Trophic diversity and evolution in Enantiornithes: a synthesis including new insights from Bohaiornithidae

Reviewer #1 (Public Review):

Understanding the ecology including the dietary ecology of enantiornithines is challenging by all means. This work explores the possible trophic diversity of the "opposite-bird" enantiornithines by referring to the body mass, jaw mechanical advantage, finite element analysis of the jaw bones, and morphometrics of the claws and skull of both fossil and extant avian species. By incorporation of the dietary information of longipterygids and pengornithinds, the authors predicted a wide variety of foods for enantiornithine ancestors. This indicates the evolutionary successes of enantiornitine during Cretaceous is very likely to have been driven by the wide range of recipes. I believe this work represented the most comprehensive analysis of enantiornithines' diet and trophic diversity by far and the first systematic dietary analysis of bohaiornithids, though the analysis themselves are largely based on the indirect evidence including jaw bone morphologies and claw and skull morphometrics. Anyway, I believe the authors did most the paleontologists could do, and I do not know whether the conclusions could be further supported by incorporating some geochemical data, as most of the specimens the authors analyzed were recovered from a small geographic area. The results also indicate that the developmental trajectories of enantiornithines, at least for jaw bones, might also have been diverse to some extent in response to the diverse ecological niches they adapted. My only concern regarding the analysis is to what extent the conclusions are convincing by comparing specimens representing various ontogenetic stages.

Reviewer #2 (Public Review):

Miller et al. take a variety of measurements and analytical techniques to assess the ecology of various species of the enantiornithine clade Bohaiornithidae. From this they suggest that the ancestral enantiornithine was a generalist and that the descendant clades occupied a breadth of niches similar to that of the radiation of derived birds after the K-Pg extinction.

I am not a statistician so I found much of the paper to be outside my ability to review. I also am not an expert on enantiornithines or cranial morphology of birds, so these areas I also am not the best reviewer.

However, I have published on bird foot functional morphology, notably that of birds of prey. This area thus is where I concentrated my efforts in the review.

Overall, I find the idea that enantiornithines had occupied a similar niche breadth to post-K-Pg derived birds to be a curious, thought provoking proposal. On methodology, I have a few questions about bird feet comparisons. Whether my comments require minor or major edits is not really possible to say since I am not commenting on e.g. the skull-based analyses.

STRENGTHS
The paper uses a multi-proxy approach to assess ecological categories. This is broader than in previous works and is to be commended. I am not well placed to comment on the specifics of the statistical methods however.

LANGUAGE
The manuscript is very well written. I don't recall seeing many or possibly any grammatical issues. That's rare these days and I commend the authors on checking their manuscript and making it readable. This said, I found the extensive use of acronyms and abbreviations to be difficult to follow. This is not much of a criticism but in a general-readership journal, perhaps not having everything abbreviated might be preferential.

The manuscript uses phrases like "superficially resembles" and "is similar to" a lot. I'm trying not to be picky, but very often these phrasings don't say how the features are similar (or not). Is it the curvature etc? Could these be expanded upon a bit more in the text please? It isn't very easy to assess similarity r dissimilarity without some point of reference.

FIGURES
The figures are generally very good, and the captions are generously descriptive. However, all figures are graphs, tables, etc. It would be nice, somewhere, to have an image or group of images showing us what a bohaiornithine is.. especially since this is a general-readership journal. I wasn't aware of the details of enantiornithine clades before reading this manuscript, and I suspect other readers would be in the same place. Can we get some images of fossils, a skeletal diagram, or something?

RAPTOR CLAWS
This is my main criticism.

The foot morphometrics suggest that there is a morphological difference between claws of raptors that feed on large prey, and those of raptors that feed on small prey. I am curious what these morphological differences are.

In our paper(s) (Fowler et al., 2009; 2011), we looked at the feet (especially the claws) of various birds of prey, and studied foot functional morphology compared with prey choice, capture and immobilization strategy. We devised a behavioural categorization that separated the behavior (mainly in subduing the prey) between "small" and "large" prey, that being whether they can be fully contained within the foot of the raptor. Most if not all raptors take small prey, and these are typically killed using constriction. Some raptors have specialized in small prey/constriction (e.g. most owls). Some raptors might also take large prey, but since (by definition) large prey cannot be fully contained within the foot then the prey item cannot be constricted and a different immobilization (kill) mechanism must be employed (which differs among clades).

We never made a morphological distinction between small and large prey specialists largely because all raptors take small prey. I am thus interested in what taxa are designated small vs large prey specialists in this study. Perhaps these authors have found characters that distinguish primarily small-prey-specialist raptors, but I do not know what they are and maybe this should be included in the text somewhere.

Owls are mainly small prey specialists. Compared with other raptors, they have a unusual foot that has (I am generalising here) short non-ungual phalanges contrasting with long ungual phalanges which are relatively low curvature. We (Fowler et al 2009) suggest that this gives owls a more tightly closable foot (short non-ungual phalanges), but maintains reach of each toe (long claw). This could be seen as indicative of small -prey specialization, but again, other raptor clades take small prey without this very specialized foot. If the "small prey specialist" category here is really just owls then it might be slightly misleading.

This is my main criticism. I would at least like some explanation of what is in this category.

Otherwise I must leave assessment of cranial functional morphology, and general statistical analysis to other reviewers.

IMPACT
As I have already stated, the idea that Enantiornithines occupied a similar breadth of niches to post K-Pg birds is thought provoking, moreso than upon initial reading. The authors note that this raises questions about the adaptations or survivorship of derived birds, and this is what I find most intriguing, and is what I think will appeal to most readers.

Reviewer #3 (Public Review):

Summary:
The authors use several quantitative approaches to characterize the feeding ecologies of bohaiornithid enantiornithines, including allometric data, mechanical advantage and finite element analyses of the jaw, and morphometric analyses of the claws. The authors combine their results with data for other enantiornithines collected from the literature to shed new insight on the ecological evolution of Enantiornithes as a clade.

The approaches used by the authors are generally appropriate for the questions being asked, their comparisons are thorough, and the interpretations are generally reasonable. However, there are a number of major flaws to the comparisons used that should at least be addressed by the authors, if not overcome by modifying the methodology. Smaller concerns/comments are provided in "Recommendations for the authors."

My first major concern is about how the presence of teeth might influence both qualitative and quantitative comparisons to extant birds. The authors should discuss how the presence of teeth might facilitate or prevent feeding strategies that might be inconsistent with (for example) patterns reconstructed using finite element analysis for a comparative sample of toothless birds.

Next, the authors should discuss the potential impact that cranial kinesis might have on the functionality of the jaws - especially with regards to the mitigation of stresses experienced by the skull. Do the quantitative approaches used here to characterize the mechanics of the jaws account for kinesis in extant birds? If so, how? If not, how do the authors' account for that mechanical difference in their interpretations?

My next concern regards potential biases introduced by the approach taken to reconstruct the bohaiornithid skulls sampled here. Using elements from closely related taxa to fill out an incomplete skull during reconstruction is reasonable, but it may influence the results of subsequent shape comparisons - especially when the "donor" skull is compared to the recipient. The authors should explain how they accounted for this possibility in their methods or their interpretations.

Next, it is unclear how or where much of the data used or generated by this study are made available. I appreciate that the authors thoroughly cite the literature from which some data (e.g., extant FEA data), but all data used should be provided in the supplement. Likewise for the FEA models generated for the newly sampled taxa. The authors indicate that some R scripts are available online (Lines 787-788), but that link is currently non-functional, so I could not verify what was made available. And unless I missed it, the authors don't indicate that other data (e.g., FEA models) are also available there. Any data used in, or generated by, this paper should be made available online - including FEA models, tree files and analysis output files.

Also pertaining to the methods, in some places, the methods the authors used to analyze their data were not specified. For example, the authors mention that "all analyses of the [MA/FEA] data were performed in R" and "scripts [are] available" online (Lines 786-787), but the authors don't specify what those analyses actually are - unless that was specified elsewhere and I missed it? I know very little about FEA or MA analyses, so maybe these approaches are well understood in those circles, but I am unable to assess the approaches here without downloading and digging into the scripts.

A broader recommendation here: in several places, I found this paper difficult to follow. That's partly understandable, the authors are discussing and comparing trends across a wide variety of data types and analyses - which is certainly both challenging and commendable. But that variety of analyses has resulted in a staggering variety of acronyms that I found nearly impossible to keep track of. Minimally, I recommend that the authors redefine the most important acronyms at the start of each major subheading.

Related to that last point, in the discussion, I often found myself missing the forest for the trees, so to speak - the authors paid much attention to interpreting the results of each analytical approach for each taxon (which I appreciate), but I found it difficult to keep track of the take-home message the authors were trying to convey. I would recommend a reorganization of the discussion that follows a backbone based on the authors' key messages - for example, a section on species-level interpretations (maybe with sub-headings for each approach discussed), followed by larger-picture discussions of Bohaiornithidae and Enantiornithes more generally. The authors included a section at the end of their discussion that already provides that larger picture for Enantiornithes, but the section on "Bohaiornithid Ecology and Evolution" includes a lot of species-level comparison that I think would be better suited for species-focused sub-sections, and I think the paper would be better served by reserving this section for a bohaiornithid-level survey.

Author Response

We thank all three reviewers for their detailed reviews, and generally agree with their feedback. To accompany the reviewed preprint of this manuscript, we wished to respond to comments from the reviewers so that they (and the public) will know what we are planning to incorporate in the revised manuscript we are currently preparing. If there are any comments on our plans in the meantime, please let us know.

• Reviewer 1, on concerns regarding identification of ontogenetic stage and comparison of taxa from different ontogenetic stages: It is fair to say that enantiornithine ontogeny is still poorly understood, though we believe all current evidence points to each specimen used in this study to being adequately mature for comparison to the extant birds used in the study. Stages of skeletal fusion are the standard method of assessing enantiornithine ontogeny (Hu and O'Connor 2017), and our comparison of histological work (Atterholt, Poust et al. 2021) to skeletal stages in Table S4 suggests a transition from juvenile to subadult in stage 0 or 1 and from subadult to adult within stage 3. Thus, the specimens we quantitatively examine in this study, all at stages 2 or 3 (Figure S10), are advanced subadults or adults. It is well-known that many living animals considered “adults” would be considered subadults or even juveniles to a palaeontologist (Hone, Farke et al. 2016). So, even if some individuals in this study are not fully skeletally mature, they should have obtained the morphology which they would possess for most of their lives and thus the morphology which undergoes selective pressure. We will add this context to the “Bohaiornithid Ontogeny” section and thank the reviewer for seeking more detail for this point.

• Reviewer 2, on need of a context figure: We have an artistic life reconstruction of a bohaiornithid in preparation, and can include that in the revised manuscript as a figure.

• Reviewer 2, on raptor claw categories: We explain these categories in-depth in a previous work (Miller, Pittman et al. 2023). However, we will now add a short summary of that explanation to this work so that this manuscript will become self-contained in this regard. In short, the “large raptor” category includes extant birds with records of regularly taking prey which cannot be encircled with the pes, while birds in the “small raptor” have no such records. As Reviewer 2 points out this does often follow phylogenetic lines, but not always. E.g. most owls specialise in taking small prey, but the great horned owl Bubo virginianus regularly takes mammals and birds larger than its pes (Artuso, Houston et al. 2020); and conversely we can only find reports of the common black hawk Buteogallus anthracinus taking prey samll enough for the pes to encircle (Schnell 2020) despite other accipiters frequently taking large prey. In both cases these taxa plot in PCA nearer to other large or small raptors (respectively) than to their phylogenetic relatives.

• Reviewer 3, on teeth vs beaks: We are not aware of any foods which are exclusive to toothed or beaked animals. There are some aspects of extant bird biology that may affect the way a certain diet may need to be adapted to which we do comment on, e.g. discussion of alternatives to the crop and ventriculus for processing plant matter in the Bohaiornithid Ecology and Evolution section. For functional studies, e.g. FEA, we have included the rhamphotheca in toothless models which serves the same role as teeth, to be a feeding surface. It should not matter, in theory, if the feeding surface is hard or soft as mechanical failure occurs in high stress/strain states regardless of the medium. If having teeth necessarily increases or decreses overall stress/strain relative to a beak (and from our work this does not appear to be the case), this would in turn necessarily limit dietary options. So, all models in our work should be directly comparable.

As an additional note on this topic, we address tooth shape in bohaiornithids at the end of the Bohaiornithid Ecology and Evolution section. We specifically note that their tooth shape is likley controlled by phylogeny in the current version, though we will add a note in the upcoming version that the morphospace of bohaiorntihid teeth overlaps that of many other clades with purportedly diverse diets, which is consistent with a hypothesis of diverse diets within the clade.

• Reviewer 3, on cranial kinesis: Our FE models should be unaffected by cranial kinesis, as these are two-dimensional and model the akinetic lower jaw only. Some mediolateral kinesis may be relevant in the mandible in the form of “wishboning” in different taxa, but its prevalence in extant birds is currently unknown. The preservation of enantiornithines (two-dimensionally and typically in lateral view) limits the ability to capture any mediolateral function regardless.

Our models of mechanical advantage do not account for any cranial kinesis. This is a necessary simplifcation. The nature of cranial kinesis in extant birds, and the role that it plays in feeding, is poorly understood. Cranial kinesis will increase gape, but we don’t yet know how/if it affects jaw closing force and speed (moreover, given the variation in quadrate and hinge morphology present in extant birds, this is also something that is likely to be highly diverse). We have therefore modelled the extant birds’ jaw closing systems as having one, akinetic out lever (the jaw joint to the bite point), to match the situation in our fossil taxa. This is a common simplification that has been used previously with success (Corbin, Lowenberger et al. 2015, Olsen 2017). However, we acknowledge that this simplification may introduce some error. Unfortunately, until the mechanics of cranial kinesis – and the variation in the anatomy and performance of kinetic structures in extant birds – are better understood, we cannot determine exactly what that error looks like. We therefore have greater confidence in the inter-species comparability this conservative, akinetic approach (in other words, we may not be making assumptions that are 100% accurate, but we are at least making the same assumption across all taxa, so it should be comparable in its error). We will add a section in the Mechanical Advantage and Functional Indices discussion calling for further research into the mechanics of cranial kinesis so future mechanical advantage work in birds can take this matter into account.

• Reviewer 3, on skull reconstruction: This issue is partly addressed in the Bohaiornithid Skull Reconstruction section, though we agree that adding more mentions of it in the MA and FEA Discussion sections and the Bohaiornithid Ecology and Evolution sections will benefit the manuscript. Most notably Shenqiornis and Sulcavis have similar ecological interpretations, but much of the Shenqiornis skull reconstruction uses Sulcavis bones. Longusunguis is the only other taxon which takes more than two bones from a different taxon, and in this case all but the quadrate are not used in any quanitative measurements. We have ensured that the skull reconstructions presented in Figure 2 show what portions of the skull come from what specimen so that as new material is discovered and phylogenetic relationships are updated it will be clear to future readers which parts of reconstructions will need to be updated.

• Reviewer 3, on data availability: All data including FEA models and raw measurement data are included in the same repository as the scripts, which we will make clear in the manuscript. Good catch on the data link being dead, we will publish it now.

As a final note, it was brought to our attention by another colleague that the original manuscript’s ancestral state reconstrction lacked an outgroup. An updated reconstruction using Sapeornis as an outgroup will be included in the revised manuscript. The addition of the outgroup does not change any conclusions of the manuscript.

We once again thank our reviewers for their valuable feedback and will submit a revised version of this manuscript for publication shortly. Please let us know if you have any additional comments after reading our response that we can take onboard in our revision.

References

Artuso, C., C. S. Houston, D. G. Smith and C. Rohner (2020). Great Horned Owl (Bubo virginianus), version 1.0. Birds of the World. A. F. Poole. Ithaca, NY, USA, Cornell Lab of Ornithology.

Atterholt, J., A. W. Poust, G. M. Erickson and J. K. O'Connor (2021). "Intraskeletal osteohistovariability reveals complex growth strategies in a Late Cretaceous enantiornithine." Frontiers in Earth Science 9: 640220.

Corbin, C. E., L. K. Lowenberger and B. L. Gray (2015). "Linkage and trade‐off in trophic morphology and behavioural performance of birds." Functional ecology 29(6): 808-815.

Hone, D. W. E., A. A. Farke and M. J. Wedel (2016). "Ontogeny and the fossil record: what, if anything, is an adult dinosaur?" Biology letters 12(2): 20150947.

Hu, H. and J. K. O'Connor (2017). "First species of Enantiornithes from Sihedang elucidates skeletal development in Early Cretaceous enantiornithines." Journal of Systematic Palaeontology 15(11): 909-926.

Miller, C. V., M. Pittman, X. Wang, X. Zheng and J. A. Bright (2023). "Quantitative investigation of Mesozoic toothed birds (Pengornithidae) diet reveals earliest evidence of macrocarnivory in birds." iScience 26(3): 106211.

Olsen, A. M. (2017). "Feeding ecology is the primary driver of beak shape diversification in waterfowl." Functional Ecology 31(10): 1985-1995.

Schnell, J. H. (2020). Common Black Hawk (Buteogallus anthracinus), version 1.0. Birds of the World. A. F. Poole and F. B. Gill. Ithaca, NY, USA, Cornell Lab of Ornithology.