Saturday, August 31, 2019

Shining a Light on Nightbird Evolution: My First First-author Paper!

As I previously mentioned on this blog, I'd been hoping to have the first part of my PhD research submitted to a journal by the time I had reason to blog about it again. I'm pleased to report that my research has now not only been submitted, but published! Diversity (the journal I submitted to) processed the article unbelievably quickly, having it reviewed, edited, accepted, and published in less than a month! I was also fortunate in that the reviewers didn't request any major changes. I certainly don't expect to go through such a painless submission experience again anytime soon.

For this study, my coauthors and I looked at the phylogenetic relationships of a remarkable group of theropods, Strisores. Many strisoreans* are well-camouflaged birds that are active at night or twilight; these include the nightjars, oilbirds, potoos, frogmouths, and owlet-nightjars. However, the diurnal swifts and hummingbirds are also members of Strisores. Most strisoreans (including the predominantly nectar-feeding hummingbirds) eat insects, but the oilbird (Steatornis caripensis) feeds exclusively on fruit.

*I used this new paper as an opportunity to make the case that "strisorean" should be the vernacular form for Strisores, for much the same reasons that I now use "enantiornithean" instead of "enantiornithine". Time will tell whether anyone else follows this...

A cartoon depiction of the major strisorean subgroups and their inferred phylogenetic relationships based on our new study.

Traditionally, the nocturnal strisoreans have been classified as one group, but recent studies have presented strong evidence that the owlet-nightjars are more closely related to swifts and hummingbirds than to the rest. (The group uniting owlet-nightjars, swifts, and hummingbirds has been named Daedalornithes.) When it comes to the phylogenetic relationships among the remaining groups, however, little consensus exists. In fact, up until recently, no two phylogenetic datasets aimed at resolving their relationships found the exact same results!

We approached this problem by combining the largest genetic and anatomical datasets that have been assembled for Strisores so far. The genetic dataset was originally put together for a different study by my coauthors Noor White and Mike Braun (accepted at Molecular Phylogenetics and Evolution but not yet published online at the time of writing) and the anatomical dataset came from a 2013 study by Dan Ksepka and colleagues.

The results of selected previous studies on strisorean phylogeny. Up until White and Braun (2019) found an identical topology to Prum et al. (2015), no two datasets produced the same result!

When we analyzed our combined dataset, we found that nightjars were best supported as the most distantly related group to other living strisoreans. The oilbird and potoos were united in one group, which was in turn closely related to a clade containing the frogmouths and daedalornitheans. This result was not only identical to what we found when we analyzed our genetic dataset on its own, but also to the findings of a previous genetic study by Richard Prum and colleagues. Although it is never wise to unilaterally declare a case closed in science, the fact that two large datasets independently recovered the same results suggests to me that this is indeed the most likely phylogenetic tree for Strisores.

In fact, we felt that the support for a group including all strisoreans except nightjars was strong enough that we chose to give it a name: Vanescaves. This name translates to "vanishing birds", partly a nod to the Emily Dickinson poem "A Route of Evanescence", which describes a hummingbird flying near some flowers. The name also references the fact that many vanescavian subgroups currently have geographically restricted ranges (oilbirds and potoos in the Neotropics, frogmouths in Australia and Southeast Asia, and hummingbirds in the Americas), but are known from the fossil record to have once lived in other regions, such as Europe. In contrast, nightjars are distributed almost globally today, but have very little of a documented fossil record.

One of the resulting phylogenetic trees we recovered in our study, scaled to geologic time. (The divergence times are largely bare minima necessary to accommodate known fossil ages and should not be taken literally.)

Speaking of fossils, one of the main benefits of combining genetic and anatomical data in our study was that it allowed us to place fossil species in the context of our phylogenetic results. Despite their small body size and delicate, sometimes literally paper-thin bones, a diverse range of fossil strisoreans have been identified in Eocene fossil deposits (33.9-56 million years old). In general, most fossil strisoreans included in our study fell out in parts of the tree that we expected them to based on previous research, but we did find a few surprises.

For example, we found that the oilbird may be the closest living relative to Fluvioviridavis, a 52-million-year-old strisorean from Wyoming, unlike previous phylogenetic analyses which found Fluvioviridavis as a close relative of frogmouths. This is a notable result given that paleornithologist Gerald Mayr previously noted features in Fluvioviridavis that are more similar to oilbirds than to frogmouths.

The holotype of Fluvioviridavis, from Nesbitt et al. (2011), under CC BY 2.5.

We also discovered that Hassiavis, a 47-million-year-old strisorean from Germany that had not been previously subjected to phylogenetic analysis, was potentially an early member of the owlet-nightjar lineage, which would make it the oldest known stem-owlet-nightjar and the first one known from outside of Australasia, identifying owlet-nightjars as yet another vanescavian group with a formerly much broader distribution. (However, it should be noted that not all of our analyses recovered Hassiavis as a stem-owlet-nightjar. In any case, we found that it was most likely a daedalornithean.)

Our research additionally allowed us to make new inferences about the evolution of strisorean anatomy. Previous studies that considered only anatomical data tended to recover nightjars, potoos, and daedalornitheans as a group. These strisoreans are often specialized for snapping up insects in flight, in contrast to the fruit-eating oilbird and the big-beaked frogmouths (which instead more commonly pounce on prey on the ground). According to previous morphology-based hypotheses, this would imply that aerial insectivory originated relatively late in strisorean evolution. However, the phylogeny we found has the oilbird and frogmouths nested among the insect-hawking groups, suggesting that they descended from ancestors that similarly hunted insects on the wing.

Nightjars and other insect-hawking strisoreans have such specialized-looking anatomy that it may seem counterintuitive that they represent the ancestral state for Strisores. However, there may be a parallel example in other flying vertebrates: bats likely also started out as aerial insectivores, and they too evolved into fruit-eaters, vertebrate predators, and nectar-feeders. Furthermore, our phylogenetic results really only imply two losses of aerial insectivory (once in the oilbird and once in frogmouths), which does not come across as an unbelievably high number to me.

The skull of a common potoo (Nyctibius griseus), viewed from the right. The giant eyes and broad palate, among other things, make this a very bizarre skull!

Is there any support for this in the fossil record? Maybe! We found that Protocypselomorphus, a small strisorean from the Eocene of Germany that was likely an aerial insectivore, may have been more closely related to the oilbird than to any other living strisorean. If this is correct, it would provide evidence that the oilbird had insect-hawking ancestors. Indeed, the often-prescient Gerald Mayr had already pointed out that Protocypselomorphus shares certain features in common with the oilbird. That being said, the majority of fossil strisoreans in our study were already very anatomically similar to their closest living relatives, so a clearer picture of what the ancestral strisorean looked like may depend on the discovery of even older fossils.

One question I've frequently received when discussing this research is whether hummingbirds and swifts evolved from nocturnal ancestors. We did not focus on this interesting topic for this study, but I personally think that we do not yet have enough information to answer this question conclusively. If one assumes that gaining and losing nocturnality are equally likely, it's true that the most straightforward interpretation is that strisoreans were ancestrally nocturnal and then became diurnal on the line leading to hummingbirds and swifts. However, it is not clear that such an assumption is correct. Furthermore, there is evidence that different groups of nocturnal strisoreans have adapted to darkness in different ways: the eyes of nightjars and potoos have a reflective layer that helps them capture more light at night, whereas other strisoreans appear to lack this feature. Thus, the notion that nocturnality originated several times in strisoreans may also be plausible.

If you've seen me present this study at conferences, you may remember that I also intended to perform divergence time estimation on Strisores. We ultimately decided to forgo that part of the study, because the massive size of our genetic dataset made divergence time estimation very computationally expensive and time consuming. However, I may still attempt such an analysis on a smaller dataset for a future manuscript. Stay tuned...

Reference: Chen, A., N.D. White, R.B.J. Benson, M.J. Braun, and D.J. Field. 2019. Total-evidence framework reveals complex morphological evolution in nightbirds (Strisores). Diversity 11: 143. doi: 10.3390/d11090143

3 comments:

  1. Just want to thank you, once again, for taking the trouble to write this in a way that is easily understood by non-academics. I really appreciate it and will look forward to reading it tomorrow.

    ReplyDelete
  2. Given how plastic extinct bird groups are in relation to modern species I don't think an ancestral niche for any given clade can be accurately reproduced. For example, that there are herbivorous Cariamiformes throws a dent on the supposed ancestral carnivorous state for the clade.

    ReplyDelete
    Replies
    1. Or those herbivorous "cariamiforms" were not really cariamiforms. As Gerald Mayr has pointed out, taxa like Strigogyps have mainly been referred to Cariamiformes based on overall similarity and have never been included in a sufficiently comprehensive phylogenetic analysis that can test that. And even if they were cariamiforms, it's at least as plausible that herbivory was specific to Strigogyps instead of being ancestral to Cariamiformes. In any event, it's a pretty big extrapolation to conclude that ancestral ecology cannot be inferred in any bird clades just based on that one example.

      Given that the oilbird and frogmouths are the only strisoreans that don't practice aerial insectivory, I would be pretty surprised if the behavior was not ancestral (assuming that the phylogeny we recovered is correct), especially if Protocypselomorphus does turn out to be a stem-oilbird.

      Delete