Mystery Bird: Araripe Manakin, Antilophia bokermanni

Ooh, pretty.

*squee!!* One of my favorite bird families! Werner was a lucky, lucky fellow getting this little man named after him.

I'm baaack! The problem seems to have gone away by itself. That is one stunning bird and one of my favourite families too, Sheri. I assume that by its name it is an endemic of this region.

What a gorgeous little bird - Can't pass a manakin without mentioning the red-capped moonwalk:

http://www.youtube.com/watch?v=T2Bsu4z9Y3k

Hey Sheri,

I hope that was for me, still blind until later this evening and now pondering the "differently crested" soldadinho-do-araripe, a critically-endangered species that commemorates Brazil's most famous herpetologist?

Ashok,

Just like flickr et al, I can't access youtube from work, but I suspect that the title of the video was a little misleading (if this is the one titled something like "King Bird of Paradise vs. Red-capped Manakin"?) which does feature a Red-capped at one point, but the "moonwalking" was displayed by a Golden-headed (Yellow-thighed) Manakin, Pipra erythrocephala.

Extraordinary, nevertheless!

Adrian, I think only about 50 individuals remain at this location, and if you can navigate a little Portuguese, this provides a little more information on the plateau (chapada).

Thanks for the link David, the English version of Birdlife gives the total population as 250-999, with a suspected 800 birds and decreasing. In 2005 only 8 nests were found, so I fear that your 50 individuals may be close to correct. As the bird was only discovered in 1998 this is another bird that is critically endangered due to deforestation and recreational use of the land. The area is classified by the Brazilian government as "sustainable use", unless you are this little gem of a bird, that is.

Adrian -- what are previous year's nest totals like? I ask because my limited experience of tropical nest-searching is that it's very difficult, even for experienced searchers. It's possible that this bird is an exception, or that there is a cadre of researchers with a good deal of experience with them specifically, but otherwise I don't think that a failure to find nests really reflects on the population.

I have to say that I was impressed with your forensics of the Western Meadowlark from a pretty messy pile of feathers on my first visit to this blog in December but seriously David, you've just identified the last four or five birds within 5 comments without even seeing the photos (I guess that also goes for every workday in prior weeks)! How do you do it and why do you keep on coming back for more?

Hello David, my mistake, that should have read 18 nests not 8. Unfortunately there are no figures previous to 2005, but there is a note that a serious fire in 2005-6 devastated an area known to hold 7 active nests. All known nests have been found in vegetation overhanging streams so I suspect a fairly accurate count can be made in an area of "only" 28sqKm. An estimated population of 800 birds is given, but which year this applies to is not stated.

Oh.. thanks Maggie! I guess I am still a child and a challenge is still a game! Anyway, gleaning from the likes of Paul, Sheri, John, Adrian et al helps fill in the [many] blanks...

Adrian/Paul- I found this little blurb on our friend: "In 2000 there was an estimated population of less than 50 individuals and it was considered as one of the rarest birds in Brazil and in the world. Only three males and one female were found until that date. In 2003 the estimations were more optimistic and BirdLife International assumed the population of 49 to 250 individuals. In 2004 it was proceed on the assumption that an estimated 783 individuals exist in the wild which was based on 43 discovered males. Unfortunately in 2000 a theme park with swimming pools and asphalted roads was build at the type locality Nascente do Farias and the largest part of its original habitat became destroyed. The cleared trees were replaced by banana plantations."

I guess someone somewhere has worked out a formula for estimating population based upon nest count or male bird sightings assuming those occurrences to be indicative of breeding success...

and some video of our subject from Chapada de Araripe

Adrian: Nice to meet another manakin fan!

Maggie: If you Google "Araripe," this bird's Wikipedia entry comes up as the second result (and the first if you add the "Werner" clue from my first post). Not that David did that to arrive at his ID, but you could. (I'm not familiar with this species, so I had to check a list of manakins occurring in Brazil.)

Paul, I am sure you are familiar with various field methods in avian ecology- I know that "point counts" are considered reliable, as is "nest monitoring" (i.e. checking known nests for egg count, fledglings, predation, etc.) but "nest counting" in of itself seems tenuous and I also can't fathom how a counted 43 males would correspond to an estimate of 783 individuals! Any ideas?

Or you could simply keep your ear to the ground and gain familiarity with those in the fields of ecology and conservation and happen to know Dr. Bokermann to have been one of the world's leading herpetologists, probably a fact not found on Wikipedia (as recently confirmed when discussing the inadequate and incomplete record of how many southern African species were named after Wahlberg)

Hi again, BLI states that the birds are monitored when the males are at their most vocal in September and October, just before the wet season. They nest shortly after and the young hatch when the trees are fruiting (76% of the trees are fruiting then). I have seen Manakin leks but can find no references as to if the unsuccessful males get to mate once the dominant male and female leave. If they do a count of males would equal the number of mating pairs, would it not?

counted 43 males would correspond to an estimate of 783 individuals

I can see an estimate like that being given from mark-recapture data - I've done worse. make it 800 and add reasonable error bars, and it's probably a decent estimate.

I'm only a statistician, so I can help with these estimates but I know the biologist really know their stuff.

Thanks Adrian and Bob,

I guess the increased singing may reflect probable mating success, as it does in species like our friend the Vermilion Flycatcher that sing longer songs after nest construction than before and when one takes into account the lekking behavior of manakin males, a good one can "pull" x number of females rather than we assume a one-to-one ratio... I have also heard that some manakin species, for example the White-bearded Manakin (Mancus mancus), are even more aggressive with their displays if they have already successfully mated so perhaps being "more vocal" may indicate a relatively reliable estimate...

Bob, I was about to conjure up a hypothetical "what if 50% of the males counted bred with x number of females each who produced clutches of x number of eggs with x percentage mortality" but the numbers would simply not have added up to anywhere near the estimate...

For example, this early 70s study on the Long-tailed Manakin (Chiroxiphia linearis) included a count of 39 nests, 15 of which contained eggs or young (clutch size 1 or 2) but where only 1 of 15 eggs followed hatched, and the only two nestlings observed failed to fledge...

This issue may be further confounded by some work showing that in fact it's closest related species, the Helmeted Mankin (Antilophia galeata) may in fact not even be as "promiscuous" as we presume!

Hi again, The BLI website says that they pair up, which seems to infer that they are monogamous. Whether the dominant males are polygamous or not isn't stated.

Sheri: umm, thank you I guess? (I was unsure as to the tone of your David reference, a man with whom I've had the privilege of working, whose gardens worldwide are astounding, and who I've watched inspire an audience of a thousand high-school students into uproar in support of sustainable design, so I guess you're good at what you do too...)

David: sorry, I know you like to hide and just get on with things but you know me!

Adrian: thank you for all the information! It seems incredible that we've only just found out about a bird so colorful and nowhere near as small as a miniature orchid or undescribed cockroach.

Bob: can you explain to me further (statistically!) how 43 becomes 800 in a species only discovered in 1998? I'm not arguing at all, I just want to undestand what parameters would be used- do we know the lekking behavior, or how many eggs hatch successfully (btw, does that 800 include young or does it assuming adukts?), or what the lifespan of manakins is?

Bob/Adrian, apparently this species does not in fact lek!

This is the account from Arthur Grosset who has several colleagues working out of Brazil and some of whose photos you have probably come across- he has some fantstic bird photos in his collection (I also find his trip reports quite interesting.)

Sheri: umm, thank you I guess? (I was unsure as to the tone of your David reference, a man with whom I've had the privilege of working, whose gardens worldwide are astounding, and who I've watched inspire an audience of a thousand high-school students into uproar in support of sustainable design, so I guess you're good at what you do too...)

Not sure why David being good at what he does bears on my own abilities (or deficiencies), but one of the things I happen to be reasonably good at (enough to make part of my living doing it) is demystifying bird identification. In this particular case, no savantitude was necessary to get to the correct ID, but this in no way implies that David is not a savant worthy of legions of fangirls. :)

yechhh! you presume and assume too much!

can you explain to me further (statistically!) how 43 becomes 800 in a species only discovered in 1998?

I assume 43 was the number of males observed. But if not all males were captured, the actual number would be higher.

There are several ways of estimating the total population size from samples. One way would be to capture the birds and mark them, and then re-catch them later (one or more times). From this you can estimate the capture probability. If you estimate it so that you only captured 10% of the males (say), then you would estimate there to be 430 males, and assuming a sex ratio of 1:1, a total population of 860 birds.

Thanks for a great link David,I'd been looking for a good account of behaviour.

No problem Adrian...

Bob, I understand the "math" but without a better idea of which methods were used, am somewhat baffled still... it would make sense if your recapture percentage was compared to subsequent observations so that 10% are recaptuerd but still 40-odd individuals are observed, but that then assumes the same ratio where recapture is 10% of observed therefore observed (or perhaps heard, which brings to mind the discrepancy of USFWS call-counts vs observations of Mourning Doves) equals 10% of population... and while I think it is feesible to assume 50/50 sex ratios for large and sustainable bird populations, we are so far under the sustainable population size for this species (which has been recently increased by a factor of 10 so that the extinction probability no longer uses 500 as the "tipping point" but 5,000) it still seems a tenuous estimate...

if one was to factor in the usual point count survey biases like observer efficiency or population density or in this case, given the dense vegetation, variables associated with cover, it becomes even more confounding...

Sheri (hope you got my email), I'm not sure where in AZ you are located but I'm sure you have been a part of the Tucson Bird Count- how is that configured and have you been able to get a measure of reliability over the years?

Adrian, you mentioned that BLI suggested the range for this species was 28km2 (which corresponds to the average CBC circle, right?)- was that based upon observations or a projected probable habitat hypothesis? (can't put my finger on the source, and perhaps I am confusing this info with another bird, but I understood the actual critical area was only 1km2!)

As far as the population estimates, I can guess at a couple of other ideas. I suppose, if you have marked birds, with repeated visits you could estimate the likelihood of seeing a particular individual on any particular visit. You could then extrapolate that to give the actual number that each detected bird would represent, and use that as a multiplier for the mean of the counts that you recorded. Or, you could intensively survey a small area, then extrapolate the density to a larger area of appropriate habitat. (That idea sounds pretty iffy, unless you have a really good idea of the bird's requirements, and tropical birds are often very, very picky.) I was going to say that, based on data from other manakin species, that you could come up with an estimate of what percentage of time a male spends on a lek, then use that to figure out how many males aren't being detected. But if this manakin doesn't lek, I guess that idea is out.

Thanks Paul!

Another issue here is what appears to have been a trend in over-estimatiing endangered bird species because of male bias (work by Paul Donald of the RSPB published in the Ibis, 2007:149, 671â692), particularly when there is also evidence that females tend to have a higher (and earlier) mortality...

I suppose, if you have marked birds, with repeated visits you could estimate the likelihood of seeing a particular individual on any particular visit. You could then extrapolate that to give the actual number that each detected bird would represent, and use that as a multiplier for the mean of the counts that you recorded.

That's the idea behind mark-recapture: on subsequent visits you only need to identify the bird, so it might be enough to spot them with binoculars, as long as you can identify the mark (e.g. the band colours). The basic ideas are well established, but there's still enough wrinkles in the methods to keep us statisticians occupied.

Or, you could intensively survey a small area, then extrapolate the density to a larger area of appropriate habitat.

I had briefly thought of that, but rejected it for the reasons you raise: you have no idea if the area you survey is representative.

Bob: now that I think about it, that was just a restating of classic mark-recapture methods, wasn't it? Ah, well...

The area of 28Km2 is based on a "remaining area of suitable habitat".

A little more research using the capture-recapture method in assessing habitat use, movement, and survival in four different manakin species- one of the more interesting findings was the seasonal elevational migration of all four species between old- and secondary-growth forest (which has serious conservation implications) and the authors confirmation of earlier studies questioning the validity of male data using this method when juvenile males appear similar or identical to females but are not included in the count...

and a little more specific work concerning the breeding period of Araripe Manikins inferred from vocalizations (where they also assume a sex ratio of 1:1 as demonstrated by the sister species, the Helmeted Manakin) from reseachers with Aquasis, the source upon which BLI seems to have relied...

Well, it took a bit of doing, but 39 pages of Portuguese later, this is the specific study (actually the Conservation Plan) which outlines the rationale for expanding the range and calculating the population- it is important to note that we are not talking about a somewhat uniform patch of land, but based upon the occurrence of birds along the small watersheds at about 47% of the 348 springs in the area, they have identified a strip of humid forest that contains those crucial 123 streams that is just under 60km long but only about 500m wide - i.e. our suggested 28km2 of "suitable habitat"...

as far as the population estimate goes, an auditory census was conducted in the breeding season (estimated in the research attached in comment #34) essentially identifying 135 males along 48 streams- using the 1:1 sex ratio assumed for this species they added a further 135 females- in each territory, each couple tolerates up to 4 other non-singing individuals described as "green" (i.e. juveniles) against which they calculated a 20% mortality to come up with 108 young- had they assumed that the remaining 75 streams were suitable they would have estimated a population of 968 birds but because they assessed that about 20 of the streams were probably unsuitable they arrived at a population estimate of about 800...

as you might have noticed, much to the chagrin of my students or my professors depending upon which role I have to play, I don't like unanswered questions and so pouring over research ends up being something of a passion- but had BLI supplied a little more substance to their data, I wouldn't have had to drink so much coffee!

(hope some of this helped!)

Good work David, It's a shame that BLI didn't link to any of this in English. Although there are references for further reading none appear to be on the net.