In principle I like this study in that it is a clear and interesting hypothesis. The results themselves seem clear enough. However, I do have comments and questions about the way the data are assembled.
My major concern is the lumping into a single metric ‘drilling frequency’ with no discrimination. Assemblage level data never seems very useful, being as it is, an amalgamation of different taxa each with their own vulnerability or attractiveness. The overall message appears to be that drilling predation in general has dropped off recently - but there must be so much more texture to the data. Evolution happens as a much finer grain size that this. Personally I would much rather see these data expressed at the species level – indeed there are relatively few species here which have an significant drilling at all and by and large they are representing distinct sizes and habits. It might make more sense to plot DF along cores for the few genuinely well represented taxa. I see very little virtue in pooling data on say Varicorbula (a chunky, shallow burrowing bivalve) and Kurtiella (a minute symbiotic bivalve, that hangs around echinoderms). Afterall the authors do make comment on the relative numbers of some of these taxa changing along the core length. We should not expect the DF that combines all these taxa to be telling the same story and it would be interesting to see that. The data exist – why not use them rather than lumping them together. Even at an assemblage level approach it would be better to divide the DF into bivalve and gastropod prey taxa. What are the predators? Can you try to differentiate muricids from naticids – again one might not necessarily expect their activities to change in unison even at this higher taxon level. But again are we really happy to think of the predator responsible for attacking the corbulid is the same as that attacking the Kurtiella?
The data in the supplementary excel files could be improved (and numbered in a way that agrees with the text). It might be good to more clearly separate molluscan classes, and I not certain why the few ‘orphan’ lines occur at the end rather than in their respective categories. Impressive as the data are, there are an awful lot of zeroes. I would find it more illuminating to have an additional table which had the handful of interesting (ie data rich) taxa, and to have a cell which have DF for that taxon, rather than forcing the reader to flit between tables to the get the number drilled, number specimens and the do it themselves.
I assume this was done as part of some other work – it is not quite clear by 4 cores were taken but only two dealt with. What does close mean? Importantly, does this mean that data from two cores at each site were amalgamated? If so what were the differences between them? I doubt they were identical?

Fig 1A seems to have three extraneous points on it – please remove or explain the relevance of them.

I have no need or wish to remain anonymous
Liz Harper