Trait dependent diversification analyses
To test whether Preponini lineage diversification has been dependent on phenotypic evolution (i.e. color pattern), we used the Equal-Splits with Simulated null model (ES-sim) method (Harvey and Rabosky, 2018). ES-sim employs a tip and rate correlation technique to test for the significance of the correlation between trait values at the tips and the phylogeny’s branching pattern in the absence of an evolutionary model. Harvey and Rabosky (2018) showed that for phylogenies with ~50 species, ES-sim performed better in the Type I error rate and had comparable power to QuaSSE (Quantitative State Speciation and Extinction, FitzJohn, 2010), a method commonly used to jointly model diversification and the change in continuous traits. In addition, ES-sim is useful for small phylogenies with few and slight changes in diversification rate, which might be the case in our study (Harvey and Rabosky, 2018). As a result of the ES-sim test, we obtained i) a Pearson’s correlation coefficient (rho), which quantifies the strength of the relationship between traits and each lineage’s diversification rate and ii) the simulation-based two tailed probability value (p-value) which allows comparison of the data to a null hypothesis.