The Bird From Hell?

By Stephen Wroe
Nature Australia, Volume 26, no 7:56 - 63

What did this gigantic bird eat? Built like a brick 'out house', with a head the size of a horse's, and military-issue jaw muscles, dare I say just about anything it darn well felt like!

With his massive forearms straddling the kangaroo carcass, the marsupial lion hurriedly scissors through his victim's tough hide with secateur-like cheek-teeth. Now with the roo's flesh exposed, he begins to bolt down steak-size slabs of meat. The size of a small Rottweiler, but far more muscular, this powerful carnivore is the top mammalian predator of his day. Even so, he can't relax. As he frantically works into the roo's hind-quarters the reason for his unease becomes apparent. A distant rumble heralds the end of his meal---and he knows it. The rumble turns into a ground-shaking thunder, accompanied by the sounds of snapping undergrowth. From 2.5 metres above, three pairs of cold, almost reptilian eyes stare him down. The gig is over. With a surly but half-hearted hiss, the 'lion' backs off. But there's no shame in giving way to these adversaries. Each weighs over 300 kilograms. Our 'lion' has just been muscled off his kill by the most formidable bipedal carnivores since the extinction of the dinosaurs. Fifteen million years ago, in northern Australia, thunder birds ruled OK!

Thunder birds or mihirungs (family Dromornithidae) were gigantic, ground-dwelling birds that roamed Australia from at least 25 million to around 26,000 years ago. Among the eight species was the largest bird that ever lived, Dromornis stirtoni, three metres high and weighing over 500 kilograms. Fossilised remains of thunder birds were first described in 1839, but until now most material has consisted of assorted post-cranial elements. Little has been known of dromornithid skull morphology. Of course, for working out both the relationships and lifestyles of extinct animals, the skull is the business end of the animal. Consequently, dromornithids have remained conspicuous but enigmatic players in Australia's faunal history.

In the absence of conclusive evidence it has traditionally been contended that thunder birds were related to ratites (Emus, Ostriches, cassowaries etc.) and that, like ratites, they were mostly herbivorous. Spectacular new evidence in the form of several skulls of a 15-million-year-old dromornithid, Bullockornis planei, has shed new light on the question of dromornithid origins. Peter Murray and Dirk Megirian from the Northern Territory's Central Australian Museum argue that these boofheaded dromornithids were in fact more closely related to the anseriform bird clan---geese, ducks and other water fowl. But this new material raises other questions. In particular, as suggested in my opening paragraph, did Bullockornis planei and perhaps other thunder birds use their powerful skulls and beaks to eat meat? Before considering this specific issue, I need to explain a little about the inference of diet in fossil animals in general, using marsupial lions and another giant terrestrial bird as examples.

Reconstructing the lifestyles of extinct animals is fraught with difficulty. One obvious and unpleasant obstacle stands in our way---they're all dead. As a result, the task of the palaeobiologist is essentially that of the forensic scientist---to erect the best supported scenario using whatever fragmentary evidence is at hand. This evidence flows from several lines of reasoning. One is phylogenetic; that is, after determining the closest extant relative of the extinct species in question, we might assume, all else being equal, that the behaviour of the deceased is reflected in that of the living. A second basic principle, that of functional analogy, operates on the assumption that, if a living animal possesses a feature present in an extinct one, then the feature was probably put to the same purpose in both. Again, by studying living forms, we gain insight into the lifestyle of the extinct. These arguments have merit, and have often been invoked by palaeobiologists. However, there are problems with both.

Clearly, the more distant the relationship between an extinct species and its living relatives, the more tenuous the assumption that they shared similar lifestyles. Regarding the second principle, the possibility always remains that an extinct animal may have used a similar feature in a living species for different purposes. Moreover, in some cases, there are simply no comparable living analogues. Because of these uncertainties, it has been suggested that a third line of evidence should be considered. This method, called 'biomechanical design analysis', involves examining the feature of interest in a fossil species and then matching its form with the most likely, hypothetical function.

In 1858, the renowned palaeontologist Richard Owen described the first specimen of a curious Pleistocene-aged marsupial and named it Thylacoleo carnifex. Examination of its premolar teeth left Owen in no doubt that he was looking at a carnivore par excellence, based on acceptance of the theory that large, vertically slicing cheek-teeth (carnassials) are the hallmarks of a mammalian predator. Owen thus drew on inference from our second 'rule of thumb'---analogy with the closest living structural counterparts. However, it wasn't long before others questioned his reasoning and proposed alternative purposes for Thylacoleo's monstrous premolars. In hindsight some were laughable, culminating in the often-ridiculed suggestion that Thylacoleo was a melon-muncher. But no matter how silly such explanations seem now, they were based on the generally accepted view that Thylacoleo was a member of Australia's great diprotodontan radiation and, because modern diprotodontans (including wombats, kangaroos, possums) are largely vegetarian, it seemed reasonable to assume Thylacoleo was too. In further support of this argument it was pointed out that Thylacoleo lacked large death-dealing canines and, instead of having large temporal muscles to control the lower jaw, like living carnivores, they had huge masseter (cheek) muscles like plant-eaters.

The debate over how Thylacoleo made a living has since gone full circle, thanks largely to a detailed form-function analysis by Rod Wells (Flinders University) and colleagues who convincingly argued that the carnassials were ideally suited to shearing flesh. Another important point raised by Wells and Co. was that Thylacoleo's massive jaw muscles and bolt-cutting cheek-teeth were seriously 'over-designed' for any other purpose but carnivory. Using Thylacoleo's cranial hardware to eat fruit, or even the hardest of nuts, would be tantamount to calling the fire brigade to put out a match. Mother Nature is a hard task mistress, and one thing she frowns on is waste. Maintaining industrial-grade hardware for a role that could be performed far more cheaply is a serious infraction. Evolutionary theory predicts that natural selection will ruthlessly weed out and exterminate features or species found guilty of such offences.

The history of interpretation of Diatryma gigantea---a giant terrestrial bird excavated from 60--55-million-year-old deposits in North America and Europe, parallels that of Thylacoleo in some important respects, but with interesting differences. Based on the huge beak and jaw muscles, Diatryma's lifestyle was likened to that of Phorosrhacos, another extinct giant, ground-dwelling bird. Phorusrhacoids were a diverse group, represented on every continent except Africa and Australia. Although some of the smaller species could fly, the larger, ground-dwelling varieties were either heavy, bulky scavengers or swift, active predators. The latter included the awesome Titanis walleri from North America, which may even have persisted as recently as 12,000 years ago, making it the closest thing to a Velociraptor ever seen alive by humans. Anyway, no-one ever doubted that all phorusrhacoids were carnivores. Consequently, Diatryma remained guilty by association until, 72 years later, an alternative theory was flagged. In 1989, Allison Andors (American Museum of Natural History) carried out a detailed biomechanical analysis and concluded that Diatryma was a herbivore. In addition he pointed out that Diatryma lacks some features common to most carnivorous birds (its beak isn't hooked and its toes are not equipped with vicious talons) and suggested that the closest comparisons among living birds are grass- and leaf-eaters. Case closed? Not quite.

In 1991, following another detailed investigation, this time by Lawrence Witmer and Kenneth Rose (Ohio and John Hopkins Universities, respectively), the flesh-ripping, bone-crunching image of Diatryma was resurrected. They pointed out that living birds can crop grass and leaves, and crack the largest of nuts (Coconuts excepted), using heads a fraction the size of Diatryma's. Unlike mammals, birds don't process food in their mouths, so for plant-eating birds, there is no advantage in increasing the absolute size of the head beyond that required to crop the grass or bust the nut of choice. Put another way, in an up-sized version of any bird the head would become smaller compared to the body, assuming that it ate the same food. For example, the Australian Palm Cockatoo (Probosciger aterrimus) has a fist-sized head and a very powerful beak that are relatively large compared to its body. With this equipment it can crack the largest of Aussie nuts. Of course, the Palm Cockatoo is a midget compared to Diatryma (or dromornithids). If we scaled the Palm Cockatoo's body up to giant-like dimensions, but kept its head fist-sized, it could still crack the largest native nuts. A bigger head and beak would provide little advantage. More to the point, this would transgress that widely accepted dictum---natural selection won't tolerate excess. Thus, as Witmer and Rose explain, the important point is not the relative size of Diatryma's head compared to its body, but the absolute size of its head relative to the proposed food of choice. Unless Diatryma was regularly busting Coconuts, they argue, it couldn't have been a plant-eater---its head was too big! Supporting this hypothesis is the fact that, in giant birds that are known to be plant-eaters (moas, Emus etc.) the head is small compared to the body. To Witmer and Rose the only way to explain Diatryma's big head was that it took larger prey.

What does all this have to do with the palaeobiology of thunder birds? Well, hopefully I've demonstrated the following points. First, inferring the biology of extinct animals through comparison with their closest living relatives is anything but conclusive when the two groups in question are only distantly related. Second, the fact that a fossil animal doesn't fit a given stereotype for all features common to a living guild is no reason to presume that it couldn't once have been part of such a guild, as long as some of its features do 'fit the bill'. Third, regarding the diets of extinct animals, the case should never be closed until thorough biomechanical design analyses have been conducted. The relevance of this to dromornithid diet is clear. Their closest living relatives (whether you take these to be ratites or geese) are very distant, separated by tens of millions of years, and a lot can happen in this time; although they lack some features present in most living carnivorous birds, they certainly possess others; and, finally, no-one has yet indulged in a detailed biomechanical design analysis for dromornithids.

Historically, the argument against carnivory in thunder birds was never strong and the new skull material presented by Murray and Megirian strengthens my conviction, at least for Bullockornis planei. The bill of Bullockornis is convex, very deep, and driven by powerful muscles (evidenced by enormous muscle-attachment sites). These are all standard equipment in most living carnivorous birds, Diatryma and phorusrhacoids. It further shares with Diatryma and phorusrhacoids a huge head, an inability to fly and nostrils placed well back on the beak. Unlike most living raptors and most phorusrhacoids, but in common with Diatryma, Bullockornis is massively built and lacks a distinctively hooked bill and talons. But then other living carnivorous or scavenging birds, such as the African Carrion Crow (Corvus crassirostris), also lack these features. The bottom line is this: Bullockornis certainly had the features required to kill and butcher large prey. Of course, this doesn't prove that it did. Some living plant-eating birds also possess many of these features. But given that the huge head and beak of Bullockornis appear to have been 'over-designed' for processing available plant material, I believe there is a strong prima facie case for the assertion of a carnivorous lifestyle. Even if Coconut-sized nuts were present in Miocene Australia, it would have to be demonstrated that they existed in sufficient abundance to sustain 300-kilogram birds in order to explain away the 'over-design' problem.

So it's time to ask again: what did this gigantic bird eat? Built like a brick 'out house', with a head the size of a horse's, and military-issue jaw muscles, dare I say just about anything it darn well felt like!

On a less frivolous note, the Bullockornis skull material raises more questions than it answers. For example, do arguments for carnivory in Bullockornis apply to other thunder birds as well? We'll need more fossil skulls to tackle that question. Certainly, the Pleistocene dromornithid Genyornis newtoni had a much smaller skull than Bullockornis, in both absolute and relative terms. Even so, a scavenging role has been suggested for this bird. Recent evidence, based on the relative proportions of different carbon isotopes present in fossil eggshells, has been interpreted as supporting the hypothesis that Genyornis was primarily a browser. However, in arriving at this conclusion the authors had assumed, a priori, that the bird was a herbivore. If we dismiss this assumption, as we must until more conclusive evidence comes to hand, then these data could just as easily support the proposition that Genyornis fed largely on browsing herbivores. Consequently, I conclude that isotopic evidence lends nothing to either pro-herbivory or pro-carnivory arguments at present.

Other evidence may count against the inference of carnivory for some thunder birds. For example, unusually for big carnivores, dromornithid fossils are relatively common. Of course, fossilisation is a fickle process and many explanations could account for this excess. For Dromornis fossils, Peter Murray suggests they are abundant because birds from a wide area congregated at water holes in times of drought.

Another question: how did such large birds manage to run down and secure their prey? Incidentally, this same question has been levelled at those who support a predatory habit for the ultimate bipedal monster, Tyrannosaurus rex (see Nature Aust. Summer 1996--97). A number of alternatives are available. Perhaps dromornithids weren't active predators, but scavengers, using their intimidating size to muscle other carnivores from kill sites. On the other hand, this query presupposes that the animals dromornithids hunted were themselves fast runners. This doesn't necessarily follow. Certainly not all dromornithids would have been ponderous slow movers. Maybe some scavenged and others chased? Form-function analyses for Diatryma and T. rex have indicated that they were not necessarily slow. Likewise an analysis by Peter Murray indicates that even the largest thunder bird, D. stirtoni, may have been capable of surprising speed, up to 35 kilometres per hour.

I don't pretend to have solved the riddle of thunder bird palaeobiology. Indeed, anything but. More realistically, I hope to have raised enough doubt to stimulate further research. There are arguments for and against carnivory in Bullockornis and other dromornithids, but to those in the anti-camp, the function of one conspicuous complex of features demands explanation. If Bullockornis didn't regularly feed on vertebrate carcasses, then what was its undeniably massive and powerful skull, beak and chewing muscles used for? Invoking the principle argued above, if it was cracking nuts or cropping grass, then it suffered from a serious 'over-design' problem. Herbivory for this bird appears to defy evolutionary theory. If this is true, then Bullockornis was the bird from hell, a truly terrifying 'reincarnation' of those long-gone ancestors of all birds, the theropod dinosaurs.

Australian Museum Mammals

The Bird From Hell?

By Stephen Wroe
Nature Australia, Volume 26, no 7:56 - 63

Further reading

Australian Museum Mammals

The Bird From Hell?

By Stephen Wroe Nature Australia, Volume 26, no 7:56 - 63

Further reading

By Stephen Wroe
Nature Australia, Volume 26, no 7:56 - 63