The Largest Mesosaurs Ever Known: Evidence from Scanty Records

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper is an interesting addition to mesosaur research, but there are several crucial points that need improvement before it can be published. Below is a short summary of my main concerns, but my specific comments can be found in the attached PDF:

1. Discussion of other hypotheses:
The authors largely overlook the body of mesosaur research published outside of their working group. Several scientifically supported hypotheses on mesosaur phylogeny, histology, ecology, and ontogeny challenge the authors' interpretations and should be mentioned and discussed accordingly. In particular, the data presented in Verrière & Fröbisch 2022 are complementary to those used in the manuscript and would be an interesting addition or point of comparison for several aspects of the study.

2.  State of preservation of the specimens:
I am very skeptical about how informative the presented fossils of mesosaur skulls are. To me, it is very likely that parts of them are sedimentary artifacts left after the original bone was destroyed, which is not a good basis for a morphometric approach. The authors should address this and temper their conclusions accordingly. Be careful of overinterpretation!

3. Superficial statistics and modeling:
The statistics used in the paper follow standard practices that should be corrected. For example, the authors should at least log their measurements and compute confidence intervals in their figures and models. I also feel that the authors do not provide a satisfactory investigation of Carroll's hypothesis for body size evolution in the early amniote, although this would add impact to their paper. I strongly suggest that they include an ancestral trait reconstruction in their manuscript to address this issue.

4. Poor referencing and presentation of the scientific context:
The introduction is rather superficial and the scientific context of their study is not properly presented. Many points need additional citations.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Overall, the quality of the author's English is adequate, with some minor grammatical errors. However, sentence construction is sometimes convoluted, jumbled, and difficult to follow. This detracts from the article's purpose and makes for a confusing read. I would encourage the authors to seek the assistance of a native English speaker to help them improve their text.

Author Response

Reviewer 1

 Comments and Suggestions for Authors

This paper is an interesting addition to mesosaur research, but there are several crucial points that need improvement before it can be published. Below is a short summary of my main concerns, but my specific comments can be found in the attached PDF:

We thank the reviewer for the kind comments; we will answer to all the referee’s comments below each one in bold cursive format.

1. Discussion of other hypotheses:
   The authors largely overlook the body of mesosaur research published outside of their working group. Several scientifically supported hypotheses on mesosaur phylogeny, histology, ecology, and ontogeny challenge the authors' interpretations and should be mentioned and discussed accordingly.

 We respectfully disagree with the reviewer’s comment as it is our standard practice to reference all relevant studies pertaining to the topic we are addressing in our manuscripts, encompassing both recent works and significant earlier contributions from respected scientists. Regarding mesosaur phylogeny, we always refer to the two more recent hypotheses, even when we have already noted our disagreement with some results. When two of the authors of the manuscript under review analyzed the mesosaur ontogeny in our previous recently published papers (i.e., Piñeiro et al. 2016; Nuñez Demarco et al., 2018, 2022) we referred to the work not only from authors that specifically studied mesosaur ontogenetic stages based on limb development (e.g., Caldwell, 1995,1996) but we also revised a long list of related researching references (92 in Piñeiro et al., 2016; 127 in Nuñez Demarco et al., 2019 and 150 in Nuñez Demarco et al., 2022), including contributions from Fröbisch, Modesto, Klein, MacDougall and others.

 In particular, the data presented in Verrière & Fröbisch 2022 are complementary to those used in the manuscript and would be an interesting addition or point of comparison for several aspects of the study.

We want to clarify to the reviewer that the work of Verrière & Fröbisch (2022) was already cited in our manuscript, but note that it does not significantly contribute to the study we propose by the following reasons:

• Almost every measurements in Verrière and Fröbisch (2022) was exclusively compared to the average size of dorsal vertebrae (ASDV) with the authors justifying this choice by suggesting that ASDV is a good proxy for determining overall Mesosaurus size (Verrière and Fröbisch 2022, p4). However, this proxy used by Verrière and Fröbisch (2022) is not optimal, given that it was demonstrated that ASDV has been shown to overestimates Mesosaurus length by including only the largest and more variable vertebrae in Mesosaurus vertebral column (Núñez Demarco et al., 2018). Concequently, most of the isometric relationships found by Verrière and Fröbisch (2022) are inapplicable, as they are based on an inappropriate proxy and cannot reliably estimate Mesosaurus size. Moreover, the allometric relationships studied by these authors are largely tied to this proxy, complicating their interpretation; at most, these findings suggest the existence of allometric relationships specifically with respect to the dorsal region. This does not refute or contradict previous findings.
• In contrast, the isometric relationships used in the Núñez Demarco et al. (2022) study were calculated using the average of all vertebrae, including cervicals and caudals, for a more accurate estimate, as recommended by Núñez Demarco et al. (2018, pp. 4–6).
• Of the 23 relationships analyzed by Verrière and Fröbisch (2022), only 4 are not linked to the ASDV; of these, only one is relevant to the present study: the skull length vs. postorbital length ratio (a measurement in which both research groups found allometry). Given the substantial amount of data collected by these authors, we decided to include their data to recalculate the relationships with greater precision, and the results are presented in the Figure 11. Additionally, we analyzed the relationship between orbit size and postorbital length ratio;a parameter previously studied by Núñez Demarco et al. (2022) but not by Verrière and Fröbisch (2022).

 State of preservation of the specimens:

2. I am very skeptical about how informative the presented fossils of mesosaur skulls are. To me, it is very likely that parts of them are sedimentary artifacts left after the original bone was destroyed, which is not a good basis for a morphometric approach. The authors should address this and temper their conclusions accordingly. Be careful of overinterpretation!

We assure to the reviewer that we know very well the morphology of the mesosaur skull and although the reviewer is right in that some regions of the skulls are badly preserved, as we already noted in the descriptions of the materials, the measurements for the morphometric studies were done on structures that can be well delimited, such as the orbital length and width and the width of the posterior border of the skull which can be taken with confidence based on the measurement of the left or right half accordingly to which of the two described skulls you are taking into account.

3. Superficial statistics and modeling:
   The statistics used in the paper follow standard practices that should be corrected. For example, the authors should at least log their measurements and compute confidence intervals in their figures and models.

Unfortunately, further statistical analyses are not feasible, as we only have five new measurements. While these are highly significant due to their substantial difference from the bulk of the mesosaur population, they are unfortunately insufficient for any robust statistical analysis.

On the appropriateness of logarithmic transformations for the data:

            As Feng et al. (2014) have demonstrated, if the data already have a normal distribution, the log transformation can produce a non-normal distribution and even increased variability. In our case, the mesosaur data have a normal distribution and then log-transformation is not recommended.

As we established in Núñez Demarco et al. (2018) following the recommendations of Feng et al. (2014), all the linear plots presented in this work use non-logarithmic data, but each plot is linked to the values calculated with logarithmic data. This also has the advantage that data and plots can be directly analyzed and interpreted (Feng et al. 2014), especially if we aim for our data to be useful for reconstructing the size of Mesosaurus. Moreover, since the data are isometric according to our results, the correct approach is to avoid applying logarithmic transformations.

We have added all this information into the Methods section of our manuscript.

Regarding confidence intervals

All data and calculations presented in this manuscript included confidence intervals and error margins. We added previously omitted intervals into the revised version of the manuscript.

I also feel that the authors do not provide a satisfactory investigation of Carroll's hypothesis for body size evolution in the early amniote, although this would add impact to their paper. I strongly suggest that they include an ancestral trait reconstruction in their manuscript to address this issue.
Indeed, we considered including an optimization of the character (size average) based on the two main tree topologies; however, we were not convinced that our paper is the right place for such a study. That is because while we have new data for the maximal size of mesosaurs, but this is not the same as average size, and there is insufficient information for this character in the other taxa examined. Therefore, as the reviewer 2 commented, we have here a very uncommon situation with mesosaurs because both the early juvenile and the mature average size of a species is virtually unknown in most Paleozoic amniote taxa. Indeed, we are here describing an amazing case within mesosaurs, which perhaps will prompt a revision of previous statements about average size in other early amniotes.

To make an optimization of the character "average size" of the taxa included in the available matrices we would need to know the average size of all the OTUs, meaning that we should have minimum two or three individuals measured for juvenile and mature individuals, a condition that is not meet is for many of the taxa in such previous attempts (e.g., Laurin 2004). Therefore, we think that the data that we obtained for Mesosaurus are not comparable with what is known for other early amniotes, on the matter of size average including early juvenile and mature individuals. Therefore, the study that the reviewer is requesting is likely a comparison between heterogeneous, not comparable data, and thus, we will prefer to not include it. Now, we think that we can keep our sections in the manuscript about body size evolution and about the Carroll's hypothesis of the small size of early amniotes because we are introducing new evidence that support the theory that not all early amniotes were as small as argued by Carroll (considering for instance, diadectomorphs as possible early amniotes or as part of the stem). As we also have calibrated a Late Carboniferous-Early Permian age for the deposits of the Mangrullo Formation containing mesosaurs through biostratigraphic relationships of other associated groups such Pygocephalomorpha, Hexapoda and Plants (see for instance Piñeiro, 2006; Piñeiro et al., 2012d; Calisto and Piñeiro, 2019) even to biotas present in Joggins, from where the earliest known amniotes were found. Thus, we have now a good argument to strengthen hypotheses that argue for a variable size in the earliest known amniotes.

Other colleagues who might have new relevant information and want to investigate the implications of mature mesosaur size shown in this manuscript should consider these problems. Our new data on mesosaurs show show that the size reached by a taxon during its life cannot be assessed without a large sample size of individuals.

4. Poor referencing and presentation of the scientific context:
   The introduction is rather superficial and the scientific context of their study is not properly presented. Many points need additional citations.

We followed the suggestions made by the reviewer in the annotated manuscript pdf.

peer-review-40662683.v1.pdf

Comments on the Quality of English Language

Overall, the quality of the author's English is adequate, with some minor grammatical errors. However, sentence construction is sometimes convoluted, jumbled, and difficult to follow. This detracts from the article's purpose and makes for a confusing read. I would encourage the authors to seek the assistance of a native English speaker to help them improve their text.

One of the authors is a native English speaker, although he also speaks in French, so we improved some of the paragraphs following his assistance and also the recommendations included into the annotated pdf file sent to us by the reviewer.

1-The Abstract is 323 words long. Please reduce it to 200 to comply with the journal's guidelines.

1-The Abstract is 323 words long. Please reduce it to 200 to comply with the journal's guidelines.

The abstract was modified accordingly

2-You should rephrase this. Yours is not the only study of mesosaur ontogeny.

The abstract was changed.

3-I disagree with this conclusion, or at least I think there are other possible interpretations. See my comments below.

The text was changed and improved.

4-Not sure if this type of consideration is relevant to the paper. This rule is widely referred to as "Cope's rule", for better or worse. For the sake of clarity, I would advise the authors to refer to it as such. Debating over the name of the rule is better suited for a different kind of paper.

There is no possible confusion in this part of our text because we explain that this is the tendency for body size to increase through time. The expression “Cope’s rule” was coined though a series of mistakes and misquotes from Cope and others. There is nothing clear about participating in historical disinformation! The debate is over as far as we are concerned (this is not Cope’s rule), and it has been argued by Polly (1998) and Bokma et al. (2016). We believe that these papers must be quoted because many colleagues seem to continue to think (erroneously) that Cope wrote on this topic, which is simply wrong.

5-What about other studies looking at early amniote groups? Do they support Carroll's claims? There are several studies on pelicosaur-grade synapsids and early reptiles that at provide elements of evidence regarding the evolution of size (see for instance Brocklehurst and Brink 2017, Brocklehurst et al. 2022). They should at least be mentioned, better yet, discussed here.

We added a brief discussion of these papers in a dedicated paragraph. Brocklehurst et al. 2022 was already cited elsewhere in the draft in a similar context.

6-This is awkwardly phrased. Did they "only confirm some" of Carroll's predictions or did they "invalidate most" of Carroll's predictions? Please rephrase more clearly

Done.

 7-Can you be more specific here? Not only is this sentence not very clear, but it also doesn't include Carroll's estimate for comparison.

This information had been provided a few paragraphs above, but we now repeat it here.

8-Do you have a reference for this?

The references were added

9-Can you specifically mention the result from this study that supports this claim? I could not find anything to support this in the paper you cite.

It was fixed.

The reference cited was correct and we added other one related to this issue.

10-The phrasing here is somewhat odd. Please try to rephrase more clearly.

The paragraph was removed.

11-Since this is the first time mesosaurs are mentioned in your paper, I feel like you should introduce the clade for the uninformed reader.

Done

12-This phylogenetic position has been widely debated in recent publications, and there is a large body of literature supporting the basal parareptile hypothesis. Why did you choose the basalmost sauropsid hypothesis over the other one as your working hypothesis? There doesn't seem to be any good justification for it. To be thorough, you should consider both hypotheses and compare the different results you get.

Our manuscript does not include a new analysis of body size evolution in early amniotes, so there are no results to compare. We only mention that it would be worth reassessing ancestral amniote body size, but that is a project in itself, so we do not cover this in the draft. However, we disagree that “There doesn't seem to be any good justification for it.” Good justifications for it were published long ago by Laurin & Reisz 1995 and much more recently, by Laurin & Piñeiro 2017, 2018.

13-Please also cite Gauthier et al. 1988, Tsuji et al. 2012, MacDougall et al 2016, 2017, 2018, Modesto et al. 2015 here

Done

14-This is not really true. Your results only show that their maximum possible body size has been underestimate, this does not affect their average body size.

This is incorrect. An average by definition depends on all data points, and ideally, these should be drawn randomly from the population of interest. This is extremely difficult to do (most often, impossible) for long-extinct taxa because of the notorious incompleteness of the fossil record. We never get the complete ontogeny. Most often, the earliest stages are missing, but in the case of mesosaurs, we demonstrate that the old adults were missing from previous samples. Of course, the impact on average body size will be less than on maximal body size, but both are impacted by our results. Note that our statement was deliberately cautious: we did not state that mesosaurs were on average twice the size as previously estimated! Nevertheless, we reformulated slightly our comment.

15-What about Modesto 1999, 2010, Canoville & Laurin 2010, MacDougall et al. 2020, Verrière & Fröbisch 2022? You need to also include and reevaluate hypotheses you have not formulated yourselves!

We cited these papers, as appropriate, in this passage. Note that MacDougall et al. 2020 implied a semi-aquatic lifestyle because caudal autotomy does not make sense in an aquatic environment, so we added it among the references suggesting that lifestyle. The others suggested a more aquatic lifestyle.

16-Could you include a schematic map of the locality?

We referred to other previously published paper.

17-Any reference for this figure?

The total vertebral length was calculated by the summatory of vertebrae of the different regions of the column following Piñeiro et al. 2021 and Nuñez Demarco et al. 2018.

18-To be thorough, you must use the different allometric relationships for mesosaurs proposed in the literature, such as Verrière and Fröbisch 2022, and compare both estimates.

This was widely explained above in this letter.

 

19-It would be informative to see a table of the original measurements and the reconstructed size proxies or skull length for each specimen, including the evaluated error.

The table can be found in Nuñez Demarco et al 2018 and 2022. The evaluated error for the measurements made in this study are already shown in the text and some in the figures.

20-This doesn't justify a complete dismissal of Verrière and Fröbisch's results, neither does it justify the fact that you do not include or compare your results to theirs. Even if you disagree with the accuracy of their size estimate, you should also include their data in your size distribution analysis and in the long bone analysis. This seems especially relevant for the specimens consisting of isolated vertebrae (Fig. 3).

This was widely explained above in this letter.

21-Besides, their conclusions and hypotheses directly contradict some of your propositions, and you should also address these claims thoroughly.

This was widely explained above in this letter.

22-I am very skeptical that these specimens can be used for morphometric analysis at all. FC-DPV 3622 is the very poor impression of a skull, and there are but bone fragments preserved. As for FC-DPV 3623, it is so damaged its shape does not even resemble that of a mesosaur skull. To me, it is entirely possible that the orange/brown areas be preservation artifacts, deposited after the bone was partially or completely dissolved. This seems very flimsy at best, and I would not base any measurements on these.

In comparison, your postcranial elements are much better preserved and form a much stronger basis to your arguments.

As explained above, measurements were taken in areas where they can be done with absolutely precision.

23-You should calculate a confidence interval for this estimate. The isolated vertebrae could correspond to the smallest dorsals as well as the largest, or a perfect average. Looking at your previous publications, you should have enough data to calculate how much each scenario would affect the estimate, which should give you an interval. Please include this here, as it may affect the extent to which these vertebrae are statistical outliers in terms of size.

That intervals are included in each related figure (9, 10, 11).

The text should have said "surpasses the population mean (5.4 mm) by more than five times its standard deviation (1.8 mm)."

The standard deviation mentioned in the line is the error of this estimate

We are unsure which other error interval the reviewer is referring to. It doesn’t make sense to add error intervals of less than 1 mm in the measurements on the graphs. All our calculations were consistently presented with their corresponding errors; we have corrected the omission of some intervals in the text.

Additionally, we included measurement errors for greater clarity, and these explanations on how the error intervals were calculated, and we have included confidence intervals in Figure 11 for greater clarity.

24-You should also compare your measurements with the results of Verrière & Fröbisch 2022, which is the most recent study on mesosaur allometry. Even if you don't agree with their approach, the measurements are worth including in your model (especially since you do not use a size proxy here).  Also, you should use logged data in your computation of allometric relationships, but see my comment of Fig. 9.

This issue was widely explained above in our responses to the reviewers.

25-Again, I am very skeptical of the dimensions of the "skulls". In my opinion, their state of preservation does not allow for accurate measurements because they do not accurately represent the original dimensions of the original bones, and you should be extra cautious when exploiting this data. Please at least include an estimate of the measurement's error and mention that these measurements are not taken on real bone or even natural molds but on bone "imprints".

The measurements are very easy to take for us because all the mesosaur skulls that we have found in Uruguay have the bones in perfect anatomical articulation and their morphology can be perfectly delimited, even when they are just impressions of partial preservation of the bones.

 26-This phrasing is a bit odd. Please cite Nunez Demarco 2022 instead, since you are relying on their findings.

We are not completely sure about this request from the reviewer, but the paragraph that is being referring, was changed.

27-How do you compute the error here?

The errors were calculated through regressions, taking into account measurement errors and errors in the linear adjustments, and in the interpretation (as you correctly mention on line 214 isolated vertebrae could correspond to the smallest dorsals as well as the largest, and this was considered).

28-What does one have to do with the other? Body proportion has nothing to do with size. Mice and whales both have long tails but don't have similar sizes. This should either be rephrased or removed entirely.

The paragraph that the reviewer is mentioning does not discuss proportion.

 29-This contradicts the results of several studies based on a wide range of evidence: morphology (Modesto 1999, 2010), histology (Canoville & Laurin 2010, Klein et al. 2019) and taphonomical (Verrière & Fröbisch 2022). You cannot simply discard it like that without addressing each argument proposed in these studies.

We are based our arguments in a very wide range of evidence: morphologic, taphonomic, ecological, physiological and biological. Therefore, we are not discharging, simplesment we dismiss because of lack of evidence.

and all the papers that the reviewer is mentioning are now included in the revised version of our manuscript.

30-Looking at your data (Fig. 7), it doesn't look like a gap so much as it does that you have found specimens that belong to the rightmost arm of the distribution. I would suggest that these are just exceptionally large specimens, not the norm for adult mesosaurs. Several specimens described in other studies are fully ossified and therefore mature but do not reach these extreme sizes. You need to address this.

But this is exactly what we point out. We now mention this possibility  

The reviewer is right and we strongly agree with this observation. That is the main reason to consider the findings described herein as extraordinary and thereby worthy of disclosure especially considering that the already described population is very far from the mean and shows significant dispersion.

These exceptionally large mesosaurs are not the norm, are exceptional in the mesosaur record. The individuals described herein represent the oldest individuals (and thus the largest ones), based on the fact that reptiles grow during all their life.

This precisely raises questions such as which could have been the size of other basal amniotes from which just one or two individuals are known.

 31-It would be also interesting to see if mesosaurs had a maximum size.

Yes, we agree, because if this is the maximum size for Mesosaurus, it would be interesting to know what could have been the maximal size of other basal amniotes from which just one or two individuals are known. However, in general, determining if growth was determinate rather than indeterminate requires histological data that we do not have on the large specimens. However, most early amniotes appear to have had indeterminate or semi-determinate growth, so there was probably no true maximum, although growth slowed drastically at a certain age.

32-Maybe these extreme specimens were just very old individuals that exceptionally reached advanced age and just kept growing.

 This is too much speculative, because we know that fossilization is not a common process and it depends on environmental and physic-chemical factors. Why do not accept the hypothesis that the big mesosaurs are just a few specimens found because they occupied a different niche where the conditions are not optimal for preservation?

33-If you want to advance the hypothesis of a gap in fossil record, you need to be able to dismiss this hypothesis. A histological analysis of some of your very large specimens would be informative.

We are not talking about a gap in the fossil record. Indeed, we have many specimens that are larger than the mean of mesosaurs found in Brazil and probably in South Africa and Namibia, but they died before to reach their maximum size, as is common to occur in most recent reptiles.

As for histological analyses, we do not have enough large and well-preserved specimens to cut, but it will be interesting to do in the future.

34-This implies that it may not apply to mesosaurs at all!

It just Implies that Bergman’s rule does not explain the large size that can be reached mesosaurs, because this rule only applies to endotherms!!!.

35-In allometry analysis, one should apply a logarithmic transformation to morphometric measurements to linearize growth patterns. There are several papers supporting this approach, for example Huxley 1932, Gould 1966, Schmidt-Nielsen 1984, more recently Pelabon et al. 2018. You should transform your data and apply the linear model to your logged data.

Also, please include error bars for your estimates.

We already have explained the reason for not to apply logarithmic transformations in this case. Please, see above. As we already explained, Feng et al. (2014) have demonstrated that if the data already have a normal distribution, the log transformation can produce a non-normal distribution and even increased variability. In our case, the mesosaur data have a normal distribution and then log-transformation is not recommended.

As we established in Núñez Demarco et al. (2018) following the recommendations of Feng et al. (2014), all the linear plots presented in this work use non-logarithmic data, but each plot is linked to the values calculated with logarithmic data. This also has the advantage that data and plots can be directly analyzed and interpreted (Feng et al. 2014), especially if we aim for our data to be useful for reconstructing the size of Mesosaurus. Moreover, since the data are isometric according to our results, the correct approach is to avoid applying logarithmic transformations

36-What did these studies conclude and what impact could the position of mesosaurs have on these results? You should elaborate on this.

New text about this issue, was added into the manuscript.

37-A transition is missing here. The link between Carroll's suggestion and mesosaurs need to be better explained.

New text on this issue was added to the manuscript.

38-The evidence you present for this "link" between mesosaurs and lepospondyls is by no mean rigorous. The elongated body is definitely a consequence of an aquatic lifestyle, since body elongation is a well-known trend in secondary aquatic adaptation. Moreover, the isometric/allometric growth measured in fossil taxa is not a valid phylogenetic character to the same extent that embryologic development patterns  can be. They show changes in proportions throughout growth, not structure formation sequences, and are therefore meaningless for phylogeny.

We are only referred to a paper that was already revised by peers and published. Isometry is for us a very important physiologic condition and must be considered for taxon relationships. However, it is reasonably to consider it as a convergent feature because it was not yet analyzed in the phylogenetic studies because the condition is unknown in most of the earliest amniote taxa.    

 

39-If anything, this is a textbook example of convergent evolution, two different clades having similar lifestyles and consequently acquiring similar body shapes, and hypothetically similar body proportion change patterns. If this has been published in Núñez Demarco et al 2022, I am baffled that it slipped through the cracks of scientific review.

We added new text of this particular concern of the reviewer.

40-This is a most incorrect way to discuss phylogeny and these claims should not be further elaborated upon unless better substantiated.

The same answer as in #38

41-The relevance of temporal fenestration in early amniote phylogeny is a highly debated topic. It seems to be highly variable in many clades, and it is likely that the old paradigm of the main amniotic lineages having their own mode of fenestration is at best an incomplete picture. See for example Haridy et al 2016 or Ford and Benson 2020.

As can be noted, we mentioned these relationships among mesosaurs and basal synapsids because it was previously proposed, but we clarified that is was not accepted because it never appears in the cladistic hypotheses. But we incorporated the work of Ford and Benson  recommended by the reviewer, into the paragraph. 

 

42-According to your own writings (line 89-93 and line 379-383), there are actually four phylogenetic hypotheses for the position of mesosaurs. Yet, you discuss only two here, including one (synapsids) that you describe yourself  as "largely dismissed by most researchers". Why? The parareptile scenario is the best supported in terms of number of studies, you have to its implications of size evolution, even if you personally disagree with it. On the other hand, I do not see the need to mention the synapsid hypothesis here, since it is not supported by phylogenetic evidence.

Because it was previously proposed, but we clarified that is was not accepted because it never appears in the cladistic hypotheses.

43-Same comment as for Fig. 9 here. Please use logarithmic scale and include error bars.

Same response as in #35

44-You do not need cite yourself since you are describing the classification defined in Soares 2003 and Oelofsen 1981.

We are confus by this comment. But while Bento-Soares and Oelofsen described taphonomic categories for the mesosaurs that they studied from Brazil and Africa, Piñeiro found other evidences for taphonomic hypotheses on mesosaur from Uruguay. The summary of the general mesosaur categories included In this manuscript was written in the mentioned papers, not by Bento Soares or Oelofsen, and refers to what is found in Uruguay. 

 45-I see your point but you should include some references linking monodominant assemblages to gregarious behavior here.

The references requested by the reviewer were included.

46-Also Rossmann & Maisch 1999, Verrière & Fröbisch 2022, Verrière et al. 2022.

The references required by the reviewer to add do not fit the end of the sentence because mesosaur fetus and hatchlings are not described in such papers, which describe juvenile mesosaurs, which is not the same. However, we added them at the end of the first part of the sentence, where juveniles are mentioned.

47-An hypothesis for preservation biases in mesosaurs was proposed in Verrière & Fröbisch 2022. They link it to life environment and distance to the coast. They also suggest that the oldest and largest individuals lived further away from the shore, mostly away from juveniles. You should discuss how this fits within your narrative and how your data may change this hypothesis.

We already discussed about this in the manuscript.

 

48-Again, a handful of big specimens do not constitute a missing ontogenetic stage. Most histological studies and ontogenetic studies on mesosaurs so far have identified smaller specimens as fully grown adults.  Looking at your own data, the very large specimens may very well be outliers.

This comment from the reviewer was responded above and was modified in the text of the manuscript.

49-This is too speculative and superficial, you need to substantiate your hypotheses with further evidence.

The evidence was explained a line below, but these are work hypotheses as those seen in all papers. We reformulated these alternative (partly mutually exclusive) hypotheses and clarified that they are hypotheses, not results. We will continue working on them looking for more information that can help to be closer to the reality, as we are used to do in paleontologic studies.

50-Moreover, how do your claims compare to those of other studies (for example, Verrière & Fröbisch 2022)?

We clarified that this is a possibility.

 

51-Would your scenario be compatible with the hypothesis that large mesosaurs lived preferentially in a more pelagic region? They could have been washed ashore by the very catastrophic events you are referring to. And what do you do of the histological evidence that support a more aquatic lifestyle (Klein et al. 2019, Canoville & Laurin 2010)?

The various adjustments to the text described above answer this concern, but we added to this paragraph a brief statement about the bone implications of bone microanatomical data described in these two papers.

52-What exactly do you mean here? It reads quite redundant.

It was modified for clarification.

53-How do you reconcile this type of taphonomy with your hypothesis that they lived on land?

This text was moved to the materials and methods section, as recommended by one of the reviewers, and concerning what is asking here, we do not propose that mesosaurs had a terrestrial life style, but that they were semi-aquatic reptiles that mature adults spent more time on land than juveniles. Besides, the incrustation of salts is produced after their death.

 

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript may be considered for publication as long as the suggested major modifications are made.

The manuscript presents relevant information about the size distribution in mesosaur populations. The authors describe some very large mesosaur bones that belong to older specimens and discuss the absence of big individuals in the mesosaur’s paleontological record. However, the low frequency of very large (or older) individuals of ectothermic organisms is what would be expected in the population and paleontological register. Even today, few individuals reach the maximum biologically permitted size as occurs with lizards, crocodiles, etc. Due to the taphonomic filter, the scarcity of very old individuals would be as justifiable as that of very young individuals.

            Thus, this article contributes by complementing a growth curve for mesosaurs; however, the data presented do not justify or support the hypothesis that very large individuals would occupy a more terrestrial environment. The authors overvalue large individuals' absence or rarity, justifying the paleontological record's lack of information within the terrestrial environment.

I am not convinced that the bones are covered by gypsum but rather by iron sulfate resulting from pyrite oxidation. One possible interpretation of the SEM-EDS analysis is that it is not gypsum but iron sulphate (FeSO4), a product of pyrite oxidation (iron sulphide). The higher Ca content comes from hydroxyapatite - Ca10(PO4)6(OH)2, which belongs to the bone composition (see Fig. 7E). Gypsum formation (if it exists in the sample) may be secondary or more recent due to the action of sulfuric acid on the calcium in the bones. It seems that the rock surrounding the bones does not show this gypsum deposition. The authors' incorrect interpretations of the gypsum presence may enhance the idea of environmental desiccation.

Drier seasonal events seem to have affected all sub-environments (lagoons) of the WhiteHill-Irati environment and do not necessarily represent a trend towards desertification. Episodic evaporation from these environments precipitated gypsum and other salts throughout the WhiteHill-Irati environment; however, I am unaware of any precipitation of gypsum exclusively on the bones of mesosaurs.

The authors do not stratigraphically locate the studied material, whether it was collected in the upper levels or not. Without this stratigraphic control, it is impossible to support the hypothesis that the mesosaurs may have become extinct due to environmental desiccation or decay of volcanic ash.

Therefore, I suggest that different parts of the text addressing these issues should be readjusted.

 

Detailed revised information

LINE

36) Abstract - The comparison with marine iguanas was not considered in the manuscript's text.

43) The key words "small size" and "large size" are very generic. I suggest “gigantism” and “ Mangullo Formation”.

109) What is the stratigraphic position of the Mangrullo Formation outcrop where the studied specimens were collected?

109-113) Remove Lines 109 to 113 from Material and Methods moving to Conclusions or Description.

119) I suggest: incomplete mesosaur skull...

It doesn't seem fragmented, but rather compressed, leading to a gentle disarticulation.

124) Gervais 1865 - Not found in references.

127) "large vertebra" is not informative; the measurements are already in the Discussion. Or include the dimensions here in the Descriptions

132) "A large" is not informative; the measurements are already in the Discussion. Or include the dimensions here in the Descriptions section.

134) Same as line 132.

135) The preservation condition is not good. It is impossible to visualize a trochlea or capitulum, perhaps use "The region corresponding to the trochlea and capitulum are proportionally larger than in younger individuals". But it seems to me that the length also follows this proportion. If ossification is progressive, the bones of older individuals are expected to be larger (more developed). In this case, I suggest not using the term "more developed".I recommend restructuring the sentence.

171) The citation Piñeiro et al. 2021 - Not found in references.

203-205) Figure 1 Caption - Specimens E and G (Fig. 1) are too incomplete to be recognized as part of a mesosaur skull.

The structure interpreted as parietal foramen (pf) may correspond to an irregularity in the rock surface’s relief or a small concretion. Note that in Figure 1 (A and C) a similar circular feature occurs in the right posterior region of the area interpreted as frontal bone. I don't know if this specimen contributes to the manuscript, perhaps it would be better to remove it.

229) Figure 1 Caption - Increase, not “evolution”.

In the same caption, indicate whether the skulls are in ventral or dorsal view.

257) The citation Piñeiro et al. 2021 - Not found in references.

263) Figure 3 Caption - Remove the numbering above the scale bar in figures B and D.

In figures B, D and F, what is the scale, 5mm or 10mm, as indicated in the caption?

265) What does "normal" mean? Wouldn't it be "average population size"?

275) Figure 4 Caption - Indicate whether it is the right or left bone.

294) Figure 5 Caption - Indicate whether it is a right or left bone.

302) The citation Piñeiro et al. 2021 - Not found in references.

 

314) Figure 6 Caption - Indicate the views: dorsal or ventral view?

333) Figure 7 Caption - It seems that in the SEM-EDS analysis, Fe has a higher value than Ca; why the presence of iron sulphate (FeSO4), a product of the oxidation of pyrite (iron sulphide) was not considered? The iron sulphate can crystallize in fine needles. Note that the hydroxyapatite composition of the original bone - Ca10(PO4)6(OH)2 - would explain the proportion of Ca. The formation of gypsum (if it exists in the sample) may be secondary, or more recent, due to the action of sulfuric acid originating from the oxidation of the sulphide.

348-350) Figure 8 Caption - This graph shows a normal distribution for vertebrae, with the largest and smallest specimens at either end of the curve. Why not consider the specimen as just an outlier in a normal population distribution curve?

379) The citation Romer & Price, 1940 - not found in references.

407) Figure 10 Caption - The "orbital size" can and should present a considerable reading error. The skull in life is curved (convex) and is usually compressed during fossilization, significantly increasing the original dimensions. This situation is particular to each specimen. This can lead to an overestimation of the dimensions of the orbital cavity. Compression usually affects the temporal region of the skull more. This may explain the extreme position of specimens 1C and 1A in the analysis.

Perhaps the accurate estimate of the skull size places it much closer to the expected mean distribution, but this seems to have not been considered.

419) The citation Piñeiro et al 2021 - Not found in references.

440) “recently”-  Not so recent. I would suggest removing the word

445) This fact does not justify parental care; this is speculative. It may simply be due to taphonomic conditions.

473) It should be noted that storm waves can suspend sediments, reworking older strata and concentrating material into a single layer by density and gravity. Usually, these layers with bones or crustaceans are deposited over massive layers (storm rock or "tempestites"). This is the product of storms, not a mortality episode caused by volcanic ash deposition.

485-487) Although gypsum deposits occur throughout the WhiteHill-Irati Sea environment, I think the interpretation of gypsum deposition on the surface of the bones of the Mangrullo Formation is mistaken (I do not believe it is calcium sulfate, as previously considered). I suggest to modify this interpretation.

514) Not variability – I suggest “variation” or “amplitude”.

521) I do not believe that this explains the extinction of the mesosaurs. Several ash deposition cycles occurred in other locations, e.g. in Brazil, and mesosaur fossils are still found above them. The filling and subsequent drowning of the basin by the entry of seawater at the end of this time continued by the deposition of the Collingham Formation in South Africa and the Serra Alta Formation in Brazil, and its consequent environmental shifts may have promoted drastic changes in the environment in which the mesosaurs lived. If this is very localized to the Mangrullo lagoon,it would have to be supported by geological regional evidence on the Mangrullo Formation, indicating the absence of mesosaurs at levels above the ash deposition. In this case, one cannot extrapolate to a continental environment with more than 1,500.000 km².

The Mangrullo Formation is a small part of the WhiteHill-Irati Sea System and is at the southernmost point of the mesosaur occurrence in South America. It is not recommended that this interpretation be extrapolated to the entire system. I suggest modifying the sentence.

Author Response

Comments and Suggestions for Authors

The manuscript may be considered for publication as long as the suggested major modifications are made.

The manuscript presents relevant information about the size distribution in mesosaur populations. The authors describe some very large mesosaur bones that belong to older specimens and discuss the absence of big individuals in the mesosaur’s paleontological record. However, the low frequency of very large (or older) individuals of ectothermic organisms is what would be expected in the population and paleontological register. Even today, few individuals reach the maximum biologically permitted size as occurs with lizards, crocodiles, etc. Due to the taphonomic filter, the scarcity of very old individuals would be as justifiable as that of very young individuals.

The reviewer is right and we strongly agree with this observation. That is the main reason to consider the findings described herein as extraordinary and thereby worthy of disclosure especially considering that the already described population is very far from the mean and shows significant dispersion.    

            Thus, this article contributes by complementing a growth curve for mesosaurs; however, the data presented do not justify or support the hypothesis that very large individuals would occupy a more terrestrial environment. The authors overvalue large individuals' absence or rarity, justifying the paleontological record's lack of information within the terrestrial environment.

We value the presence of the very large individuals described herein because of the reason that the reviewer correctly described above, but the hypothesis that they may have had a more terrestrial lifestyle as adults than as juveniles is presented as a hypothesis (not a fact) and is justified by solid evidence that was partially described in previous works (e.g., vertebral patterns, leg ratio proportions) and that the features of the new material presented herein strengthen somewhat (e.g., bad preservation, in comparison with the juvenile and younger adults, and the fact that these mature individuals are found in shallow, nearshore deposits, based on the presence of well preserved plant and insect remains). We added additional comments to strengthen our point, although we consider the possibility that these individuals lived farther from the shore, as suggested by some previous studies, we do not found evidence to prove it which strongly support the nearshore hypothesis.

I am not convinced that the bones are covered by gypsum but rather by iron sulfate resulting from pyrite oxidation. One possible interpretation of the SEM-EDS analysis is that it is not gypsum but iron sulphate (FeSO4), a product of pyrite oxidation (iron sulphide). The higher Ca content comes from hydroxyapatite - Ca10(PO4)6(OH)2, which belongs to the bone composition (see Fig. 7E).

Pyrite is very common in some levels, it is plausible that iron sulphite (FeS) may be present in the sediments or fossils; however, the mineral found in these deposits is gypsum. It is noteworthy that one of the authors of our paper (PND) is a geologist and a geophysicist specialized in mineralogy and petrology. Moreover, geochemical analyses of Araújo et al. (2001) and sedimentological profiles by Xavier et al (2018) describe evaporitic intrasedimentary growth of gypsum crystals and dolomitization in the Irati Formation of Brazil and Oelofsen (1981) also describes gypsum and halite crystals in the dolomitic and shale deposits of the Whitehill and Irati formations. All the mentioned studies agree in that the presence of layers if gypsum crystals in these units is indicative of subaerial exposure, hypersaline shallow water bodies and evaporite conditions (see also Hachiro 2000; Holanda et al., 2019; Dos Anjos et al. 2010). We found gypsum crystals associated to the bones described herein in the Mangrullo Formation, which was described and figured by Piñeiro et al. (2012b) as a Konservat-Lagerstätte.

Also, Xavier et al. (2018) indicates that evaporitic sediments can be diagenetic and form within already deposited sedimentary levels. This raises the possibility that the evaporitic minerals are not strictly contemporaneous with the fossils. However, the evaporitic deposition we discuss herein corresponds to specific desiccation events particularly affecting the last sediments deposited, which is also described by many works for the Brazilian an African areas of the Paraná and the Karoo basins, and we added this evidence to the revised version of our manuscript. Concerning what we observe in the Mangrullo Formation, the thinness of the evaporitic layers and their rarity are also indicators that these events were sporadic and particularly affected the last sedimentary layers where the fossils were found. If diagenetic, the phenomenon would have been widespread and persistent over time and it would have affected multiple layers or left a significant deposit, which is not the case for what we see in the Uruguayan Mangrullo Formation. This implies two things: 1) the deposition and the evaporitic events are close in time, and thus mesosaurs lived under such evaporitic conditions and were affected by them judging for the high amount of skeletal remains associated to the gypsum crystal levels (see the new Figure 8), and 2) the depositional environment already involved a shallow water body near the coast. This connects these mesosaurs to a coastal environment rather than a deep-sea environment and under this hypothesis we can suggest that mesosaurs found in Uruguay are associated with more coastal than open marine environments.

Likewise, the Phanerozoic sediments of the Uruguayan Norte Basin (=Paraná and Karoo basins) are heavily affected by secondary ferrification and silification due the Gondwana breakdown (Piñeiro et al. 2012; Panario et al 2014), coating the crystals with iron oxide films.

Gypsum formation (if it exists in the sample) may be secondary or more recent due to the action of sulfuric acid on the calcium in the bones. It seems that the rock surrounding the bones does not show this gypsum deposition. The authors' incorrect interpretations of the gypsum presence may enhance the idea of environmental desiccation.

Correct, the mineralization is not solely and exclusively associated with the bones; however, part of the samples shows intergrowth of gypsum crystals on the surface and within cavities.

Drier seasonal events seem to have affected all sub-environments (lagoons) of the WhiteHill-Irati environment and do not necessarily represent a trend towards desertification. Episodic evaporation from these environments precipitated gypsum and other salts throughout the WhiteHill-Irati environment; however, I am unaware of any precipitation of gypsum exclusively on the bones of mesosaurs.

We corrected these expressions and improved the text to clarify this matter as well as we included a stratigraphic profile (New Figure 8) showing the position of large mesosaurs and the gypsum crystal levels. Other studies on mesosaurs rarely locate them stratigraphically, as the samples generally come from isolated quarries where the material is mixed. By fortune we have a complete sequence of the Mangrullo Formation and can recognize the different lithologies from which the samples originate. Considering the mineralogy associated with the fossils, we can also determine not the specific level, but the sedimentary environment in which the organisms lived.

The authors do not stratigraphically locate the studied material, whether it was collected in the upper levels or not. Without this stratigraphic control, it is impossible to support the hypothesis that the mesosaurs may have become extinct due to environmental desiccation or decay of volcanic ash.

We added a new figure (Figure 8 NEW) modified from Piñeiro et al., 2012, fig. 6 to show the stratigraphic levels where the herein described materials, were found.

Therefore, I suggest that different parts of the text addressing these issues should be readjusted.

As we explained, what we have interpreted in this manuscript is related to the implantation of adverse environmental and climatic conditions, particularly associated to the precipitation of brines in warm water bodies when evaporation exceeded the water inflow. All these hypotheses were already supported by evidence presented in previous papers (e.g., Piñeiro et al., 2012b, Nuñez Demarco et al. 2018; 2022 for Uruguay; Oelofsen 1987 for Africa; Araújo 2001, Xavier et al 2018, Hachiro 2000; Holanda et al 2019; Dos Anjos et al. 2010; Bastos et al., 2021, for Brazil), and the conditions in the lagoon should have been difficult, mainly for aquatic or semiaquatic animals. As a consequence, we described the mesosaur and pygocephalomorph environment as a hypersaline lagoon for the last moment that mesosaurs colonized the Norte Basin. The environment was affected by the installation of climatic arid and dry conditions and experimented successive drougts (Piñeiro et al., 2012b). The stratigraphic profile (modified from Piñeiro et al., 2012b, fig. 6) that we added illustrates the general mesosaur and pygocephalomorph distribution as well as the position of the herein described fossils thorough the Mangrullo Formation of Uruguay and the episodes of high mortality associated to the precipitation of the gypsum crystals and their disappearance before the end of the evaporite deposits.

 Detailed revised information

LINE

36) Abstract - The comparison with marine iguanas was not considered in the manuscript's text.

The reviewer is right, indeed we had to remove this from the abstract to reduce it to 250 words.

43) The key words "small size" and "large size" are very generic. I suggest “gigantism” and “ Mangullo Formation”.

Done and thanks!

109) What is the stratigraphic position of the Mangrullo Formation outcrop where the studied specimens were collected?

We added a stratigraphic profile (modified from Piñeiro et al., 2012b, fig. 6) to show the position of the described fossils and the general distribution of the mesosaur remains thorough the Mangrullo Formation of Uruguay

109-113) Remove Lines 109 to 113 from Material and Methods moving to Conclusions or Description.

Done.

119) I suggest: incomplete mesosaur skull...

It doesn't seem fragmented, but rather compressed, leading to a gentle disarticulation.

Yes, right. The change was done.

124) Gervais 1865 - Not found in references.

It was added.

127) "large vertebra" is not informative; the measurements are already in the Discussion. Or include the dimensions here in the Descriptions

Yes, the issue was solved.

132) "A large" is not informative; the measurements are already in the Discussion. Or include the dimensions here in the Descriptions section.

Fixed.

134) Same as line 132.

Done.

135) The preservation condition is not good. It is impossible to visualize a trochlea or capitulum, perhaps use "The region corresponding to the trochlea and capitulum are proportionally larger than in younger individuals". But it seems to me that the length also follows this proportion. If ossification is progressive, the bones of older individuals are expected to be larger (more developed). In this case, I suggest not using the term "more developed".I recommend restructuring the sentence.

Okay, but indeed, the humerus of mesosaurs is the bone that experimented more morphological changes from the hatching to the adult stage and of course, the development of the entepicondyle and the foramen that pierces it in a young adult continue to increase, possibly at the same rate into the older and more mature individuals. The throclear and capitular regions are notable in comparison to the humerus in younger specimens.

171) The citation Piñeiro et al. 2021 - Not found in references.

The reference was added.

203-205) Figure 1 Caption - Specimens E and G (Fig. 1) are too incomplete to be recognized as part of a mesosaur skull.

This specimen corresponds to part and counterpart of an incomplete large skull preserving the postorbital cranial region enough clear. It is very common to find this region separated from the rest of the skull in the Mangrullo Formation. 

The structure interpreted as parietal foramen (pf) may correspond to an irregularity in the rock surface’s relief or a small concretion. Note that in Figure 1 (A and C) a similar circular feature occurs in the right posterior region of the area interpreted as frontal bone. I don't know if this specimen contributes to the manuscript, perhaps it would be better to remove it.

As we explained enough in the manuscript, these large mesosaur specimens are very badly preserved, and concerning the skull, only two specimens were found until now. Therefore, we prefer to retain this evidence in the manuscript; perhaps some minerals could have precipitated into the parietal foramen and distorted its outline. We added an interrogation sign for denoting uncertainty in the identification of the pf, but this is the position where this foramen is present in the mesosaur skull.  

229) Figure 1 Caption - Increase, not “evolution”.

In the same caption, indicate whether the skulls are in ventral or dorsal view.

Done.

257) The citation Piñeiro et al. 2021 - Not found in references.

Reference added.

263) Figure 3 Caption - Remove the numbering above the scale bar in figures B and D.

In figures B, D and F, what is the scale, 5mm or 10mm, as indicated in the caption?

Okay, all was fixed and clarified in the figure caption.

265) What does "normal" mean? Wouldn't it be "average population size"?

Modified as the reviewer suggested.

275) Figure 4 Caption - Indicate whether it is the right or left bone.

Done

294) Figure 5 Caption - Indicate whether it is a right or left bone.

Done.

302) The citation Piñeiro et al. 2021 - Not found in references.

 Reference added.

314) Figure 6 Caption - Indicate the views: dorsal or ventral view?

333) Figure 7 Caption - It seems that in the SEM-EDS analysis, Fe has a higher value than Ca; why the presence of iron sulphate (FeSO4), a product of the oxidation of pyrite (iron sulphide) was not considered? The iron sulphate can crystallize in fine needles. Note that the hydroxyapatite composition of the original bone - Ca10(PO4)6(OH)2 - would explain the proportion of Ca. The formation of gypsum (if it exists in the sample) may be secondary, or more recent, due to the action of sulfuric acid originating from the oxidation of the sulphide.

The high spike in the SEM analysis marks the presence of oxygen, not Fe. However, pyrite is very abundant in the Mangrullo shales and pyrite and gypsum are very frequently found together.

 348-350) Figure 8 Caption - This graph shows a normal distribution for vertebrae, with the largest and smallest specimens at either end of the curve. Why not consider the specimen as just an outlier in a normal population distribution curve?

Indeed, the large mature specimens are part of a normal population and represent possibly the oldest individuals denoting a continuous growth during life. The position of the youngers and the oldest at the ends in the curve reflects such a continuous growth.    

379) The citation Romer & Price, 1940 - not found in references.

Reference added.

407) Figure 10 Caption - The "orbital size" can and should present a considerable reading error. The skull in life is curved (convex) and is usually compressed during fossilization, significantly increasing the original dimensions. This situation is particular to each specimen. This can lead to an overestimation of the dimensions of the orbital cavity. Compression usually affects the temporal region of the skull more. This may explain the extreme position of specimens 1C and 1A in the analysis.

Perhaps the accurate estimate of the skull size places it much closer to the expected mean distribution, but this seems to have not been considered.

Given that in almost all mesosaurs that we measured the skull is compressed in a similar way, this problem does not affect significantly the resulting data, to the extent that we expect that it affected orbital length in the same way. Therefore, the giant mesosaurs are indeed more than three times larger than the average population size previously considered, and this conclusion is not coming from an overestimated measure.

419) The citation Piñeiro et al 2021 - Not found in references.

Reference added.

440) “recently”-  Not so recent. I would suggest removing the word

Okay, done.

445) This fact does not justify parental care; this is speculative. It may simply be due to taphonomic conditions.

This is probably not a taphonomic artefact, because we have a lot of specimens consisting in adult mesosaurs associate with bones that belong to hatchling individuals (see Piñeiro et al., 2012). However, speculation is a common in paleontology, particularly when determining behaviors of very ancient groups. However, aggregation is the most plausible reason to find such kind of association in mesosaurs. We modified the text accordingly.

473) It should be noted that storm waves can suspend sediments, reworking older strata and concentrating material into a single layer by density and gravity. Usually, these layers with bones or crustaceans are deposited over massive layers (storm rock or "tempestites"). This is the product of storms, not a mortality episode caused by volcanic ash deposition.

Yes, this can be for the cases where there is evidence of the action of storm waves and jumbled bones, which is absent in Mangrullo Formation, in which many articulated specimens are found. Instead we have bentonite layers covering the bones, thus indicating a case of high mortality by volcanic activity aggravated by the drought of the sea in the basin. We rephrased here to avoid confusing interpretations.

485-487) Although gypsum deposits occur throughout the WhiteHill-Irati Sea environment, I think the interpretation of gypsum deposition on the surface of the bones of the Mangrullo Formation is mistaken (I do not believe it is calcium sulfate, as previously considered). I suggest to modify this interpretation.

We have already widely answered to this reviewer concern (see above).

514) Not variability – I suggest “variation” or “amplitude”.

Okay, fixed.

521) I do not believe that this explains the extinction of the mesosaurs.

We clarified that our hypothesis is mainly suggested for what could have occurred with mesosaurs living in Uruguay, but considering the last research about the climatic and environmental changes that experimented the Irati-Whitehill sea, it can be expected that all mesosaurs were affected by these same catastrophic conditions.

Several ash deposition cycles occurred in other locations, e.g. in Brazil, and mesosaur fossils are still found above them.

Many Brazilian colleagues have informed to one of the authors (GP) that the stratigraphy of the Irati Formation is being revising in the last years but biostratigraphic relations are not completely known. Thus, we cannot be sure about what the reviewer is stating regarding mesosaur stratigraphy. However, we added the information obtained by several recent papers from the Iratí Formation that support our interpretations.  

The filling and subsequent drowning of the basin by the entry of seawater at the end of this time continued by the deposition of the Collingham Formation in South Africa and the Serra Alta Formation in Brazil, and its consequent environmental shifts may have promoted drastic changes in the environment in which the mesosaurs lived. If this is very localized to the Mangrullo lagoon,it would have to be supported by geological regional evidence on the Mangrullo Formation, indicating the absence of mesosaurs at levels above the ash deposition. In this case, one cannot extrapolate to a continental environment with more than 1,500.000 km².

For now, we are presenting data that come from Uruguay, which could be used as a reference for new stratigraphic studies in the areas of mesosaur distribution in Brazil and Africa. But we added relevant information from recent studies that corroborate our data from Uruguay.

The Mangrullo Formation is a small part of the WhiteHill-Irati Sea System and is at the southernmost point of the mesosaur occurrence in South America. It is not recommended that this interpretation be extrapolated to the entire system. I suggest modifying the sentence.

We did not extrapolate our interpretation to the entire system, on the contrary! Please, see that we noted that our hypothesis is for mesosaurs living in the Uruguayan region of the Parana and Karoo basins, and although the Mangrullo Fm. is just a “small part” of the Whitehill-Irati-Mangrullo-Huab sea, it is a big producer of complementary information, which is available for stratigraphic, biostratigraphic, paleoenvironmental, palebiologic, and paleoclimatic reconstructions in all these units. We modified the referred paragraph of our draft, adding more references to recent studies that even though our results are consistent with data from stratigraphic distribution and paleoenvironment indicators in correlative deposits of Brazil and Africa.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

First, I acknowledge that the authors have adequately addressed several of my previous comments, particularly in expanding the discussion of related works and including additional data. While I may not entirely agree with their approach to dismissing certain research, I recognize that this version provides better justification for their perspectives. The introduction and discussion sections have also been significantly improved, offering a more comprehensive overview of the current state of research. Most of my comments are therefore for minor revisions.

However, two critical issues remain in this manuscript that I believe require immediate attention:

1. I strongly disagree with the authors' rationale for not logging their data. While Feng et al. (2014) caution against unnecessary data transformations, this does not apply to the present case. Allometric relationships are inherently exponential (see Huxley 1932; Gould 1966; Schmidt-Nielsen 1984; Pelabon et al. 2018), making it essential to log-transform the data before performing linear regression analyses. Please revise your analyses accordingly, as the omission of this step could significantly affect your results.

2. The discussion on the phylogenetic position of mesosaurs in Section 4.3 continues to misrepresent the field. The notion of a "basal synapsid" hypothesis is simply inaccurate and has not been substantiated in the literature. I strongly urge the authors to either rephrase or remove references to this hypothesis, as its inclusion risks perpetuating confusion in future works. This point is particularly perplexing because it has little bearing on the manuscript’s results or primary conclusions. Resolving this issue would greatly enhance the clarity and credibility of the discussion.

I kindly request that the authors consider these points carefully, as they are crucial to improving the rigor of the manuscript.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The English language has been greatly improved in this new version of the manuscript. However, section 4.3 deserves special attention, as there is a noticeable shift in style and quality of the English there.

In general, I would also advise you to pay attention to your wording: keep a neutral and scientific style throughout, be respectful of previous authors when mentioning methods or results with which you disagree, and avoid self-congratulation.

Keeping this in mind will considerably improve this manuscript and any future ones.

Author Response

Reviewer’s comments on Page 4: Please provide the complete species name here and cite Gervais 1869.

Done, we have added  Gervais 1865.

Reviewer’s comments on Page 5: "Mesosauria" ist too unspecific (also I am not sure to what extent this is valid, as it is a monospecific clade). For clarity, please write Reptilia or Sauropsida here instead, as it would encompass most phylogenetic hypotheses for mesosaurs.

Done, we used Sauropsida

Reviewer’s comments on Page 6: establishment

Done, thanks.

Reviewer’s comment on Page 9: While Feng et al. 2014 is a valuable questioning of the standard practice of logging data to decrease variability and force a normal distribution, this is not really applicable to morphometric data. Regardless of the distribution of your data, logging morphometric data is motivated by the exponential nature of allometric relationships. Logging "flattens" allometric relationships to force a linear relationship between your variables, which can be modelled and tested more easily than a non-linear one.

Also, there is circular reasoning here. The fact that you measure an isometric relationship using non-corrected data cannot be the justification for not correcting data. All in all, I would still request that the data be logged here.

The explanation regarding the log-transformed data might have been unclear to the reviewer because to facilitate interpretation, the data was presented without logarithmic transformation. However, all analyses and calculations were conducted using log-transformed data. Notably, in this case, the isometric result (obtained using log-transformed data) confirms that the data are linear and can appropriately be represented in non-logarithmic form. At no point were logarithms excluded from the calculations; they were omitted solely for visual clarity in the data presentation. In Section 2.2.1, we explicitly stated: "the allometric constants are calculated from logarithmic data." We referred to Feng et al. for the visual presentation of the information. Figure 11 clearly indicates that the plots on the right were calculated using the isometric equation, which relies on logarithmic data, even though the plots at the left present the data in linear form. Moreover, the figure explicitly clarifies that the data was "calculated for log-transformed data."

Reviewer’s comment on Page 10: in this case, allometric

Yes, thanks. It was fixed.

Reviewer’s comment on Page 14: something's missing here

It was fixed.

Reviewer’s comment on Page 15 (1): This section should end with a calculation  of the former and new average body size for mesosaurs, before and after including your new specimens. This would be of particular interest in your discussion.

It is true that the mean remains almost unchanged because there are only a few recorded individuals from the older and giant population. However, this is precisely the core of the discussion. The presence of these individuals suggests the existence of an entire underrepresented etarian stage of the population—spanning the gap between these individuals and those smaller, already documented. Please note that we are not specifically referring to the giant mesosaurs but rather to everything missing in the gap between them and the individuals of less than 1 m, already recorded. This underrepresentation implies that the calculated mean is inaccurate, raising the question of whether the mean should be adjusted, potentially through weighting, to account for estimates of this missing population.

Reviewer’s comment on Page 15(2): Allometry is the correct term here. Isometry is a specific type of allometry.

Yes, but we are here testing the isometric type of allometry in mesosaurs.

 Reviewer’s comment on Page 15(3): True, but it could also be due to different methods is computing linear models, especially since they worked on logged data and you did not. I would also mention difference in methodology here.

Both groups used the same data, and both Núñez Demarco et al. and Verrière and Fröbisch worked with log-transformed data. The only difference lies in how the data is presented: we use a linear format, while Verrière and Fröbisch present it logarithmically. Some results are identical, but Verrière and Fröbisch did not account for error intervals, which reveal that many of their results are isometric. No amount of elaborate statistics can replace or override the simplest mathematical truths.

Reviewer’s comment on Page 15 (4). Verrière & Fröbisch 2022 statistically test allometry coefficients for significant difference with isometry, and the tests take into account confidence intervals in their calculation. This mention does not accurately represent their methods.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.

Reviewer’s comment on Page 16(1): Also cite MacDougall et al. 2020 here.

Done.

Reviewer’s comment on Page 16(1): Rephrase here, otherwise it reads as if the Mangrullo Formation is the only place with great preservation.

The Mangrullo Formation has indeed provided unique fossils, but also the Iratí Formation has been considered as a candidate for a konservat lagerstätte (Soares, 2003; Piñeiro et al., 2012b). 

 Reviewer’s comment on Page 18: While I appreciate the work you have done to reformulate this, I am not convinced of the relevance of this paragraph. First, where is the evidence that mesosaurs were larger than their "ancestors"? If you mean the common ancestor to amniotes in general, that would be too large a phylogenetic range to consider Bergmann's rule. Second, even if you consider this to be only pure speculation, it doesn't bring anything to your conclusions. Third, this type of pure speculation creates a risk that future studies might cite your paper as scientific evidence that Bergmann's rule might apply to ectotherms, which it clearly does not.

I recommend deleting this section altogether.

We prefer to keep this section because the Berman’s rule has been recently questioned as useful to apply only in ectotherm groups. Therefore, to support such an interpretation should be good to know more specific examples. Even though, we modified a little the text for clarification. 

Reviewer’s comment on Page 19: Although this is an improvement compared to the first version, i still don't think this paragraph is at best confusing, at worst misleading. You cannot put phylogenetic analyzes and a resemblance in growth patterns on an equal footing. I fully appreciate the value of highlighting a potential convergence between mesosaurs and recumbirostrans on this particular trait, but it doesn't have the same weight as a phylogenetic analysis, and convergence is no evidence of phylogenetic closeness.

In our opinion, phylogenetic approaches are important when are based on characters whose states are enough checked; so anatomical comparisons are necessary tools for analyze relationships for avoid to obtain unstable threes that change all the time. Therefore we prefer to keep this section as it is. We have already widely explained that we have based this paragraph in a peer-reviewed paper that consider our conclusions as possibly valid, requiring further studies. The present contribution is going in that line of research. 

 Reviewer’s comment on Page 20(1): Again, you are misrepresenting the state of mesosaur research here. Their position as basal Synapsida is not supported by any analysis, it has never been proposed. You are jumping from a character, that many authors consider too variable to be properly informative, to a systematic classification. This is not how it works. There are a lot of other morphological characters that support their reptilian affiliation other than temporal fenestration. Feel free to mention a convergence, but please delete claims that there is such thing as a proposed basal synapsid position for mesosaurs. This is simply not the case.

In our opinion, phylogenetic approaches are important when are based on characters whose states are enough checked; so anatomical comparisons are necessary tools for analyze relationships for avoid to obtain unstable threes that change all the time.

In this case, a possible relationship between mesosaurs and basal synapsids has been suggested by other authors previously, including Piñeiro (2002, 2004 and some preliminary contributions presented in scientific events). If these comparisons are wrong, other authors should report the mistake in their own papers. As we explained at the beginning of this article we have not yet convinced about what are the most probably mesosaur relationships, thus, it is valid to investigate all the possibilities. Also, we already explained the reasons for which we do not support a mesosaur-parareptile relationship in more than one previous papers, which are included as references in this contribution.. Therefore we prefer to keep this section with some modifications for clarification.

Reviewer’s comment on Page 20(2): Same comment as before: this is a misrepresentation of scientific results on early amniote phylogeny. There is no "proposed" basal synapsid position, just the (debated) presence of temporal fenestration.Moreover, and this hasn't changed from thee previous version, but I don't understand why you don't debate what your results would be like in the framework of the parareptile hypothesis. The scope of your paper is not phylogenetic relationships, why not discuss all proposed options (instead of focusing on the least supported one!) Please add this to your discussion.

Please see our previous answer.

Reviewer’s comment on Page 20(3): This section is oddly off-topic. I don't see how the discussion of mesosaur reproductive techniques really contributes to understanding average size in early amniotes. First, there is no need to detail the nature of the mesosaur embryos, citing the original publications will do. Second, when it comes to connecting both topics, which is done in the last three lines, you mention that "the eggs and hatchling [...] do not support such hypothesis" but do not detail why. Please shorten this section to what is relevant to your paper and clearly explain what your arguments are.

We have clarified why the eggs and hatchling mesosaurs do not support the Carroll’s scenario. For the other concern of the reviewer in order to reduce this section, we consider that it contains just a brief review of a paper that was published some years ago, in 2012, and it is a relevant information for the topic of this paper in the little modified current form.

Reviewer’s comment on Page 23: Everything from this point until the end of the section should be condensed and rephrased. You spend too much time detailing what analyzes you did not perform.

Respectfully, this is an explanation to your previous suggestion of to include a paragraph on the optimization of the character “average adult size” in early amniotes. Therefore, we believe that the explanation of why we decided not to include it is pertinent particularly to the readers that may want to see it.

Reviewer’s comment on Page 23: These paragraphs should also be rephrased. The potential reach of these few specimens and how they might affect future results seems overly speculative, especially since you do not provide a comparison between average mesosaur size with and without your new specimens. Alternatively, please include average body size for mesosaurs before and after the inclusion of your extreme specimens. My guess is that it would not vary that much if your sample is big enough. Regardless of the results, it would be interesting to discuss it.

The reviewer’s thoughts are right and we have provided an explanation for this in previous answers, but we will repeat it here for your convenience. “The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data.  

Reviewer’s comment on Page 25: these few specimens and how they might affect future results seems overly speculative, especially since you do not provide a comparison between average mesosaur size with and without your new specimens. Alternatively, please include average body size for mesosaurs before and after the inclusion of your extreme specimens. My guess is that it would not vary that much if your sample is big enough. Regardless of the results, it would be interesting to discuss it.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.

Reviewer’s comment on Page 26: Please avoid excessive self-citation.

The reviewer previously asked why we consider the Mangrullo Formation as a relevant litostratigraphic unit for the paleontological research, and we offered a response. We work in the Mangrullo Formation, so, its citation is unavoidable. 

Reviewer’s comment on Page 34: Rather than present two regressions based on data source, you should show the difference results produced before and after the inclusion of your extreme points. Treat the data as one body of evidence, it will be more informative.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize this, because the data from those specimens naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.

Reviewer 2 Report

Comments and Suggestions for Authors   For the authors   Given that the article presents relevant information on the size distribution in mesosaur populations, I recommend the manuscript for publication without any further corrections.

Author Response

Reviewer’s comments on Page 4: Please provide the complete species name here and cite Gervais 1869.

Done, we have added  Gervais 1865.

Reviewer’s comments on Page 5: "Mesosauria" ist too unspecific (also I am not sure to what extent this is valid, as it is a monospecific clade). For clarity, please write Reptilia or Sauropsida here instead, as it would encompass most phylogenetic hypotheses for mesosaurs.

Done, we used Sauropsida

 Reviewer’s comments on Page 6: establishment

Done, thanks.

Reviewer’s comment on Page 9: While Feng et al. 2014 is a valuable questioning of the standard practice of logging data to decrease variability and force a normal distribution, this is not really applicable to morphometric data. Regardless of the distribution of your data, logging morphometric data is motivated by the exponential nature of allometric relationships. Logging "flattens" allometric relationships to force a linear relationship between your variables, which can be modelled and tested more easily than a non-linear one.

Also, there is circular reasoning here. The fact that you measure an isometric relationship using non-corrected data cannot be the justification for not correcting data. All in all, I would still request that the data be logged here.

The explanation regarding the log-transformed data might have been unclear to the reviewer because to facilitate interpretation, the data was presented without logarithmic transformation. However, all analyses and calculations were conducted using log-transformed data. Notably, in this case, the isometric result (obtained using log-transformed data) confirms that the data are linear and can appropriately be represented in non-logarithmic form. At no point were logarithms excluded from the calculations; they were omitted solely for visual clarity in the data presentation. In Section 2.2.1, we explicitly stated: "the allometric constants are calculated from logarithmic data." We referred to Feng et al. for the visual presentation of the information. Figure 11 clearly indicates that the plots on the right were calculated using the isometric equation, which relies on logarithmic data, even though the plots at the left present the data in linear form. Moreover, the figure explicitly clarifies that the data was "calculated for log-transformed data."

Reviewer’s comment on Page 10: in this case, allometric

Yes, thanks. It was fixed.

Reviewer’s comment on Page 14: something's missing here

It was fixed.

Reviewer’s comment on Page 15 (1): This section should end with a calculation  of the former and new average body size for mesosaurs, before and after including your new specimens. This would be of particular interest in your discussion.

It is true that the mean remains almost unchanged because there are only a few recorded individuals from the older and giant population. However, this is precisely the core of the discussion. The presence of these individuals suggests the existence of an entire underrepresented etarian stage of the population—spanning the gap between these individuals and those smaller, already documented. Please note that we are not specifically referring to the giant mesosaurs but rather to everything missing in the gap between them and the individuals of less than 1 m, already recorded. This underrepresentation implies that the calculated mean is inaccurate, raising the question of whether the mean should be adjusted, potentially through weighting, to account for estimates of this missing population.

Reviewer’s comment on Page 15(2): Allometry is the correct term here. Isometry is a specific type of allometry.

Yes, but we are here testing the isometric type of allometry in mesosaurs.

Reviewer’s comment on Page 15(3): True, but it could also be due to different methods is computing linear models, especially since they worked on logged data and you did not. I would also mention difference in methodology here.

Both groups used the same data, and both Núñez Demarco et al. and Verrière and Fröbisch worked with log-transformed data. The only difference lies in how the data is presented: we use a linear format, while Verrière and Fröbisch present it logarithmically. Some results are identical, but Verrière and Fröbisch did not account for error intervals, which reveal that many of their results are isometric. No amount of elaborate statistics can replace or override the simplest mathematical truths.

Reviewer’s comment on Page 15 (4). Verrière & Fröbisch 2022 statistically test allometry coefficients for significant difference with isometry, and the tests take into account confidence intervals in their calculation. This mention does not accurately represent their methods.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.

Reviewer’s comment on Page 16(1): Also cite MacDougall et al. 2020 here.

Done.

Reviewer’s comment on Page 16(1): Rephrase here, otherwise it reads as if the Mangrullo Formation is the only place with great preservation.

The Mangrullo Formation has indeed provided unique fossils, but also the Iratí Formation has been considered as a candidate for a konservat lagerstätte (Soares, 2003; Piñeiro et al., 2012b). 

Reviewer’s comment on Page 18: While I appreciate the work you have done to reformulate this, I am not convinced of the relevance of this paragraph. First, where is the evidence that mesosaurs were larger than their "ancestors"? If you mean the common ancestor to amniotes in general, that would be too large a phylogenetic range to consider Bergmann's rule. Second, even if you consider this to be only pure speculation, it doesn't bring anything to your conclusions. Third, this type of pure speculation creates a risk that future studies might cite your paper as scientific evidence that Bergmann's rule might apply to ectotherms, which it clearly does not.

I recommend deleting this section altogether.

We prefer to keep this section because the Berman’s rule has been recently questioned as useful to apply only in ectotherm groups. Therefore, to support such an interpretation should be good to know more specific examples. Even though, we modified a little the text for clarification. 

Reviewer’s comment on Page 19: Although this is an improvement compared to the first version, i still don't think this paragraph is at best confusing, at worst misleading. You cannot put phylogenetic analyzes and a resemblance in growth patterns on an equal footing. I fully appreciate the value of highlighting a potential convergence between mesosaurs and recumbirostrans on this particular trait, but it doesn't have the same weight as a phylogenetic analysis, and convergence is no evidence of phylogenetic closeness.

In our opinion, phylogenetic approaches are important when are based on characters whose states are enough checked; so anatomical comparisons are necessary tools for analyze relationships for avoid to obtain unstable threes that change all the time. Therefore we prefer to keep this section as it is. We have already widely explained that we have based this paragraph in a peer-reviewed paper that consider our conclusions as possibly valid, requiring further studies. The present contribution is going in that line of research. 

Reviewer’s comment on Page 20(1): Again, you are misrepresenting the state of mesosaur research here. Their position as basal Synapsida is not supported by any analysis, it has never been proposed. You are jumping from a character, that many authors consider too variable to be properly informative, to a systematic classification. This is not how it works. There are a lot of other morphological characters that support their reptilian affiliation other than temporal fenestration. Feel free to mention a convergence, but please delete claims that there is such thing as a proposed basal synapsid position for mesosaurs. This is simply not the case.

In our opinion, phylogenetic approaches are important when are based on characters whose states are enough checked; so anatomical comparisons are necessary tools for analyze relationships for avoid to obtain unstable threes that change all the time.

In this case, a possible relationship between mesosaurs and basal synapsids has been suggested by other authors previously, including Piñeiro (2002, 2004 and some preliminary contributions presented in scientific events). If these comparisons are wrong, other authors should report the mistake in their own papers. As we explained at the beginning of this article we have not yet convinced about what are the most probably mesosaur relationships, thus, it is valid to investigate all the possibilities. Also, we already explained the reasons for which we do not support a mesosaur-parareptile relationship in more than one previous papers, which are included as references in this contribution.. Therefore we prefer to keep this section with some modifications for clarification.

 Reviewer’s comment on Page 20(2): Same comment as before: this is a misrepresentation of scientific results on early amniote phylogeny. There is no "proposed" basal synapsid position, just the (debated) presence of temporal fenestration.Moreover, and this hasn't changed from thee previous version, but I don't understand why you don't debate what your results would be like in the framework of the parareptile hypothesis. The scope of your paper is not phylogenetic relationships, why not discuss all proposed options (instead of focusing on the least supported one!) Please add this to your discussion.

Please see our previous answer.

Reviewer’s comment on Page 20(3): This section is oddly off-topic. I don't see how the discussion of mesosaur reproductive techniques really contributes to understanding average size in early amniotes. First, there is no need to detail the nature of the mesosaur embryos, citing the original publications will do. Second, when it comes to connecting both topics, which is done in the last three lines, you mention that "the eggs and hatchling [...] do not support such hypothesis" but do not detail why. Please shorten this section to what is relevant to your paper and clearly explain what your arguments are.

We have clarified why the eggs and hatchling mesosaurs do not support the Carroll’s scenario. For the other concern of the reviewer in order to reduce this section, we consider that it contains just a brief review of a paper that was published some years ago, in 2012, and it is a relevant information for the topic of this paper in the little modified current form.

Reviewer’s comment on Page 23: Everything from this point until the end of the section should be condensed and rephrased. You spend too much time detailing what analyzes you did not perform.

Respectfully, this is an explanation to your previous suggestion of to include a paragraph on the optimization of the character “average adult size” in early amniotes. Therefore, we believe that the explanation of why we decided not to include it is pertinent particularly to the readers that may want to see it.

Reviewer’s comment on Page 23: These paragraphs should also be rephrased. The potential reach of these few specimens and how they might affect future results seems overly speculative, especially since you do not provide a comparison between average mesosaur size with and without your new specimens. Alternatively, please include average body size for mesosaurs before and after the inclusion of your extreme specimens. My guess is that it would not vary that much if your sample is big enough. Regardless of the results, it would be interesting to discuss it.

The reviewer’s thoughts are right and we have provided an explanation for this in previous answers, but we will repeat it here for your convenience. “The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data.  

Reviewer’s comment on Page 25: these few specimens and how they might affect future results seems overly speculative, especially since you do not provide a comparison between average mesosaur size with and without your new specimens. Alternatively, please include average body size for mesosaurs before and after the inclusion of your extreme specimens. My guess is that it would not vary that much if your sample is big enough. Regardless of the results, it would be interesting to discuss it.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize that the data from the giant mesosaurs were not included in the calculation of these trends. Instead, the data naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.

Reviewer’s comment on Page 26: Please avoid excessive self-citation.

The reviewer previously asked why we consider the Mangrullo Formation as a relevant litostratigraphic unit for the paleontological research, and we offered a response. We work in the Mangrullo Formation, so, its citation is unavoidable. 

Reviewer’s comment on Page 34: Rather than present two regressions based on data source, you should show the difference results produced before and after the inclusion of your extreme points. Treat the data as one body of evidence, it will be more informative.

The extreme data points, corresponding to the giant mesosaurs, were neither considered nor included in the isometric calculations. In the case of orbit size vs. post-orbital skull length (Figure 11A), the measurements for giant mesosaurs clearly align with the expected pattern. It is important to emphasize this, because the data from those specimens naturally align with the isometric trend established with Nuñez Demarco et al. (2022) and Verrière and Fröbisch (2022) data. This reinforces the notion that isometry is preserved even in these giant mesosaurs. In the case of orbit size vs. maximum skull width, the giant mesosaurs do not appear to follow the general pattern. This discrepancy could be attributed to ontogenetic changes, the limited number of observed data points, skull distortion, or a combination of these factors. In Figure 11C, the original skull lengths of the giant specimens are not available, so it is impossible represent their precise placement in the graph. Nonetheless, the figure allows for an estimation of the total skull size of these specimens, assuming that the data follow a isometric pattern — an assumption supported by Figure 11A.