Monday, May 10, 2021

Angels of the High Seas: The Mystifying Affinities and Origins of Tropicbirds

I have never had the good fortune of seeing living tropicbirds in person, so I can only imagine that they must look ethereal. These seabirds range from pigeon- to chicken-sized, but their long, pointed wings give them wingspans of a meter or more. Both male and female tropicbirds are clad mostly in white plumage and possess a central pair of extremely long, narrow tail feathers (as long as or longer than the rest of their body), which they flaunt to each other during aerial courtship displays.

Red-billed tropicbird in its natural state, photographed by Dominic Sherony, under CC BY-SA 2.0.

True to their name, tropicbirds spend most of their lives flying over the tropical oceans of the world, far from land. Although there are other seabirds (like albatrosses and frigatebirds) that frequent the skies above the open ocean, many of these groups undertake their long pelagic journeys by soaring effortlessly on air currents. Tropicbirds, on the other hand, travel primarily by sustained flapping flight, punctuated by only brief periods of gliding and soaring. From the air, they spot their prey (mostly small fishes and squids), and capture them by plunging from the sky with a splash before taking off again.

When these elegant aeronauts need to rest, they alight on the surface of the ocean. However, despite having totipalmate feet (webbing between all the toes on each foot), their hindlimbs are so small that they are not particularly good at active swimming. Their feeble hindlimbs also render them awkward when they return to land to breed, requiring them to shuffle along the ground on their belly, sometimes with the aid of their beak and wings.

Red-billed tropicbird resting on the ocean's surface, photographed by Steven Mlodinow, under CC BY-NC 4.0.

Fortunately, tropicbirds nest on remote islands that are typically safe from land predators. They often adopt crevices in rocky ledges as nest sites, though flat ground or forks and hollows in trees may also be used. Tropicbirds are probably at their least angelic on their breeding grounds, as competition over suitable nest sites can be extremely fierce, and many adult tropicbirds bear scars on their head as a result of such disputes.

Three species of extant tropicbird are currently recognized. In order of increasing size, they are the white-tailed tropicbird (Phaethon lepturus), the red-billed tropicbird (Phaethon aethereus), and the red-tailed tropicbird (Phaethon rubricauda). Although they are low in diversity and restricted to equatorial regions today, the fossil record reveals an ancient history and more widespread distribution for total-group tropicbirds in the past.

What may be the oldest known stem-tropicbird fossils were described by Mayr and Scofield (2016), these being a partial humerus (upper arm bone) and carpometacarpus (fused wrist and palm bones) from the Paleocene Waipara Greensand of New Zealand (about 60.5–61.6 million years old), similar in age to some of the oldest known stem-penguins. A potentially even older fossil bird that has been likened to stem-tropicbirds is Novacaesareala, known from fragmentary forelimb bones found in the Hornerstown Formation of the eastern United States (Mayr and Scofield, 2016). The age of this formation is disputed (it may date to either the Late Cretaceous or early Paleocene), but in any case the available specimens of Novacaesareala are probably too limited for a well-founded identification (Mayr, 2017).

Partial humerus of a possible stem-tropicbird from the Waipara Greensand (A–D) compared to the humerus of a white-tailed tropicbird (H–K), from Mayr and Scofield (2016).

There are, however, much better-represented stem-tropicbird taxa that are only slightly younger (by geologic standards) than these records, namely Lithoptila from the late Paleocene–early Eocene of Morocco and Prophaethon from the early Eocene of England. (In fact, Lithoptila is the oldest crown bird known from Africa.) The placement of both these genera as stem-tropicbirds has been supported by phylogenetic analyses, though results differ on whether they form a clade with one another that excludes crown tropicbirds (Bourdon et al., 2005; Smith, 2010).

Lithoptila and Prophaethon were around the same size as extant tropicbirds, but had a longer pelvis and longer feet, suggesting that they were more capable swimmers. Although they were probably strong fliers, details of their wing skeleton indicate that they may not have been aerial specialists the way modern tropicbirds are. It has been proposed that Prophaethon lacked the elongate tail feathers of its closest living relatives, based on the shorter transverse processes (sideways projections) on its tail vertebrae (Mayr, 2015). The abundance of remains known from Lithoptila and Prophaethon suggest that they approached continental shorelines more frequently than extant tropicbirds do, maybe even forming large breeding colonies in these habitats (Bourdon et al., 2008a; Mayr, 2015). Taken together, these lines of evidence paint these stem-tropicbirds as having been less pelagic than modern tropicbirds, but potentially having spent more of their time in the water. Mayr (2015) speculated that they caught prey while swimming on the water's surface, similar to albatrosses. The aquatic inclinations of stem-tropicbirds may provide a plausible explanation for why extant tropicbirds retain totipalmate feet.

The foot of Prophaethon (a) compared to that of a white-tailed tropicbird (b), from Mayr (2015). The scale bars = 5 mm; as can be seen, Prophaethon had a much longer foot than modern tropicbirds, despite being similar in body size.

Another early stem-tropicbird was Zhylgaia, known from partial humeri dating to the late Paleocene of Kazakhstan that were originally mistaken for those of a shorebird (Bourdon et al., 2008a). Mayr and Scofield (2016) further noted that a tarsometatarsus (fused ankle and foot bones) described as "Tshulia" from the same time and place probably belongs to Zhylgaia. Other possible stem-tropicbird specimens similar to Lithoptila and Prophaethon have been reported from the late Paleocene of the eastern United States (Olson, 1994) and the middle Eocene of Belgium (Mayr and Smith, 2002). Additionally, the supposed ibis "Proplegadis", known from a fragmentary humerus from the Eocene of England, is likely a specimen of Prophaethon (Mayr, 2015).

Possibly more closely related to extant tropicbirds was Phaethusavis from the early Eocene of Morocco, based on a partial humerus (Bourdon et al., 2008b). After the Eocene, the completeness of the tropicbird fossil record appears to drop off sharply, perhaps corresponding to the onset of more pelagic tendencies. Nonetheless, there are a few younger fossil tropicbirds that have been described from regions where tropicbirds no longer occur, such as Heliadornis ashbyi from the middle Miocene of Belgium and the eastern United States (Olson, 1985; Olson and Walker, 1997), as well as possibly Heliadornis paratethydicus from the late Miocene of Austria (Mlíkovský, 1997) and Heliadornis minor from the late Pliocene of Slovakia (Kessler, 2009).

Why tropicbirds vanished from higher latitudes is unknown. Mayr (2015) suggested that competition and predation from other organisms may have driven them to specialize in living off nutrient-poor tropical waters, where food sources are relatively patchy and scarce. The Neogene disappearance of tropicbirds from high latitudes and their current restriction to the tropics bring to mind similar biogeographic histories in some terrestrial bird groups (such as mousebirds), which have been attributed to global climate change. It makes me wonder whether tropicbirds represent a marine example of this recurrent pattern.

One of the most longstanding mysteries surrounding the evolutionary history of tropicbirds concerns their phylogenetic position. Traditionally, tropicbirds were grouped together as "pelecaniforms" alongside pelicans, frigatebirds, boobies, cormorants, and anhingas, with which they share totipalmate feet. Tropicbirds have long been recognized as the "odd ones out" in this collective, however. For example, they lack a bare throat pouch, a long hallux (innermost toe), or reduced external nostrils, and have a covering of downy feathers upon hatching (as opposed to being naked). Some researchers instead argued for a closer relationship between tropicbirds and procellariiform birds (albatrosses, petrels, and kin), which was supported by several phylogenetic analyses based on morphological characteristics (Mayr, 2003; Bourdon et al., 2005; Smith, 2010).

Red-tailed tropicbird chick, covered in fluffy down, photographed by Forest & Kim Starr, under CC BY 3.0.

The rise of large-scale molecular phylogenetic analyses has greatly clarified the relationships among modern bird groups, and one of their major findings has led to the breakup of the traditional "pelecaniforms". It turns out that though frigatebirds, boobies, cormorants, and anhingas do form a clade (now known as Suliformes), pelicans are more closely related to herons, ibises, shoebills, and hamerkops, with the name Pelecaniformes now applied to this latter assemblage. Both suliforms and pelecaniforms, however, belong to a large group of mostly aquatic birds called Aequornithes, which also contains storks, procellariiforms, penguins, and loons.

A surprising commonality among the results of these analyses was that tropicbirds were not found as members of Aequornithes, meaning that they are not especially closely related to pelecaniforms, suliforms, or procellariiforms. However, early molecular phylogenetic analyses found little consistent support for what tropicbirds actually are closely related to. It was only more recently with the advent of increasingly large datasets that a more widely corroborated answer to that question has emerged, and it is an unexpected one: the closest living relatives of tropicbirds are likely the eurypygiforms.

... Well, it might be hard to find that shocking if one is not familiar with eurypygiforms. They are a group of birds that includes just two living species: the heron-like sunbittern (Eurypyga helias) from the Neotropics and the nearly flightless kagu (Rhynochetos jubatus) from New Caledonia, both of which exhibit little if any obvious anatomical similarity with tropicbirds. Furthermore, though the sunbittern regularly forages for prey in freshwater environments, neither species shows any inclination towards a marine lifestyle.

A sunbittern, which does not look or behave a whole lot like a tropicbird, photographed by Brent Moore, under CC BY 2.0.

Yet time and time again, a close relationship between tropicbirds and eurypygiforms (first recovered by McCormack et al., 2013) has been supported by recent large-scale analyses of genetic data (Jarvis et al., 2014; Prum et al., 2015; Reddy et al., 2017; Kuhl et al., 2021). Most (though not all) of these studies also place the eurypygiform–tropicbird clade (which has been called Phaethontimorphae) as most closely related to Aequornithes, forming an expanded "waterbird clade". (At present, no formal name has been assigned to the group uniting phaethontimorphs and aequornitheans in technical literature, though John Boyd has adopted the old name Ardeae for this clade on his Taxonomy in Flux website. Personally, I think another old name, Natatores, would also be a reasonable option, as it has been applied to a very similar assemblage of birds in at least one 21st Century paper.)

The current consensus of phylogenetic relationships in the expanded "waterbird clade", plotted against geologic time.

The eurypygiform–tropicbird relationship has not gone ignored in paleontological studies. Despite this, essentially no anatomical features that clearly characterize this clade have been identified, and even the existence of well-represented stem-tropicbird fossils has so far been of limited help in this regard (Mayr, 2014; Mayr, 2019). If anything, the aquatic adaptations of stem-tropicbirds seem to render the the close relationship with the terrestrial eurypygiforms even more baffling. Prophaethon does share with eurypygiforms long, slit-like nostril openings in the skull (which are retained in extant tropicbirds only as juveniles), but this feature is also found in most aequornitheans (Mayr, 2014). 

Skulls of Prophaethon (A), a red-tailed tropicbird (B), a juvenile brown pelican (C,  Pelecanus occidentalis), and a sunbittern (D), from Mayr (2014). Arrows indicate the extent of bony nostril openings.

Perhaps the biggest missing piece of the puzzle though is the virtual absence of known fossil eurypygiform specimens. There is a fossil from the Eocene Green River Formation of the western United States that has been proposed to be a stem-sunbittern; however, it has not yet been formally described. A group of Paleogene birds called the messelornithids were originally suggested to have been eurypygiforms, but all recent studies on their phylogenetic affinities instead find that they were probably gruiform birds, more closely related to cranes and rails.

So maybe the discovery of unambiguous stem-eurypygiforms will further clarify the evolutionary connection between eurypygiforms and tropicbirds. Such fossils may also have the potential to shed light on other aspects of tropicbird evolution, like whether their marine ecology was inherited from the last common ancestor of the expanded "waterbird clade" or arose independently in total-group tropicbirds and aequornitheans.

For now, there is clearly much we have yet to learn about tropicbirds. None of the three extant species are considered to be under immediate threat of extinction, though many of their breeding grounds are at risk from introduced predators, and their pelagic habits make their population trends difficult to monitor. However, if effective conservation measures are established and maintained, it is likely that we will have the opportunity to marvel at these angels of the high seas long into the foreseeable future.

White-tailed tropicbird in flight, photographed by hokoonwong, under CC BY-NC 4.0. Fly on, mysterious angel.