Traditionally, gastornithids and dromornithids have been argued as carnivores due to their massive, deep bills. Modern examples of large flightless birds, the ratites, have rather shallow bills, and indeed their herbivorous and/or omnivorous diets do fine with such simple jaws. Thus, it’s been traditionally argued that the crushing maws of these gigantic stem-fowl were ridiculously overbuilt for a herbivorous diet, that carnivory, be it in the form of scavenging or active predation, was the only way such a beak would evolve and be selectively pressured to be retained.
Alas, things are not as simple as they seem. Besides actual isotope studies directly indicating that Gastornis and Genyornis were herbivores, and the utter lack of predatory features dooming any sincere assertions of carnivory for other members of either clade, the fact is that a number of other flightless herbivorous birds also developed such “overbuilt” bills. The most obvious case is Sylviornis, the mysterious New Caledonian stem-fowl, which developed equally as deep and powerful a beak as it’s distant dromornithid and gastornithid, in fact being directly compared in terms of morphology to those of the former. Yet it has always been clear that this bird was primarily herbivorous bird, with a penchant for granivory. Hitting even closer to home are the Hawaiian Moa-Nalos, which also developed proportionally massive and usually deep bills, yet clearly as means to forage on the various types of ferns and other vegetation in the older islands of the archipelago. Among living birds, the Takahe of New Zealand made the already fairly deep rail jaws into a hatchet like instrument, all just to cut through thick grasses. Even the famous Dodo seemingly went down this path: from the rather unassuming, thin pigeon jaws came a seriously impressive pair of pincers that quickly became infamous for their nasty bite, all to aid in foraging in Mauritian forest undergrounds.
Even ratites seemingly went down this path: Pachyornis possess a powerful, deep beak that is thought to have aided the animal to feed on harsh scrub vegetation, while Aepyornis developed one of the largest skulls known in all fossil birds; it is very unlikely it was anything other than a browser with frugivorous tendencies, though it would have surely resorted to animalivory occasionally. Outside of Aves, one could argue that oviraptors represent yet another case where a powerful beak has evolved in correlation to specialized herbivory.
So, as it seems, several large, flightless birds with primarily herbivorous diets developed “overbuilt” jaws. While the exact mechanics and reasons for this specialization aren’t entirely well understood, it seems that there do seem to be several reasons as to why “overbuilt” jaws are pressured to evolve. In Pachyornis and Moa-Nalos, correlation between diet and the development of their beaks has been somewhat established: the former fed primarily on hard scrub vegetation, twigs and bark forming a component, while the latter seemingly had a preference for foraging on ferns. The former would naturally favor more powerful bites, while the latter seemingly proved a challenge in terms of wearing down the jaws, which is presumably while Moa-Nalos also developed unique serrations in their bills. Granivory has been suggested as the reason as to why Sylviornis‘ jaws developed, at at times also implicated for Gastornis and the dromornithids, though evidence of granivory has only been acquired for Gastornis — and even then, in the form of rather generic seeds, not the expected coconut like seeds suggested. The Takahe’s jaws clearly evolved for breaking down large grasses and reeds, while the Dodo’s beak is thought to have evolved in response to a pig-like foraging lifestyle in Mauritius’ forest undergrowth. With the exception of the Takahe and the Dodo, most of these powerful beaks seem to correlate with browsing.
Thus, “overbuilt” beaks seem encouraged in many contexts within the herbivorous flightless bird biological archetype. The fact that modern ratites lack such jaws is therefore not indicative of a correlation between avian herbivory at large sizes and small heads, especially when even a few extinct ratites produced powerful jaws. One may indeed wonder if many extinct birds without well preserved skulls, like Eremopezus or eogruiids, also had powerful beaks.
Sorenson et al. (1999): Relationships of the extinct moa-nalos, flightless Hawaiian waterfowl, based on ancient DNA. Proceedings of the Royal Society.
Williams, D. L. G. (1981) Genyornis eggshell (Dromornithidae; Aves) from the Late Pleistocene of South Australia
Burrows, C. J. 1980. Some empirical information concerning the diet of moas. New Zealand Journal of Ecology 3, 125–130.