in this dramatic example.
My main objective in targeting this stub was to observe it over time, more for what might happen as the decay advanced and whether it might become a target for ivorybills; it’s the type of “stump” that Pearson described as being favored by ivorybills after his visit to the Singer Tract in 1932, though Pearson’s “stump” (scroll down in the linked article) was much longer dead. The bark scaling, while interesting, was in the “could have been anything” category. Getting this data on squirrels was a pleasant surprise, one that I should have anticipated based on the small bits of cambium on the ground. My bias came into play, as I ascribed them to a smaller woodpecker. Between Wylie and the trail cam results, I’ve learned a lesson. In terms of the bigger picture, however, the results so far suggest that squirrels are not the source of the putative Ivory-billed Woodpecker scaling on hickories.
Summary/Bottom Line Up Front
5 of our 8 trail camera deployments (a 9th is planned) are targeting unscaled hickories, stubs and live trees that show signs of damage or decline. The main goal is to capture images of Ivory-billed Woodpeckers by identifying potential foraging trees and targeting them in advance of their being scaled (a long shot but the best strategy I can think of). A secondary purpose, especially with targets that already show some scaling, is to test my feeding sign hypothesis by documenting work by Pileated Woodpeckers and Eastern gray or fox squirrels (a much easier task). These are the only other local species that have the capacity to remove large quantities of tight bark from the boles of mature living and recently dead trees, though I suspect this capacity is limited. Results from two deployments, one on a still living but compromised and extensively scaled hickory (4/29-6/6/2018) and the other on a sweet gum stub (5/3-6/4/2018), suggest to me that neither Pileated Woodpeckers nor squirrels are the source of the initial, extensive hickory scaling.
Part 1 will address Pileated Woodpecker activity on the hickory, followed by a brief discussion of my visit to the search area from June 6-11. Part 2, which should be up within a week or so, will focus on squirrels, especially their activity on the sweet gum stub, which I think is analogous to the Pileated Woodpecker activity on the hickory. These are preliminary findings that may be contradicted by data obtained in future, but thus far they support the suggestion that Ivory-billed Woodpeckers are the source of the initial scaling.
Our PlotWatcher Pro trail cameras are dedicated time-lapse (as opposed to motion- activated) units. We set them to capture images at 20 second intervals, approximately between sunrise and sunset. The proprietary software makes it possible to create QuickTime movie versions of the time-lapse stills. The QuickTime versions are worth watching and are illustrative of behavior, but it’s important to remember that they are compilations of still images, not actual videos. Thus, it’s also worth stepping through them frame by frame. This will be an image heavy post but with a lot of analysis. I hope you’ll stay with it.
I am not planning to do posts on every review of every trail cam deployment but will post results that I think are a relevant or interesting to readers – PIWOs and squirrels spending time on scaled areas or removing bark, most or all woodpecker hits, mammals that are infrequently seen, identifiable birds, but not squirrels running up and down tree trunks.
Before turning to a discussion of the trail cam imagery, I’ll take this opportunity to restate and elaborate on my hypothesis. The imagery and discussion will be familiar to some readers, but I hope the new data will shed additional light, even for those who know the material.
Initially, I hypothesized that certain types of bark scaling might be diagnostic for ivorybill. Bear in mind, my focus has been on identifying a diagnostic, not covering the full range of what Ivory-billed Woodpeckers might do, so my criteria, especially this specific restatement thereof, may be overly narrow. Over time, and based on trail cam captures of Pileated Woodpeckers and a squirrel doing some scaling that seemed to match Tanner’s descriptions, I have refined and limited the hypothesis as follows:
A particular and distinctive looking type of extensive scaling (large contiguous areas with bark removed) with associated insect tunnels on bitternut and pignut hickory boles – live trees, snags, and stubs – may be diagnostic for ivorybill. For recent work, the presence of large bark chips at the base of such trees is a related potential diagnostic.
The tree on the home page (shown below) is one example. Numerous insect tunnels, some with signs of having been expanded by woodpeckers, must be visible. The appearance of the scaled areas is such that I can recognize older work even in the absence of chips. Because standing hickories (in drier areas at least) are slow to decay and the bark can remain tight for years, older examples persist, and I’ve found many of those over the years.
The hypothesis is founded on Tanner’s descriptions of scaling (although bitternut and pignut hickories were absent from the Singer Tract) and on the anatomical characteristics of Campephilus woodpeckers versus those of Dryocopus or Melanerpes, which appear to lack the physical capacity to do this type of work or do it extensively. The characteristics of hickory bark – chemical composition, strength, tightness, and thickness – are also central. I think the same limitation applies to squirrels, at least on hickories. This is not to exclude squirrels and PIWOs from coming along after the initial scaling and expanding it.
It recently struck me that grain may play a major role in limiting Pileated Woodpeckers and squirrels from scaling on hickories. Both species seem to follow the grain when stripping bark. Bitternut and pignut hickory bark differs from that of many bottomland hardwoods in having criss-crossed grain, making it considerably harder to remove, except in narrow strips like these, presumably removed from the target tree by a PIWO last year.
Based on the recent observations, PIWOs are more likely to dig through hickory bark to reach the sapwood than they are to scale it. I suspect it’s the case with the presumed 2015 PIWO “scaling” on hickory shown below, which is the only example of that type of work I’ve found. The differences may appear subtle in the photo, but they are more pronounced in situ; the work is patchy and discontinuous, and there are few or no exit tunnels. The chips at the base of this tree were small and included sapwood, indicating that this was actually shallow excavation with associated bark removal, something that is often and easily confused with true scaling.
Since 2013, we have found approximately 20 hickories that appear to have been recently or freshly scaled. Below are additional examples, plus some that were presumably several years old when photographed. In the case of the tree on the home page, the scaling was very fresh when found. I have not found any recent looking work of this type since March of 2015, though Phil found some last year that were probably scaled during 2016; one of these is the target tree in the sequences shown below. Steve Pagans has found several examples, but I’m not sure of any recent ones. Note the expansion of the exit tunnels, which makes it evident that woodpeckers have been active on the scaled surfaces.
A Cerambycid infesting the trees has been identified as Parandra or Hesperandra polita, a medium-sized, bole dwelling species with a 3-5 year life cycle.
H. polita remains, apparently of 2 adults, were found in one of the ivorybill stomachs examined by Cottam; it was collected in West Carroll Parish, August 19, 1903 and is referenced by Tanner. The adults of this species remain under the bark for some time after metamorphosis.
There may be additional subtle features, including chip characteristics and general appearance, that are relevant to other tree species and may be indicative if not diagnostic. A very nearly dead sweetgum, which had an intriguing cavity, from the old search area, is one example.
Trail Cam Captures – Hickory – April 29 -June 6, 2018
Over the course of this deployment, Pileated Woodpeckers were photographed on the target tree on 10 of 36 days, sometimes briefly and sometimes for periods of over 20 minutes. On some days in the latter part of the deployment, intervening foliage partially or fully blocked the lens, so it is possible there were additional visits. The camera is trained on the base of the scaled area, which extends to the broken top of the bole, and some of the sequences make it clear that pileateds are visiting the upper part of the scaled area as well. Phil Vanbergen found this tree last season; it is still living, although the sapwood is already somewhat punkier than what I’ve found on many longer dead hickories.
The first important point is that very little bark was removed over the course of this deployment, despite considerable time spent by Pileated Woodpeckers on the scaled area. The first two images are from the first and last days of the deployment for purposes of comparison. The new work is so minimal that it is only noticeable in the field upon close examination of the edges.
Pileated Woodpeckers were photographed on May 1, May 3, May 8, May 12, May 21, May 25, May 28, May 30, June 1, and June 4. A Hairy Woodpecker was captured on April 30 and on May 12. There were no squirrel captures on the scaled area. Individual frames and time-lapse films (when birds were present for more than a couple of captures) of the Pileated sequences follow. If you choose to watch/step through just one clip, I suggest the one from May 28, in which two birds are present, but viewing all of them will provide a better sense of what took place.
In addition to removing only a modest quantity of bark along the already scaled edge, despite being present on the scaled surface for over an hour altogether, it appears that the Pileateds are not actually stripping the bark but instead are excavating through it. This is also suggested by the quality of the bark chips on the ground. (The caption predates this imagery.)
The same appears to be the case for the presumed Pileated “scaling” on hickory in the preceding section. It’s unclear what the woodpeckers were seeking in terms of food, since there is no evidence of current or recent Cerambycid tunnels. In any event, the sequences give evidence of what Pileated Woodpeckers do on hickories, and just how inefficient they are at removing bark.
In the images we captured last year from a nearby deployment (on another scaled tree Phil found, approximately 100 yards away), the Pileated appeared to be scaling bark rather than digging through it, and this too was a very inefficient process. It involved removing long narrow strips, distinctly different from the chunks of bark shown above. In a sequence Frank and I captured on the homepage tree back in 2013, a Pileated appeared on that very recently scaled tree but removed no bark at all; a Hairy Woodpecker also appeared and removed a very small piece. (Scroll down to the bottom of that post for the relevant images.)
The trail cam captures validate my prediction that the source of this most recent work is Pileated Woodpecker.
Thus, we have documented multiple instances of Pileated Woodpeckers foraging on trees suspected to have been initially scaled by ivorybills. None of the imagery thus far shows pileateds removing bark rapidly, efficiently, or extensively. Such bark as they do remove is from areas that are already compromised, although I have found a single example of hickory bark presumably removed solely by a Pileated Woodpecker; it differs markedly from what I’ve hypothesized is Ivory-billed Woodpecker work. While it is difficult to prove a negative, the data gathered thus far suggests that Pileated Woodpeckers do not or cannot extensively scale tight bark from hickories. I suspect they lack the ability to do so.
Other Interesting Trail Cam Images
While these trail cams have limited usefulness, except at close range, all of our cameras are close to their main targets. They are positioned close enough for me to recognize lizards and even insects on the trunk but not identify species. Identification problems can exist for small birds too. I suspect the bird in the image below is a Brown Creeper, though I’m prepared to be corrected as it is so poorly resolved. If it is a Brown Creeper, it would be the latest record for the species in Louisiana.
And this series shows what appears to be a bobcat, at the lower right.
Some Brief Comments on the Trip
Nothing strongly suggestive of Ivory-billed Woodpecker was seen or heard – no ambient sounds or responses to ADKs. Tommy Michot joined me on Saturday, Sunday, and Monday morning. We deployed the 8th trail cam, aiming it at the bole of a sweet gum snag that has a scaled area; I suspect Pileated Woodpecker was the source, and squirrel is also possible, but the decay state is within the range Tanner associated with ivorybill, so this will be a long-term deployment.
Low water made it possible for us to explore some previously unvisited areas. Tommy did several stakeouts where Matt made the recordings this year and where I heard wingbeats and had a possible sighting. On Sunday evening, he observed a Pileated near the sycamore roost where I photographed one in April, and we watched it flying away from the roost on Monday morning. No wingbeats were audible in either encounter, as would be expected for Pileated Woodpecker as opposed to ivorybill.
Temperatures reached the 90s by noon every day, depressing avian activity and making fieldwork unpleasant. There were heavy rains on the afternoon of Sunday, June 10, so the 11th was particularly challenging in terms of heat, humidity, and suddenly high water levels. I had a close encounter with a cottonmouth that dropped from a tree limb within a foot of my head, landing in a slough that had been dry the day before.
One personal highlight was the abundance of chanterelles, especially after Sunday’s rain. I came home with several pounds, though transporting them was a challenge, and I couldn’t keep them clean. Despite a little grit, they’re delicious.
Here are some additional photos. There’s one new example of extensive bark scaling from the upland area shown in detail; it is interesting, but I would not consider it highly suggestive. There was also some recent scaling on a sweet gum that may be more suggestive. I was using a rented 4/3rds camera and a lens with an effective 200-800 mm reach. I had trouble adjusting to the electronic viewfinder, but it was good to get a closer look at one of the cavities found this season. I also found feathers from a Pileated Woodpecker that was presumably recently taken by a raptor; this was in an upland area within 50 yards of where I parked. This is the first time in all my years in the field the that I’ve found PIWO remains.
Stay tuned for Part 2. Like this one, it will initially be password protected so that colleagues can review it before publication.
Thanks to everyone who responded to my request for comments about the blog and its future. It’s very gratifying to have received such positive feedback and to know that my work is appreciated. With that in mind, the comments and emails have convinced me that I shouldn’t make any major changes, even if it would have been wiser to have deliberated more before posting and amending the recent post about a possible sighting.
Apologies to those who were unable to view this post due to the temporary password protection while it was being reviewed.
As mentioned previously, we currently have six cameras aimed at potential feeding trees. Tommy Michot will be deploying the seventh (on the tree discussed below) in the near future. This will put us at or near what I consider full capacity in terms of servicing the cameras and reviewing the imagery. Since Frank’s death, the task of reviewing images and working with the cameras had fallen almost entirely on Phil Vanbergen’s shoulders, and we’ve only changed cards when there’s been evidence of woodpecker activity on the target tree but have otherwise allowed them to overwrite. Recently, several people have volunteered to review the imagery, and Tommy and I have gotten a better handle on deploying the cameras. Thus, we’ll be able to shift to replacing the cards every few months and (hopefully) reviewing all the images we capture.
On the last trip, Tommy and I found recent bark scaling on a sweet gum stub, and although the bark was loose, we decided to move one of our cameras and target the stub.
The camera had previously been aimed at a hickory that lost its top in a storm last year, a location where Peggy, Phil, Tommy, and I heard a double knock last year. For some reason, the overwrite function failed, so the card was full when Tommy and I redeployed the camera. I brought it home to review. I found some interesting imagery that I’ll share and discuss. I think that at least one of the frames is suggestive of ivorybill, although it is far from conclusive, and no one I’ve shared it with is nearly as interested by it as I am. Regardless, I’m now convinced it’s a good idea to check every card and am very glad that Geoffrey McMullan encouraged this approach.
First a few comments on trail cams, which will update some of Frank’s observations from this post on the subject written in January 2015.
Our trail cams are made by Day 6 Outdoors. The Plot Watcher Pro is the only dedicated time-lapse mass market trail camera. It’s lightweight, easy to use, and inexpensive compared to Reconyx, the second best time-lapse option. The accompanying software makes it possible to step through the captures frame by frame or to view them as video at various speeds. (Our time lapse interval is 20 seconds.) This simplifies the reviewing process, at some cost to image quality, although this has always been a problem with trail cams.
Trail cam sensors are relatively low resolution; only in the last year or so have 20 megapixel cams gone on the market, and many brands are in the 10-12 MP range. The bigger problem is that these cameras are designed to photograph terrestrial creatures, mostly large mammals and Wild Turkeys. It’s also almost impossible to avoid backlighting over the course of an entire day. Thus, unless the subject is at fairly close range, the chances of getting an identifiable image are greatly reduced, as will become clear from the examples below. So let’s look at some of the captures. I used Let’s Enhance, an internet based, automated image enhancement program.
The card ran from January 24 – May 1, and I worked my way backwards. During that time, there was one clearly identifiable woodpecker capture – a PIWO that spent under 40 seconds on the target hickory.
Note that WordPress has changed the way photographs are presented, at least in the format I’m using. To enlarge the images in this section, click on the image, and it will appear in a separate frame. There’s an option to see the full-sized image at the lower right. You can further enlarge the full-sized images by using the cursor.
The first large woodpecker capture was on March 16. The bird, which is not clearly identifiable, was present from ~3:25-3:46 pm, though it was not visible in all frames. I’ve included a QuickTime video covering the entire period (and about 10 additional minutes), along with a few enhanced screen caps with a box around the bird, which first appears on the right fork of the beech tree to the left of the target, then flies to a high branch in the background and moves around before disappearing.
The next series that interested me is from the morning of March 9. A bird in flight is visible in one frame captured at 8:22 am. It appears near the beginning of this 50 frame QuickTime animation.
This is an enhanced JPEG of the relevant frame with a box around the object of interest.
While others disagree, I think the object in the box is the upper (dorsal) surface of a bird’s wing. (One reviewer suggested it could be a squirrel’s tail.) There is nothing to suggest the bird in question is a woodpecker (if it’s a bird at all, which it appears to be based on the time lapse). To my eyes, however, it has a pattern of black and white that is consistent with the wing of an ivorybill, if the white is on the trailing edge. Given the distance from the camera and the extent of the white on the wing, I’m confident it is not a Red-headed Woodpecker. I’m also confident that this is actually an object, based on a review of the surrounding frames and on experimenting with color, exposure, and white balance. I’ve included a few examples, in a tiled mosaic to illustrate. Again, click on the individual images to embiggen.
A third frame that I found somewhat less interesting was captured on the morning of February 16th. I’m less intrigued by this image. It has been suggested that it might show a Pileated Woodpecker wing or a Red-headed Woodpecker on the target tree. I’m not sure what’s going on, but I think it may show a woodpecker with a lot of white on one of the smaller trees behind and to the right of the target (which would make it too large to be a Red-headed Woodpecker). I suspect that the bright patch of white on the bird is an artifact, since it matches some of the highly reflective leaves below and to the right.
Regardless of whether any of these captures involves an Ivory-billed Woodpecker and despite the problems with image quality (which I hope are even more apparent to readers of this post), the images highlight the importance of reviewing all cards and having as many cameras functioning as is possible. I’m hopeful that this more extensive coverage, which is at or near the maximum that’s practical for us, will be productive in the future and may solve the mystery of what’s doing the initial scaling on the hickories. But as I noted recently, this will require patience, and it may take a few years.
Regular readers are no doubt familiar with some of the images shown below. The “neck bird”, which was photographed in our old search area in August 2009, has been discussed in a number of posts. I think it likely shows a female Ivory-billed Woodpecker, something that became clearer once I satisfied myself that what appeared to be red in the crest (in the wrong place for either Pileated or ivorybill) was likely an artifact and that the crest appears to be all black, as shown in the enhanced image below. The original captures (the first taken a minute before the neck bird appears and the second showing the neck bird) are immediately below that for those who haven’t seen them.
Years after the capture and probably after the first post about this image in 2014, I noticed that an object suggestive of a light colored bill was visible in both frames, apparently protruding from the lower cavity in the snag to the right of center. While I have shared this information privately with a number of people and did a vague Facebook post about it a couple of years ago, I’ve hesitated to blog about it or discuss it in detail. That changed after I showed it to Jay and Erik before we parted company on my last trip to Louisiana. When Erik suggested that the object might be a vine or some other intervening vegetation, I decided to go back through my files. I discovered that Frank had sent me several additional captures from the same deployment. I examined these frames and found that the apparent bill was absent from all of them.
Below are details from frames 1095 and 1096 showing the apparent bill, which changes position slightly from one frame to the next. The time lapse interval between images was 1 minute. Again, the cavity in question is the lower one (below the fork) in the snag to the right of and behind the one on which the neckbird is seen in 1096. These snags are black willows (Salix nigra), and the neck bird snag (with the large cavity apparently being used by a squirrel) fell between November 2009 and January 2010. I’m also posting the close-ups in tiled mosaic format so they can be viewed side-by-side.
For this round of image processing, I used Let’s Enhance, which enabled me to retain a large format for cropped and zoomed versions.
Next are two details from images captured a few days later. The possible bill is nowhere to be seen. The same is true for the other captures from this deployment. Thus intervening vegetation and artifact can be ruled out.
To summarize, the following two photos show what appears to be a bill in the cavity:
2009-08-11 7:48 am (image 1095.jpg)
2009-08-11 7:49 am (image 1096.jpg – the neck bird photo)
The following photos show a cavity with no apparent bill:
2009-08-14 6:19 am
2009-08-14 6:20 am
2009-08-14 6:21 am
2009-08-14 4:06 pm
2009-08-14 4:07 pm
2009-08-15 6:20 am
If this is a bill, it appears to be large and light colored, consistent with Ivory-billed Woodpecker. Both Erik and I noticed that, in frame 1095, a topmost part of the white dorsal stripe also may be visible. When Jay first saw the photos, he was reminded of the Neal Wright photos from Texas. Some images from the Singer Tract also come to mind.
Thus, this apparent bill resembles those of known ivorybills in cavities – in size, shape, orientation, and contrast. It is present only in frames 1095 and 1096 (the latter of which shows another possible ivorybill); it changes position over the course of a minute, from one frame to the next. There is no way to be sure images 1095 and 1096 show an Ivory-billed Woodpecker in a roost hole, but these facts, especially taken together, suggest that they may.
When I look back at what transpired in the old search area between August 2009 and November 2010, when the adjoining parcel was logged, it’s extraordinary. I may revisit those events in a future post or two.
For now, I’ll close by tying this into the Bits and Pieces series. The old search area is not one that would be deemed suitable under most habitat models. The images above were captured in a stand of black willows at the edge of a bean field. The other trail cam capture, where I had a sighting, was also within perhaps 30 yards of that field. When I look back at my assessment of the habitat from the time, I think I somewhat naively overstated its quality; however, there was a good deal of dead and dying timber, and it was in close proximity to several much larger habitat patches. If we did indeed capture ivorybills with our trail cams, their presence in this area may point to how the species has been able to adapt to more fragmented habitats.
Thanks to Erik Hendrickson for his input on this post and his help in making it clearer.
Part 1 is here. Part 2 is here. I encourage you to read those posts first, as this one only adds a little to what’s been documented and discussed. Phil Vanbergen found an additional sequence this morning, one that shows the balance of the scaling that was done on the limb between December 22nd and January 26th.
In this sequence, Pileateds do most or all of the work that I had ascribed to squirrels. As in the other two sequences, the birds spent approximately fifteen minutes foraging, removing a couple of small patches from the lower part of the limb before working somewhat more extensively on the upper left. As in the other sequences, it appears that one bird did the vast majority of the work, but I haven’t found any frames that are well enough resolved to determine whether it was the female or the male. Per Phil, squirrels can be seen in additional sequences, “often poking around the scaled patches”.