I read this article with great interest as the authors explore the very timely question of individual thermoregulatory strategies in the context of animal personality (also known as consistent individual behavioral differences). The main results are interesting and fit well into the accumulating body of results we have regarding individuality in reptile thermoregulation. The strongest aspect of the study is the massive amount of data (100 readings for 30 individuals) the authors gathered by biologging crocodiles over 15 years, giving them substantial statistical power to explore individual variation in thermoregulatory strategy.

However, I believe that the study misses a great opportunity by concentrating solely on among-individual variation in the studied traits. The above-mentioned high number of repeats per individual makes separating among- and within-individual behavioral variation (i.e., behavioral type and behavioral predictability) possible by applying a double-hierarchical modeling approach. These models are now widely used and relatively easy to implement in R (see Hertel et al. 2020 Mov Ecol, Hertel et al. 2021 J Anim Ecol for a detailed tutorial). By doing so, the authors would be able to address further questions on whether i) individuality in behavioral predictability is present or not; ii) behavioral type and predictability are linked; and iii) if sex and size affect behavioral predictability. This aspect would broaden our understanding regarding crocodile thermoregulation in the wild, which would provide valuable insight on thermoregulatory strategies, as we generally lack data from wild-ranging reptiles.

I also have a minor, but perhaps equally important, comment. The authors combined sex and SVL to use it as a fixed factor in their model, but I advise against doing so. For one hand, the interpretation of the results is not that easy this way, as we don’t get information on the main effect of sex and SVL; these variables may affect behavior separately. Also, there are no medium-sized females, which makes the categories somewhat unbalanced.

Review of ’Cooling down is as important as warming up for a large-bodied tropical reptile’ (RSPB-2024-1804)

I greatly appreciate the authors’ effort in applying the recommended DHGLM methodology. In my view, the paper’s strongest message is the substantial individual variation and consistency over time in both the average and predictability of behavioural thermoregulation. This finding underscores that thermoregulatory behaviour is an integral part of an individual’s behavioural strategy (i.e., animal personality), and should be considered as important as other more traditionally studied ‘personality traits’ in poikilotherms. The research is particularly exciting because, to my knowledge, these aspects have never been tested in a large-bodied tropical reptile, unlike in small-bodied helioterms.

Although the paper has significantly improved with the new analysis, I believe the authors should further emphasize the animal personality aspect by discussing their findings on inter- and intra-individual variation in thermoregulatory behavior and why it is important to integrate this trait into the study of ectotherms. I’m not suggesting a complete shift in focus, but expanding on these aspects could be beneficial. The following recent papers might be helpful in this regard: Goulet et al. 2016 Behav. Ecol. 27, 1635–1641; Goulet et al. 2017. Ecol. Evol. 7, 710–719; Goulet et al. 2017 J. Anim. Ecol. 86, 1269–1280; Michelangeli et al. 2018 Funct Ecol 32, 970–981; Horváth et al. 2024 Sci Rep 14, 14200

I have only two minor additional comments.

1. It should be clarified early on whether Tb or Tr was analysed. The Abstract suggests that Tb is the dependent variable, but it turns out that it was corrected for water temperature, making Tr the dependent variable. Please clarify this and ensure consistency in terminology and abbreviations.

2. In addition to repeatability and CVP estimates, I would like to see the random effects table (intercept and sigma_intercept from ‘group level effects’) included in the model output. As it stands, Suppl. Table 1 provides information only on the fixed effects.

Once these revisions are made, this paper will be an excellent contribution to PRSB, and I look forward to citing it in the future.