Thursday 16 August 2018

Those terrific pelagornithids

Miocene pelagornithid Pelagornis chilensis parents and chicks, what fantastic animals they must have been. Note the lack of pseudoteeth on the chicks, recent work suggests they didn't develop until the cusp of adulthood. Not that I drew them in and removed them at the last minute when researching this post, of course.
Given how often I've written about giant pterosaurs, it's peculiar that I've never thought to cover the only flying animals which have ever come close to challenging their size, the pelagornithids: long-winged, often gigantic birds which attained wingspans exceeding 6 m. And yes, a 6 m wingspan is a metric that many pterosaurs - not even just the big azhdarchids - would find endearingly cute, but it's the largest wingspread of any bird and well above the size of any living flying animal. In public engagement pelagornithids are mostly wheeled out to gawk at their size and weird pseudoteeth before being put away again, but there's lots of fascinating anatomy under those big wings, and they deserve a longer period in the spotlight.

Pelagornithids - which are sometimes called (the now defunct name) "pseudodontorns" - were a long-lived and globally distributed group, their fossils ranging from Palaeocene - Pliocene rocks of Eurasia, both Americas, Africa, Antarctica and New Zealand (Mayr and Rubilar-Rogers 2010; Bourdon and Cappetta 2012). They were a group of large-bodied pelagic soarers, seemingly adapted for extended periods of flight over seas and oceans. Most of their fossils are - as is typical for birds - pretty fragmentary, but a number of species are are relatively well represented, especially members of the genus Pelagornis. Their soft-tissue anatomy is virtually unknown, save for primary wing feather impressions associated with the holotype of P. orri (Howard 1957).

Many pelagornithids are known from single bones or a few pieces of skeletal shrapnel. In having good cranial and limb material, the Oligocene taxon Pelagornis sandersi is among the better known species. Note the difference in size of the humerus (e-f) vs. the hindlimb bones (j-q; femur is j-k, tibiotarsus is l-m, and carpometatatsus is o-p). From Ksepka (2014).
Much uncertainty and confusion surrounds the composition of pelagornithid taxa with numerous genera being considered invalid or synonymous with others. This problem is rooted in over-enthusiastic naming of undiagnostic, often crushed scraps of bone as well as a lack of comparable anatomies between holotypes. Evaluation of Neogene pelagornithid material has suggested that most genera are weakly supported, leading Mayr and Rubilar-Rogers (2010) to suggest that all Miocene and Pliocene taxa (which includes Osteodontornis, Pseudodontornis, Neodontornis and possibly Cyphornis) should be sunk into Pelagornis. It's a little difficult to say how speciose Pelagornithidae is given the fluidity of their taxonomy but, as a rough figure, it seems to be composed of a little over a half-dozen Neogene Pelagornis species and a handful of Palaeogene representatives. All pelagornithids share the same basic "long-winged and pseudotoothed" bauplan, but the characteristic anatomies and proportions of the group are most expressed in later taxa, such as Pelagornis.

The confusion over pelagornithid systematics is not confined to generic relationships. Their placement among other birds has been the source of much discussion and controversy, and it's perhaps best to regard their affinities as currently uncertain. Initially regarded as possible relatives of Pelecaniformes (classically thought to contain pelicans, cormorants, gannets and so on - the situation has changed since then), Procellariiformes (tube-nosed birds, including albatrosses) or Ciconiiformes (storks and allies), Bourdon (2005) found stronger evidence linking pelagornithids with Anseriformes - the same group that includes ducks, geese and screamers. Numerous features of the skull and forelimb support this affinity, as do some features of skull development (Louchart et al. 2013). An affinity with waterfowl might seem bizarre for these ocean-going giants but Anseriformes have a long and varied evolutionary history: this is the same branch of avian evolution that (probably) begat the giant, flightless gastornithids and mihirungs, as well as the wader-like Presbyornis. Viewed from a geological perspective instead of a modern one, Anseriformes are not just birds that honk and quack.

But while an anseriform affinity for pelagornithids has not being dismissed out of hand, the idea is not without critics. Some pelagornithid anatomies - such as their sterna - are not anseriform like (Mayr et al. 2008), and other features imply a position outside the anseriform-galliform clade (Galloanserae: crudely, the duck-chicken group) without qualifying for entry into Neoaves (all living birds but ratites and galloanserans) (Mayr and Rubilar-Rogers 2010; Mayr 2011). So while these studies broadly agree that pelagornithids emerged from fairly rootward stock among Aves, and that they are not closely related to any modern soaring birds, further work, and maybe more fossils, are needed to clarify their actual phylogenetic position.

Pelagornithid primary wing feather impressions associated with the holotype of Pelagornis orri. It's not known is these represent the longest feathers of the wing, but they still have a useful role to play in reconstructing pelagornithid wingspans. From Howard (1957).

Size-off: Pelagornithids vs. Argentatvis

All pelagornithids are characterised by large size with even the earliest, smallest taxa being comparable to big albatrosses in wingspan (Bourdon 2005). But how big did they get? I know several readers are already sharpening their comment knives about my introduction suggesting that pelagornithids are avian wingspan record holders, thinking I've forgotten about the giant, 7 m wingspan Miocene teratorn Argentavis magnificens. But that's not a mistake: pelagornithids really should be considered the record holders for avian wingspans, and Argentavis isn't as big as most people imagine.

Classic image of teratorn researcher Kenneth E. Campbell posing with a 25ft wingspan (7.62 m) silhouette model of Argentavis magnificens at the National History Museum of Los Angeles. Alas, Argentavis wasn't quite as big as depicted here. From Campbell (1980).
Some giant pelagornithids, such as Pelagornis chilensis, are unusual among giant fossil fliers in being represented by relatively good skeletal material and their feathered wingspan estimates of 6 m or more can be considered trustworthy, reliable figures. Giant Argentavis magnificens, on the other hand, are known from fragmentary remains and some degree of uncertainty surrounds their wingspans: estimates have ranged from 5.7 to 8.3 m (e.g. Campbell and Tonni 1983; Chatterjee et al. 2007). Recent workers have suggested that the lower range of these estimates is more likely. When describing P. chilensis, Mayr and Rubilar-Rogers (2010) noted that the 82 cm long humerus of their pelagornithid was vastly bigger than the 57 cm long Argentavis humerus, and that scaling the latter to proportions seen in smaller teratorns yields a wing skeleton length of 183 cm. If so, the bony wing spread of the largest Argentavis might have struggled to reach 4 m, and this is before assuming any flex in the wing bone joints. And no, the addition of feathers does not bring Argentavis into record-breaking territory. Ksepka (2014) predicted that the primary feathers of Argentavis would need to be about 1.5 m long to reach a 7-8 m wingspan, a length that would exceed the primary feather: wingspan ratio of all living birds as well as contradict the observation that primaries tend to scale with negative allometry against wingspan. Accordingly, Ksepka (2014) suggested Argentavis was more reliably sized at a 5.09 - 6.07 m wingspan, with estimates at the lower end of that range being predicted in most models. In contrast, the wing skeletons alone of P. chilensis and P. sandersi easily exceed wingspans of 4-5 m, and the addition of conservatively estimated feather lengths easily raise these wingspans into the 6-7 m range.

Skeletal reconstructions of giant pelagornithids: the holotype of Pelagornis chilensis (ventral view) and P. sandersi (dorsal view). That bird to the right of the image is a little thing called the wandering albatross, which has the largest wingspan of any extant flying bird. Pelagornithids must have been amazing to see in life. Images from Mayr and Rubilar-Rogers (2010) and Ksepka (2014).
Despite their size, pelagornithids were not heavy animals. Mass estimates for giant pelagornithids are in the region of 16–29 kg for P. chilensis and 21.9–40.1 kg for P. sandersi (Mayr and Rubilar-Rogers 2010; Ksepka 2014). Given that these birds are twice the size of albatrosses, the fact that their predicted masses match, or only double, the maximum masses of extant flying birds are surprising. Remember that mass increases by a factor of eight for every doubling of a linear dimension so, if we scaled wandering albatross (using masses given at Wikipedia) to Pelagornis proportions we’d expect a mass of 50-96 kg - well above those predicted figures. The pelagornithid weight-watching secret seems to lie in their unique wing proportions: even more than albatross, pelagornithids have extremely long wings compared to the rest of their bodies, and can thus attain giant wingspans while keeping their masses low. Similar tactics were also exploited by giant pterosaurs: maximising wing area while keeping the body small is a great way to maintain volancy at large size. Extremely thin walled bones also helped pelagornithids maintain low masses (and also explains why so many pelagornithid fossils look like they've been hit with a bulldozer).

Biological sailplanes

The largest pelagornithids were of a size which exceeds some theoretical flight limits for albatross-like birds (e.g. Sato et al. 2009), though the plainly obvious flight adaptations of their skeletons suggest this problem lies with our calculations and not the concept of pelagornithid flight itself. Indeed, glide analyses of P. sandersi indicate a supreme soaring capability with a very low sink rate (the rate at which altitude is lost during gliding) and high glide speeds, a combination that would facilitate extremely wide-ranging, energy efficient flight (Ksepka 2014). Their flight performance seems generally more akin to that of albatross than other pelagic birds, so reconstructions of pelagornithids riding air currents between waves, buzzing along the water surface and cruising on ocean winds seems sound. Reduced hindlimb proportions indicate that pelagornithids were probably not capable walkers or runners however, and we might envisage them only landing infrequently, perhaps most commonly when nesting. Curiously flattened and wide toe bones recall those of birds which use their feet as air brakes when landing (Mayr and Rubilar-Rogers 2010; Mayr et al. 2013), and may also have aided stabilisation on land (Mayr et al. 2013).

Predictions of glide ability and lift:drag ratios in P. sandersi from Ksepka (2014). Note how both models compare very well to albatross flight (black), but less well with frigate bird (red) or raptor flight (green).
Maintaining flight is a relatively easy part of aerial locomotion: how pelagornithids became airborne is trickier to fathom. This is mostly because of several indications of a limited flapping ability in the largest Neogene species, which are also the ones that would struggle the most with launch. Scaling of muscle energy availability predisposes all large flying animals to a relatively limited flapping capability and, like all fliers operating at the upper limit of their respective bauplan, pelagornithids likely relied on short-lived bouts of powerful anaerobic muscle activity to perform flapping (Ksepka 2014). But there is some question over whether they could flap their wings at all: several osteological features suggest pelagornithids had reduced shoulder/humeral motion (including a lack of rotatory capability) and lessened downstroke musculature (Mayr et al. 2008). A lack of dynamism is also seen elsewhere on the wing in that the articulation for the alulua was weakly developed, prohibiting spread of this structure during takeoff and landing (Mayr and Rubilar-Rogers 2010). The alula, when extended, allows the wing to function at higher angles of attack (the angle of the wing relative to the direction of airflow) and is thus very useful in initiating flight, controlling landing and general aerial manoeuvrability. Its immobility in pelagornithids would have impacted their range of flight dynamics quite considerably.

The large size of pelagornithids means that a very limited, maybe absent flapping ability may not be as detrimental as we intuitively predict. Flapping motions - both frequency and amplitude - reduce against increasing wing area and flight speed (the latter being predicted as high for any giant flier) so, as the largest flying birds of all time, pelagornithids may not have missed flapping as much as you'd think. But nonetheless, a significantly reduced flapping capacity and limited alula motion may have demanded fairly specialised launch and landing behaviour. Pelagornithids may have been limited to launching by simply extending their wings and using running, gravity or headwinds to find sufficient glide velocity. Landing, by contrast, may have involved low-angle approaches, slowing as much as possible (a dangerous game, as slower gliding also brings higher sink rates) and ditching to the ground. I can entirely believe that undignified semi-crash landings were common in this group.

If our understanding of pelagornithid flight is accurate, typical seabird behaviours like cliff-nesting - demonstrated here by northern gannets (Morus bassanus) - can be ruled out. Long winged seabirds are not the most agile fliers, but many still have enough control over their initiation and cessation of flight to land on small ledges. Pelagornithids trying this may have ended up splattered, Wile E. Coyote-style, on the side of a cliff. Photo by Georgia Witton-Maclean.
Being so light relative to wingspan would assist in both takeoff and landing, but nevertheless question marks hang over their ability to achieve flight in some conditions, such as escaping water (Ksepka 2014). Perhaps, like frigate birds, pelagornithids avoided entering water (though the former struggle with water escape because of waterlogged feathers rather than restricted flapping kinematics). I wonder if this is the case however, it being historically proposed that (unrealistically lightweight) giant pterosaurs could achieve flight from water by simply spreading their wings and catching wind (e.g. Bramwell and Whitfield 1974). The predicted pterosaur masses, wingspans and wing area models used in these old pterosaur studies are not far off those modelled today for giant Pelagornis: if so, could pelagornithids have escaped water with the same wing-spreading trick? It would be interesting to see this modelled biomechanically for a pelagornithid-specific model. On land, we may assume that pelagornithids favoured open space that permitted full 6-7 m wing spreads for launching and landing, and it would not be surprising if they favoured windy, elevated coastal regions that provided environmental launch assistance. I'm not sure what their prospects for flight in continental habits are, but it probably wasn't good: they almost certainly stuck to oceanic soaring, as suggested by the skew of their fossils to marine sediments.

They're only pseudoteeth, but I like them

We’ve made it all the way through this post without discussing the other characteristic anatomy of pelagornithids: their ‘pseudoteeth’. These structures are bony outgrowths of the jaw bones which strongly resemble actual dentition, though histological studies have verified that they lack all tissues associated with true teeth (Howard 1957; Louchart et al. 2013). It’s thought that pseudoteeth compensated for a well-developed hinge in the lower jaw that permitted wide horizontal bowing during feeding. This allowed for large prey to be swallowed, but compromised overall jaw integrity and potentially risked the loss of slippery seafood prey. A set of variably sized spikes along the margins of the beak is a great way to ensure snagged foodstuffs - thought to be mainly surface-seized fish or squid - did not slip from their beaks. Though many of the spikes are hollow, jaw bone surficial textures indicate that the entire jaw - pseudoteeth and all - was covered with a cornified sheath typical of other bird beaks (Louchart et al. 2013). As we discussed when looking at mammal horns, that’s a pretty potent combination for maximising lightness with strength, though the bone forming the pseudoteeth was mechanically weak and, despite their ferocious appearance, they were not adapted for tackling large, formidable prey (Louchart et al. 2013).

Holotype skull of P. chilensis in lateral view: check out those pseudoteeth. From Mayr and Rubilar-Rogers (2010).
That pelagornithid teeth functioned well as fish-grabs is suggested in their similarity in size and distribution to the dentition of other fish eaters, including certain crocodylians, large predatory fish, pterosaurs and temnospondyls. Quite how pelagornithids caught their prey is not well understood: if they could enter the water, they may have foraged from the water surface or dived; if not, they may have snatched prey from the water surface or stole it from other birds. Further research into pelagornithid flight capabilities and launch kinematics would narrow down this range of possibilities.

Recent studies have shown that pseudoteeth erupted from the jaw relatively late in pelagornithid growth (Louchart et al. 2013), meaning juvenile Pelagornis would have looked like regular, cute baby birds before developing their toothy smiles as adults. This has several interesting implications for pelagornithid growth and ecology. The first is that the cornified beak tissue covering their jaws must not have hardened until after the teeth had fully developed (recall from a previous post that cornified sheaths, on account of being inert, dead tissue, can’t be easily modified once deposited). This characteristic is not common among birds, but occurs in a number of Anseriformes. This observation is not a deal clincher for the pelagornithid-anseriform phylogenetic hypothesis, but it's an interesting connection nonetheless. Secondly, studies show that the emerging pseudoteeth were relatively delicate and potentially unable to withstand stresses imparted by thrashing fish or squid until late in development. This being the case, Louchart et al. (2013) proposed that pelagornithids might have been altricial, feeding regurgitated food to their offspring until they were fully grown and able to forage for themselves; or else that the juveniles were foraging on different foodstuffs. Altriciality would be unusual behaviour for a stem-neoavian as most bird species of this grade have precocial offspring that feed themselves straight after hatching. Insight into these hypotheses would be provided by fossils of juvenile pelagornithids but these remain extremely rare. I wonder if these animals were like living pelagic birds and nested atop cliffs in isolated offshore settings? If so, I wouldn’t hold your breath waiting for fossils of their hatchlings.

Enjoy monthly insights into palaeoart, fossil animal biology and occasional reviews of palaeo media? Support this blog for $1 a month and get free stuff!

This blog is sponsored through Patreon, the site where you can help online content creators make a living. If you enjoy my content, please consider donating $1 a month to help fund my work. $1 might seem a meaningless amount, but if every reader pitched that amount I could work on these articles and their artwork full time. In return, you'll get access to my exclusive Patreon content: regular updates on research papers, books and paintings, including numerous advance previews of two palaeoart-heavy books (one of which is the first ever comprehensive guide to palaeoart processes). Plus, you get free stuff - prints, high quality images for printing, books, competitions - as my way of thanking you for your support. As always, huge thanks to everyone who already sponsors my work!


  • Bourdon, E. (2005). Osteological evidence for sister group relationship between pseudo-toothed birds (Aves: Odontopterygiformes) and waterfowls (Anseriformes). Naturwissenschaften, 92(12), 586-591.
  • Bourdon, E., & Cappetta, H. (2012). Pseudo-toothed birds (Aves, Odontopterygiformes) from the Eocene phosphate deposits of Togo, Africa. Journal of Vertebrate Paleontology, 32(4), 965-970.
  • Bramwell, C. D., & Whitfield, G. R. (1974). Biomechanics of Pteranodon. Phil. Trans. R. Soc. Lond. B, 267(890), 503-581.
  • Campbell Jr, K. C. (1980). The world's largest flying bird. Terra, 19(2), 20-23.
  • Campbell Jr, K. E., & Tonni, E. P. (1983). Size and locomotion in teratorns (Aves: Teratornithidae). The Auk, 390-403.
  • Chatterjee, S., Templin, R. J., & Campbell, K. E. (2007). The aerodynamics of Argentavis, the world's largest flying bird from the Miocene of Argentina. Proceedings of the National Academy of Sciences, 104(30), 12398-12403.
  • Howard, H. (1957). A gigantic" toothed" marine bird from the Miocene of California. Santa Barbara Museum of Natural History, Department of Geology Bulletin, (1), 1-23.
  • Ksepka, D. T. (2014). Flight performance of the largest volant bird. Proceedings of the National Academy of Sciences, 111(29), 10624-10629.
  • Louchart, A., Sire, J. Y., Mourer-Chauviré, C., Geraads, D., Viriot, L., & de Buffrénil, V. (2013). Structure and growth pattern of pseudoteeth in Pelagornis mauretanicus (Aves, Odontopterygiformes, Pelagornithidae). PloS one, 8(11), e80372. 
  • Mayr, G. (2011). Cenozoic mystery birds–on the phylogenetic affinities of bony‐toothed birds (Pelagornithidae). Zoologica Scripta, 40(5), 448-467.
  • Mayr, G., Hazevoet, C. J., Dantas, P., & Cachão, M. (2008). A sternum of a very large bony-toothed bird (Pelagornithidae) from the Miocene of Portugal. Journal of vertebrate Paleontology, 28(3), 762-769.
  • Mayr, G., & Rubilar-Rogers, D. (2010). Osteology of a new giant bony-toothed bird from the Miocene of Chile, with a revision of the taxonomy of Neogene Pelagornithidae. Journal of Vertebrate Paleontology, 30(5), 1313-1330.
  • Mayr, G., Goedert, J. L., & McLeod, S. A. (2013). Partial skeleton of a bony-toothed bird from the late Oligocene/early Miocene of Oregon (USA) and the systematics of neogene Pelagornithidae. Journal of Paleontology, 87(5), 922-929.
  • Sato, K., Sakamoto, K. Q., Watanuki, Y., Takahashi, A., Katsumata, N., Bost, C. A., & Weimerskirch, H. (2009). Scaling of soaring seabirds and implications for flight abilities of giant pterosaurs. PLoS One, 4(4), e5400.


  1. That bit about Argentavis was particularly interesting. I didn't realize they were that much smaller than Pelagornis.

  2. what about their feathers being a meter long?

  3. Also, might have they been actually flightless and simply used their wings as sails like the Tualapi?

  4. "As we discussed when looking at mammal horns, that’s a pretty potent combination for maximising lightness with strength, though the bone forming the pseudoteeth was mechanically weak and, despite their ferocious appearance, they were not adapted for tackling large, formidable prey (Louchart et al. 2013)."

    On the other hand, the fact the bone-teeth didn't project outwards like the procumbent teeth of typical ornithodires, suggests such would have risked injury to the weak bone-teeth: and therefore that the prey was comparatively large enough to damage its jaws. Fish can slap around when caught but squid do not because of the squid pen. mmm...

  5. The hindlimbs of birds do most of the work to generate thrust for the launch (i.e. a stationary take-off is initially mostly hindlimb-driven). So the reduced shoulder/humeral motion of pelagornithids might not have been such a problem for a stationary take-off. Even a rudimentary flight stroke could possibly have been sufficient for the 'climb-out' phase of the launch. So a running take-off might not have been necessary for pelagornithids (contra Ksepka, 2014), unless they were launching from water.
    Then again, maybe pelagornithid hindlimbs were too short to generate sufficient thrust for the launch - and/or too short to allow the long wings to clear the ground during the first downstroke of the launch.

  6. Among tested taxa in the large reptile tree, Puffinus the shearwater, nests most closely to Pelagornis. Next is Morus, the gannet.

  7. My gosh that's a lot of spam....

    More to the point, what are current weight estimates looking like for Argentavis? Is it still considered more massive than the big pelagornithids?

    1. It should be around the same weight as P. sandersi.

  8. So if pelagorns didn't nest on cliffs, where did they nest?

  9. This comment has been removed by a blog administrator.