Diet of Mesozoic toothed birds (Longipterygidae) inferred from quantitative analysis of extant avian diet proxies.

Background: Birds are key indicator species in extant ecosystems, and thus we would expect extinct birds to provide insights into the nature of ancient ecosystems. However, many aspects of extinct bird ecology, particularly their diet, remain obscure. One group of particular interest is the bizarre toothed and long-snouted longipterygid birds. Longipterygidae is the most well-understood family of enantiornithine birds, the dominant birds of the Cretaceous period. However, as with most Mesozoic birds, their diet remains entirely speculative.

Results: To improve our understanding of longipterygids, we investigated four proxies in extant birds to determine diagnostic traits for birds with a given diet: body mass, claw morphometrics, jaw mechanical advantage, and jaw strength via finite element analysis. Body mass of birds tended to correspond to the size of their main food source, with both carnivores and herbivores splitting into two subsets by mass: invertivores or vertivores for carnivores, and granivores + nectarivores or folivores + frugivores for herbivores. Using claw morphometrics, we successfully distinguished ground birds, non-raptorial perching birds, and raptorial birds from one another. We were unable to replicate past results isolating subtypes of raptorial behaviour. Mechanical advantage was able to distinguish herbivorous diets with particularly high values of functional indices, and so is useful for identifying these specific diets in fossil taxa, but overall did a poor job of reflecting diet. Finite element analysis effectively separated birds with hard and/or tough diets from those eating foods which are neither, though could not distinguish hard and tough diets from one another. We reconstructed each of these proxies in longipterygids as well, and after synthesising the four lines of evidence, we find all members of the family but Shengjingornis (whose diet remains inconclusive) most likely to be invertivores or generalist feeders, with raptorial behaviour likely in Longipteryx and Rapaxavis.

Conclusions: This study provides a 20% increase in quantitatively supported fossil bird diets, triples the number of diets reconstructed in enantiornithine species, and serves as an important first step in quantitatively investigating the origins of the trophic diversity of living birds. These findings are consistent with past hypotheses that Mesozoic birds occupied low trophic levels.