Response to comment on Mutualism weaken the latitudinal diversity gradient among oceanic islands

TL;DR

The study demonstrates that mutualistic relationships significantly weaken the latitudinal diversity gradient among oceanic islands, with plants involved in mutualisms contributing disproportionately to this pattern.

Contribution

It provides new evidence that mutualisms reduce the latitudinal diversity gradient on oceanic islands, challenging previous assumptions and addressing critiques with robust analysis.

Findings

Mutualist-associated plants contribute to reduced LDG on islands

Models show mainland diversity predicts island diversity

Critique reanalysis is flawed and does not alter original conclusions

Abstract

In our original paper (Delavaux et al. 2024; https://www.nature.com/articles/s41586-024-07110-y), we find that the latitudinal diversity gradient (LDG) in plant species richness is reduced on oceanic islands worldwide. Moreover, we find that plants that associate with mutualists, including pollinators, AM fungi, and N-fixing bacteria, disproportionately contribute to this reduction. To make this assertion, we fit models of continental diversity to absolute latitude and compared mainland diversity estimated by these models to the observed diversity of individual islands. The critique of our work by Pitcher and Hartig (arXiv:2411.15105) questions our results through a reanalysis of the data. In their reanalysis, they add longitude as a predictor of richness on islands. However, we have several concerns about this reanalysis. In particular, we question (1) the biological/physical basis of predicting species richness with longitude as well as modelling non-linear relationships potentially leading to spurious relationships and (2) the justification for how this comment calculated island deficits from the data. Additionally, (3) described analyses and provided code (https://github.com/MaximilianPi/IslandLDG) are in many cases insufficient to understand or meaningfully engage with this work, precluding productive discussion about cited analytical concerns. Given these major issues with the reanalysis, we reject the conclusions from this critique that our prior results are due to statistical artifacts. We therefore do not have reason to question our original conclusion that the latitudinal diversity gradient is weakened by mutualisms among oceanic islands worldwide.