This is an interesting study about the role of predators in driving the morphological pattern known as the “island rule” in a set of avian species, according to which mean body size increases in small species and decreases in large species upon island colonization. The authors find evidence for the island rule itself and for the significant effect of the presence of avian predators, so that the island rule effect is stronger when raptors are absent on the island. This is a predictable result but the hypothesis was worth testing, and I find the result highly relevant for the fields of island biogeography, island ecology and evolutionary biology in general. The paper is generally well written (see some edits below), and the analyses seem adequate. I have a few concerns regarding methodological issues that I present below.

In addition to the role of predation, the authors also test the role of inter-specific competition in affecting the island rule by using the presence of congeneric species as an indicator of competition. They found no effect, although a more appropriate test would include ecologically similar species (congeners or not) while controlling for phylogeny. However, I agree with the authors that ecological data on many island species is currently unavailable so that selecting potential competitors for each species may not be feasible at present.

I am surprised to see that the dataset used is composed of only 120 insular bird species, when data on many more species are readily available. For example, all the data from more than 400 species used in Benitez et al. 2021 are available in a public repository. Please explain. Using a more complete dataset would perhaps improve the power to detect the role of latitude/climate, which was found to be significant in Benitez et al. 2021 but not here.

Several details on how the data were analyzed are missing in the methods section. For example, regarding the selection of islands to analyze, did you include both oceanic and continental islands? Did you set a minimum distance from the island to the mainland? Regarding the bird data, how did you deal with sexual dimorphism? Did you average measurements from both sexes? Did you take into account subspecific morphological variation across the species range?

Finally, I would suggest changing the title from “Strength of the ‘island rule’ in birds is negatively associated with presence of avian predators” to “Strength of the ‘island rule’ in birds is positively associated with absence of avian predators” which seems more intuitive to me given that it is the absence of raptors on islands that gives rise to the pattern, not their presence.

Minor suggestions:
L73: Increased
117: co-occuring
L187: removed,
L190: increased
L190: species,
L191: requirements
L192: where food
L198: fewer competitors
L202: potentially
L208: e.g.,
L214: Please consider: We did not find a latitude effect in our models

I am satisfied with the authors' replies to my questions and concerns, and I congratulate them for an excellent work. Just a few additional suggestions:

L81: large-bodied species
L81: associated with
L82: as it has been shown in ungulates
L101: were located at least 50 km from
L108: used
L109: the subspecies geographically closest to the island
L210: a stronger island-rule effect