New insights into the palaeoecology of Miocene Eurasian rhinocerotids based on tooth analysis

Hullot et al. submitted a very interesting paper entitlted "Spatio-temporal diversity of dietary preferences and stress sensibilities of early and middle Miocene Rhinocerotidae from Eurasia: impact of climate changes"
Since very few study focus on the paleocology of Miocene Rhino, this paper is a great contribution that should be published.

In general, this paper is very rich, well written with supplementary material that provide a lot of data to better understand the study.
However some points should be improve either to facilitate the understanding of the study or to moderate some affirmations :

- In the method, the authors explain which data they choose to investigate hypoplasia. However they never explain the reason behind this selection. Hypoplasia remains quite a mysterious phenomenon and I am wondering for example what information can be extrapolate from the distance between the defect and the collar of the teeth. Could the author detail more their reasoning ?

- In the discussion, "Interactions with co-occurring herbivores" part. I find the author quite rush in their affirmation. They seem to forgot that they are studying palimpsest with a very high-resolution method. There is a discrepancy between the precision of the method and the context in which it is applied which should be taken into account and reported in the text.

These two points in particular seem to me very important to consider. Otherwise, I have pointed out in the text (see attachment) some minor errors/changes/suggestions that the authors may or may not take into account. 

I hope these few comments will be helpful for the authors.

Overview

This paper includes an interesting mixture of dental microwear texture analysis and documentation of dental hypopasias in Neogene Eurasian rhinos. The dental microwear analyses adopt established methods. Likewise, the distribution of hypoplasias are documented soundly. However, the results and discussion sections wander, generate unsupported paleodietary interpretations and make loose associations of the results with paleoenvironmental conditions. While the data themselves are analyzed with sound statistical methods, the authors take very large leaps from the statistical results regarding differences between the various samples to generalizations about diet and paleoecology.

The paper states that its aim is to assess "dietary preferences using dental microwear texture analysis, and stress sensibility via the study of enamel hypoplasia". (lines 65-67). However no hypothesis was provided, leading to an unfocused presentation. I think the authors would be better served by generating a testable hypothesis or series of hypothesis which could them be used to give the investigation purpose and focus. What ecological paleoecological questions concerning Neogene rhinos can be addresssed by these data? Answering this question may also lead to a more coherent research design. Why are dental microwear data and dental hypoplasias being presented together? Is there some hypothesis that they can both address? Presently, it seems as if microwear reveals more about diet, while hypoplasias have more to do with ecological stress. These aren't mutually exclusive phenomena. Is there some interesting question that could be answered by examining both. If not, the authors may be better off publishing these two studies separately. 

Throughout the DMT results and discussion sections and summarized in table 6, a variety of interpretations are made about diet and feeding. Most of these interpretations use the classic browser, mixed-feeder, grazer categories, but other ad hoc terms and special qualifications are thrown about including "folivore, "browser favoring leaves", "soft food items", "soft browsing", "browsing preferences with the inclusion of hard objects, probably fruits".The statistical methods used to test for differences in the data give a false impression of scientific rigor. But the actual dietary interepretations have seemingly been made out of thin air. How were these interpretations made? Extant rhino microwear data are discussed in a subsequent section but it is not clear how the modern data are being used to interpret diets of extinct species. What are the confidence levels of these dieatary assignments?  How are these dietary categories defined? How were the lines drawn between browser, mixed-feeder and grazer in terms of microwear data and diet? What aspects of the microwear data lead you to make conclusions about diet? What about these other special categories, such as folivore, soft browsing, and other special exceptions (e.g. browser favoring leaves, soft browsing etc)? What elements of the data reveal these?  It is already challenging enough to differentiate classic browsers, mixed feeders, and grazers using microwear data and these additional diet characterizations, I suspect, are actually beyond the powers of the available dental microwear data to differentiate. 

In several areas, this paper makes strong conclusions about diet and other aspects of paleocology. For example, in the conclusion section (lines 663-665):"Though, DMTA results suggested only browsers and mixed-feeders (no grazers nor frugivores) in the studied rhinocerotid sample, they unraveled clear niche partitioning through food resources at several diachronous localities..."  Niche partitioning is a process by which natural selection drives competing species into different patterns of resource use or different niches. Some microwear differences found in a small number of teeth do not constitute clear evidence for niche partitioning or any other complex evolutionary and ecological process. The dental microwear differences merely suggests different diets or feeding environments. It does not reveal why those diets are different. With regard to niche partitioning, it can only be concluded from this analysis that it cannot be falsified as an explaination for the differences. Perhaps it would be helpful here, and, in fact, for all the conclusions derived in this paper, to consider what is the null hypothesis regarding ecological differentiation of these species to begin with? This is one example of numerous in the manuscript where your conclusions need substantial softening.

Specific comments (see comments in manuscript for additional comments

The first two sentences of the abstract seem unecessary to me. Of course major changes are documented in most of not all geological ages. Rhinoceroses are interesting, according to who? They are abundant and diverse according to what metric?

In both the abstract and the main document, proxy data are confused with diet (or other ecological parameters of interest). For example, line 22 states,  "Our results suggest a clear niche partitioning based on diet at Kumbi 4..." Replace "diet" with "DMT data". The evidence for niche partitioning is not based on diet, it is based on DMT data, which is not diet. Also see comments above about niche partitioning. 

Likewise on lines 24-25, "All rhinocerotids studied were browsers or mixed feeders, and none had a grazing nor frugivore diet." Avoid such definitive conclusions. To be more precise, it would be better to write, "the DMT data are consistent with browsing and mixed feeding diets, but not grazing". 

Line 27: If the prevalence of hypoplasias document the local conditions, those conditions need to be stated, otherwise omit this sentence.

Line 90: A figure of the grinding and shearing facets sampled needs to be provided. A person who is not very familiar with rhinocerotid molars will need a figure to understand this.

Line 171-172: The statement about avoiding classic thresholds is contradicted in the results tables. Table 2 appears to use P=0.1 as a threshold. Table 3 explicitly uses P=0.05 as a threshold. A keyword search for the word "significant" is needed to be sure that it is not used in ways that contradict the statements about significance in the methods (see line 666)