The Evolution of Poisonous Birds

tags: , , , , , , , , ,

The Hooded Pitohui, Pitohui dichrous,
endemic to New Guinea, is very unusual
because it has poisonous plumage and skin.

Image: John Dumbacher.

I have been in love with New Guinea since I first read about it as a kid. Everything about this tropical island is exotic and fascinating to me, from the large numbers of endemic bird and plant species to the tremendous number of spoken languages -- more than anywhere else on the planet. So I was immediately interested to learn about Jack Dumbacher's adventures there between 1989 and 1991. At the time of his first visit, he was a grad student in ornithology who was catching birds of paradise as part of a National Geographic Society expedition -- what I wouldn't have given to be part of that! As the story goes, Dumbacher removed several fiesty orange-and-black birds that had become accidentally entangled in his mist nets when he stopped to lick the wounds on his hands. Shockingly, his lips and mouth became numb: he had been poisoned.

"I was scared and I tried not to swallow," he recalled. "I figured I had probably brushed up against some poisonous tree."

Fortunately, his symptoms subsided after four hours. However, roughly two weeks later, a colleague also experienced these same symptoms after handling this same bird species. That was when Dumbacher began to suspect that the source for the poison was not the vegetation at all, rather, it was the bird that was poisonous.

One year later, Dumbacher returned to New Guinea to verify his suspicion.

"I performed this very complicated and sophisticated experiment. I clipped off some feathers and popped them in my mouth" -- taking care not to swallow. His lips, mouth and tongue became numb almost immediately. Clearly it was the pitohui itself -- not some plant or tree -- that was the culprit.

He spoke to the local tribespeople about this species and learned that they consider the bird to be taboo. Dumbacher recalled that some reported that the pitohui was "good for nothing, a rubbish bird."

Dumbacher received permission to bring four hooded pitohuis, Pitohui dichrous, back to the United States for further study. He and John Daly, a chemist at the National Institutes of Health in Bethesda, Md., isolated and identified the poison, which turned out to be one of the Batrachotoxins (pictured, right). Batrachotoxins are extremely potent cardio- and neurotoxic steroidal alkaloids found in South American poison dart frogs and Melyridae beetles -- and remarkably, in the skin and feathers of five of the six pitohui species of New Guinea. Astonishingly, another avian species, Ifrita kowaldi, was later discovered to have batrachotoxins in its plumage, too.

Given this suite of poisonous birds, Dumbacher wondered whether all poisonous birds closely related to each other. Based on morphological characters, it was thought they were closely related, but to learn more about the evolutionary history of this group of birds, it was necessary to reconstruct their relationships based on their DNA.

So Dumbacher and a team of molecular biologists sequenced the DNA from 55 museum specimens comprising three major polytypic Pachycephalidae genera (Colluricincla, Pitohui, Pachycephala) and most of the monotypic Pachycephalidae genera (Rhagologus, Aleadryas, Eulacestoma, Falcunclus, Oreoica). They analyzed these DNA sequences and recovered this phylogenetic tree (figure 1);

Fig. 1. Most likely phylogenetic tree, based upon PAUP likelihood analysis using a 15-partition model of evolution. Bayesian posterior probabilities from MrBayes 3.1.1 were multiplied by 100 and appear above each node; parsimony bootstrap percentages appear below the node. Taxa currently classified in the genus Pitohui are noted in bold.
DOI: 10.1016/j.ympev.2008.09.018.

As you can see from this phylogeny, the six Pitohui species do not form a monophyletic group. Instead, Dumbacher and his team found that P. nigrescens is allied with the whistlers (genus Pachycephala); P. ferrugineus and P. incertus are allied with the shrike-thrushes (genus Colluricincla). Pitohui cristatus is the closet relative -- "sister" -- to Aleadryas, and these two species are sister to the Crested Bellbird, Oreoica gutturalis, of Australia, while the two remaining Pitohui species (P. dichrous and P. kirhocephalus) form a distinct clade. Based on these data, Dumbacher and his team are currently revising the genus Pitohui and reviewing the application of taxonomic names.

This phylogeny also shows that the two most toxic Pitohui species, the hooded pitohui, P. dichrous, and the variable pitohui, P. kirhocephalus -- which also happen to be the most conspicuously colored -- form a monophyletic group with strong phylogenetic support.

Another finding is that the Morningbird of Palau, Pitohui tenebrosus, is actually a whistler, and according to the above phylogeny, this species may be closely related to the widely distributed Pachycephala pectoralis. Interestingly, these two sister species do not resemble each other physically, which provides more support for the new concept of ''islands as engines" of evolution, as first proposed in 2005 by two of my postdoctoral colleagues at the AMNH, Rob Moyle and Chris Fillardi [read more about it here].

To answer his second question regarding the evolution of toxicity in these birds, Dumbacher's team constructed a second phylogeny that maps which species have batrachotoxins in their skin and plumage (figure 2);

Fig. 2. Summary of BayesTraits analysis of toxin evolution. This partial cladogram roughly depicts the mapping of toxicity onto the Pitohui phylogeny. The pie diagrams associated with key nodes denote the relative support for a toxic (black) or non-toxic (white) ancestor.
DOI: 10.1016/j.ympev.2008.09.018.

These data reveal that toxicity is not ancestral: instead, it has evolved several times in this particular group of birds, and it has evolved convergently in five of the six Pitohui species -- the single exception is the White-bellied Pitohui, P. incertus, which lacks batrachotoxins in its skin and plumage.

Interestingly, the five toxic Pitohui species exhibit dramatic behaviors that likely are the result of their toxicity. They are gregarious and vocal, and they are the leaders of large mixed flocks of birds as they forage in the forest understory. Their closest relatives, which are not toxic, do not share any of these behaviors as to their unpalatability.

The toxic Pitohui species are also deceptively similar in size, morphology and ecology -- so much so that they were originally classified into one genus. However, the DNA data suggest that these birds evolved convergently to resemble each other on the basis of one shared character; toxicity. This independent evolution of morphological and behavioral similarities among these birds may function as Müllerian mimicry, where multiple toxic species resemble each other and thus, share the cost of "educating" predators.

This research elegantly demonstrates that the evolution of just one character -- in this case, toxicity -- can profoundly affect the evolution of a suite of other characters, ranging from body size and behavioral traits to ecological niche.


J Dumbacher, K Deiner, L Thompson, R Fleischer (2008). Phylogeny of the avian genus Pitohui and the evolution of toxicity in birds. Molecular Phylogenetics and Evolution, 49 (3), 774-781 DOI: 10.1016/j.ympev.2008.09.018.

More like this

The fact that disparate species are all synthesizing the same toxin tells me that it's a pretty minor genetic switch change to be able to do so. What usual bodily secretion is being altered to form this toxin? Are they being formed in a skin gland?

Convergence is easily the coolest effect (and one of the best proofs) of natural selection.

I wonder if they've raised any in captivity; and if so, whether or not they're still poisonous. If so, this is truly fascinating.

Yes, this is fascinating.

Another thought just came to mind. I don't have my New Guinea field guide handy. It's got to be some accipiter driving this convergence, yes? Or something just like one?

What usual bodily secretion is being altered to form this toxin? Are they being formed in a skin gland?

Apparently the birds don't make the toxins themselves, and neither do the poison-dart frogs -- both just happen to eat melyrid beetles.

By David MarjanoviÄ (not verified) on 05 Dec 2008 #permalink

"Apparently the birds don't make the toxins themselves, and neither do the poison-dart frogs -- both just happen to eat melyrid beetles."

I suspected some sort of explanation like that when I looked at the phylogenetic tree and then read about the similarity in their ecological niche's. (Convegenent evolution in 5 groups of birds yeilding the synthesis of a specific toxin just seems too implausible)

Just wondering what affect this might have on parasites.

4 & 5:

You would think that (most) predator species that eat the frogs/beetles would find it toxic. It seems to me that unless the larger phylogenetic group is able to tolerate this toxin, then this observation only shift the evolutionary convergence from generating the toxin, to being able to tolerate it and/or excrete it via the skin. (This may also not be true convergence, if they use different means of tolerating or excreting this toxin.)

(Taking nothing away from your observation: seems a good catch to me, assuming you're right--this is way off my patch so I can't verify it!)

By BioinfoTools (not verified) on 06 Dec 2008 #permalink

It seems to me that unless the larger phylogenetic group is able to tolerate this toxin, then this observation only shift the evolutionary convergence from generating the toxin, to being able to tolerate it and/or excrete it via the skin.

That's of course correct, and remarkable enough, but much less so than producing not just any alkaloid but the exact same hemibatrachotoxin as a poison-dart frog would be.

By David MarjanoviÄ (not verified) on 06 Dec 2008 #permalink

Oh! Oh! Congrats on the Fark! posting! I read here a lot, but I'm stoked to see you posted there!

The paper implies evolved toxicity, as opposed to aquired toxicity, though I agree with posters that this is the simpler answer to this extraordinary example of convergence.

Do these poisonous beetles exist on this "island"? If so, question settled. Exuding toxins from the skin seems an ordinary response. If not, this toxic compound must be a simple genetic jump from some other benign biologic substance.

Where does eating distasteful, poisonous beetles begin the walk down this evolutionary path? It's a nasty toxin, how does this begin? Something along the lines of eating the toxic beetle makes the individual wobbly and more easily predated, but taste so bad the predator swears off?

Accipiters have to be involved in this.


I agree with your point and realised that as I wrote it, but I wanted to try keep my post compact and not waffle! :-)

I have other thoughts on this, too. One more ;-) That the toxin is excreted by via the skin, and assuming the toxin excreted via the skin is exactly the same toxin as the birds ingest from eating the beetles, indicates it is absorbed into the body unmodified, and is not processed by gut microbes. The latter could be one possibility that might work around the evolutionary convergence issue in a variety of ways.

Now, I just expect someone to pop up and tell me that this toxin *is* processed by microbes and that this might (in part) counter for the evolutionary convergence issue! :-)

By BioinfoTools (not verified) on 06 Dec 2008 #permalink

On Friday evening, after dinner, I was musing out loud that I knew of countless poisonous insects, fish, molluscs, et cetera; that I knew of poisonous reptiles, amphibians and the venom of the platypus; that I was aware of toxins secreted by unicellular organisms and of the co-opting of cnidarian weaponry by Aeolidioidea, but that I did not know of any poisonous birds (and that I found this a bit strange).
Just two days later I'm sitting here browsing through ScienceBlogs -- just wandering around, clicking on whatever link headers sound interesting (as one does) -- when I come across this blog entry (written on the very same day as I was wondering about it all, how's that for serendipity) and it kicks me off on a couple of hours of investigating what people know or hypothesise about it all. Another little hole in my knowledge of the world filled in. Another pleasant few hours of learning, and of marvelling at the way the encyclopaedia of scientific knowledge continues to expand, pushing back the boundaries of what we know and, crucially, making that progress available for the world to know, if it chooses to.
Thank you, GrrlScientist, and thank you, ScienceBlogs. I don't know if there is any way to express my gratitude to all of you bloggers at once but this is what I always hoped that the Internet would be, and I thank you for it. Don't stop researching, and don't stop telling us about it. I am not able to donate anything other than my appreciation, but you have that, and daily.

By FreestyleScientist (not verified) on 07 Dec 2008 #permalink

"Eh? Are they the ones who run the Illuminati now?"

No, accipiters act quickly and near thoughlessly. They are all talon-jerk response. See it, killit.

The Illuminati think long-term, they plan ahead. Like... What's for breakfast tomorrow?

Do these poisonous beetles exist on this "island"?


By David Marjanović (not verified) on 08 Dec 2008 #permalink

OK, I am forced to agree the toxicity is aquired.

Thus: A diverse array of creatures aquire toxicity from the ingestion of melyrid beetles, with fascinating evolutionary effects.

This raises a number of questions.

Are there co-evolutes wherever these beetles are found?

What is the chemistry of the neuro-tolerance and the exuding of this toxin through the skin? Is it the same across species?

How does the warning coloration/appearance/behavior correlate with the level of toxicity in different places under different predatory pressures?

Hypothesis: Map the melyrid beetles, and more toxic co-evolutes will be found.