In general, I totally agree with the comments associated with the user rating posted by ErikiLund. Our study did examine only a few taxon pairs, and thus we were careful to explicitly acknowledge this point in several places in our article. Data from additional taxon pairs is indeed highly desirable.

That said, I want to point out that generating the types of data presented in the article is difficult, even for a single taxon pair. In particular, isolating how many niche dimensions are subject to divergent selection in the wild requires manipulative experiments in nature, experiments that not only measure selection, but also isolate the actual agents of natural selection.

In short - further replication is required, but the existing data does support the niche dimensionality idea, using the required morphological, genetic, and experimental data.