The paper by DeLeo et. al., is a very nice study that describes the evolution of bioluminescence and depth in anthozoans, octocorals in particular, using phylogenomic methods. The manuscript is clear and well written; the figures are interesting, and the supplementary information complements well the main document. Please find below some suggestions to improve the quality of the manuscript.

For bioluminescence tests, the fragments that glow are certain to have it, but I am not sure that the fact that they don’t glow is because they lack bioluminescence. There may be other factors that can cause coral fragments, exposed to physical and physiological stress, to not glow even though they are bioluminescent. ¿How many of the results can be false negatives? ¿Is it possible to use a statistical method to estimate the proportion of false negatives?

It is well known that the accuracy of estimating ancestral character states depends on the density of taxon sampling, the character states linked to those terminal taxa, and the tree topology. Although I’m certain that the tree topology used in this study is robust and as complete as it can be at the moment, the authors must include and discuss the probable uncertainties in their reconstructions due to the selection of terminal taxa (for depth and bioluminescence), and to the doubts of the particular character state for some taxa (for bioluminescence).

Other comments

It is not clear which trees represent UCEs topology or mixed UCEs + mtMutS topologies. Please include this information in the legends of both the main document and supplementary figures.

Figure 1. Acanella arbuscula has an “s” at the end; please delete.

Figure S3. Include more information in the legend, particularly regarding the outgroup used to root the tree. Those taxa are not octocorals.