Identifying common trends in speciation requires a comparison of species divergence history across many replicate species pairs.
However, a focus on dense geographic sampling of populations has the drawback of diverting resources away from contrasting speciation histories across many replicate species pairs.
Analysis proceeded in three stages: 1. Full data set: Following exploratory analysis, the final model used the fixed terms: site, species, and species × site interaction, as well as the random terms: provenance (species), provenance (species) × site, and replicate × species (site) (Table 4). 2.
Some publications partially or wholly replicated the species names from the Materials list in the phylogenetic matrix.
Our analysis is therefore limited to studies (sometimes subsets of the full data from each experiment) that replicated both species richness and composition.
Species composition is usually best treated as a random effect because replicate species compositions are usually only a small sample of all the possible combinations.
This produced a set of 100 bootstrap replicate species trees and their greedy consensus tree.
In total, our collection has 200 replicate species trees generated under four model conditions (M1, M2, M3, and M4).
For MLBS, we used the greedy consensus of 200 replicate species trees, each computed on an input consisting of one bootstrap replicate tree per gene.
By removing the need for phylogeographic sampling, the triplet method can help expedite the study of comparative species divergence across replicate species pairs.
