This manuscript builds on data and dominance calculations published recently in Ecology Letters (Matas-Granados et al. 2024). It extends by testing if functional traits explain two types of dominance: (1) local abundance and (2) regional frequency across four different rainforest habitat types in the western Amazon.

The idea that dominance can be explained by traits is appealing, but I have a number of major concerns about the data and analyses presented in the manuscript. I outline these below, and provide detailed mark-up on an attached PDF document.

Data: The authors use available species-level trait data (largely from existing databases) but given the enormous diversity in these systems do not achieve <30% trait coverage for all but DBH max (estimated directly from their plot data).

Arbitrary dominance thresholds: Whilst I acknowledge that the definition of dominance (species that together account for 50% of the total relative abundance of each habitat type) has been established in the literature, it may be more appropriate to keep local and regional dominance as continuous variables, like you do with the two mixed-effects models.

Dominance metrics should be kept separate: I would recommend removing any analyses that pooled the two types of dominant species (locally abundant and regionally frequent) – this is problematic considering expectations that these two dominance types are functionally distinct (hypothesis 2). Your results from the mixed-effects models affirm this expectation and as such, it is likely that this pooling of dominant species obscures trait-based dominance patterns in your results.

Comparing against pooled non-dominants: An additional issue with the ANOVAs and PCA analyses that are referred to in the main text (Figure 3 and 4), is that dominant species from each habitat type were compared against non-dominant species pooled across all habitat types. While dominant and non-dominant species within each of the four habitat types were analysed separately in supp analyses (Figure S8), these results were mentioned fleetingly and should have been presented in place of Figure 3. Pooling all non-dominants is problematic as it does not allow you to draw inferences about functional differences between dominants and non-dominants within habitat types.

Mixed-effects models: These models are in my opinion a much stronger approach to assess the role of traits in driving different types of dominance. However, it was unclear whether they included all species (both non-dominant and dominant) or only included dominant species. The results communicated in the paper suggest that only dominant species were included, however, this was not explicitly stated. It would make most sense to include all species so that (1) you are undertaken an honest test that dominance (treated as a continuous variable from very low to high values) is explained by traits and (2) the slopes between dominance metrics and traits can be estimated most effectively because the dominance variables span the full range of values.
Habitat type descriptions: It would be useful to provide a more detailed description of the habitat types and their most limiting resources (e.g. light, moisture, oxygen) in order to drive your hypotheses and justify functional differences between habitat types.

Other points: It may be useful to acknowledge environmental variation within each of the habitat types (e.g. elevation gradients across terra firme forests), which is likely to drive strategy variation, perhaps even more so than habitat type??? Consequently, trait-based dominance patterns for the habitat types may be obscured by trait-modulated responses to within-habitat environmental variation.

The paper had two research questions that were consistently referred to throughout, however, the authors had many hypotheses related to each question that I felt were too ambiguous. It was therefore difficult to link the proposed ideas, founded in the literature, with the results and conclusions.