This manuscript explores the evolution of female ornamentation in relation to parental costs in females, using cross-species comparisons in a phylogenetic context. While other studies have used similar methodologies to look for broad patterns, this paper is important in that it focuses on parental costs (as either egg volume for females or parental care in males). The author/s took care to develop thoughtful methodology and present the results in a way that is easy to understand and read.

My main concern is related to the discussion in lines 324-330 and more generally, studies like this one which seek to look at phylogenetic correlates of female coloration. As the author/s discusses, one of the strongest correlates of female coloration is male coloration, demonstrating the importance of genetic linkage. The author/s discuss that shared selective pressures on both sexes may produce the same effect. However, this framing leaves out a major possibility – divergent types of selection in males and females can select for similar ornamentation in both sexes. The genetic basis for shared ornamentation may in fact be pleiotropy or genetic linkage with coloration in the opposite sex, but this does not negate an adaptive function in both sexes. In other words, genetic correlation is a not necessarily an alternative hypothesis to any adaptive explanation for female coloration, and linkage does not imply non-adaptivity.

The problem is most apparent in the decision to statistically “correct” for correlated plumage similarities in males (lines 197 – 199). I’m not saying that this is completely incorrect, but I do think it stems from some big assumptions that we know not to be true in many cases. Including male coloration as a covariate in the model effectively will remove significant amounts of variation that other variables (such as egg production and paternal care) could account for. The question is, does female coloration need to vary independently from male coloration to be adaptive? I believe the answer is “no”. In a sense, the author/s methods should be most accurate if we assume that NONE of the correlated female coloration with males could have been selected for by the other variables. Say for example female coloration in a species is adaptive because of low egg production levels AND female coloration is genetically linked with the male’s coloration. In this case we lose the ability to detect the explanatory power of egg production because the model will account for most of the variation through male coloration. Perhaps this is why some of the results were relatively weak – the methods in this paper are highly conservative, perhaps too conservative to detect real selective mechanisms.

Should the author/s remove male coloration from the model? That’s not a great way to do things either, because this makes the opposite assumption – that ALL female coloration is potentially adaptive even when it could be a non-adaptive by-product of male coloration. However, I see no reason why the assumption that all correlated coloration is adaptive (removing male coloration in the model) is any more egregious than assuming none of correlated coloration is adaptive (including male coloration in the model). As many studies have shown, female ornamentation has adaptive value. The only study I know of that was able to look at correlations between sexes (when selection on females is convincingly excluded) is Clark’s 2020 paper (doi.org/10.1111/evo.13881), which found that non-adaptive correlations exist but they are weak.

I don’t have a perfect solution to this issue, and I think this is a general problem with phylogenetic analyses of sex-specific ornamentation that seek adaptive explanations, rather than a specific criticism of this paper. A separate analysis in which male coloration is excluded from the model could be informative, though again, this has similar issues to including it. One method is too conservative, the other is too liberal. At the very least, much more emphasis on this implicit model assumption needs to be addressed throughout the paper but especially in the discussion.

A few line-by-line minor comments:

25-26: Though I agree with this point historically, especially before 2000, at this point there’s a solid 20+ years of good research on female ornamentation. Not enough, but a lot.

29-67: These two paragraphs are nicely written, but they read as a bit of a history lesson and could be cut for space if need be.

68: What’s a female-related force? I would say that males also have a high cost of reproduction.

181: “We used estimates of contributions when available, otherwise we estimated them.” This line reads a bit funny. The rest of the paragraph is the process of estimation, correct? If so, just needs a little clarification in the first sentence.

191-194: So were cases of male-only care actively removed or were there just no cases of it? The equations provided for calculating parental care values seem like they should be adequate for placing male-only care on this spectrum. So if the data exists I see no reason for exclusion.