Biogeographic dependent diversification analyses
To test if the Andean uplift or closure of the Panama Isthmus influenced Preponini diversification rates we used a similar likelihood approach as for time dependent diversification analyses implemented in RPANDA. In this case, we evaluated the fit of six models allowing speciation and extinction rates to be dependent on the paleo-elevation of the Andes (three models) or the degree of connectivity between Central and South America (three models; Morlon, 2014; Morlon et al., 2011; see Table 1).
We used Andean paleo-elevation from present time until 32 Ma based on the database provided by Lagomarsino et al. (2016) and information from Garzione et al. (2008). The closure of the Panama Isthmus was estimated using reports of migration rates for several groups of organisms by Bacon et al (2015). In the latter study they reported that migration rates between the two continents increased sequentially at four points in time, 41.1, 23.7, 8.7 and 5.2 Ma. We used these migration rate estimates to compute the probability of observing at least one migration event in each time period (i.e. 50 – 41.1, 41.1 – 23.7, 23.7 – 8.7, 8.7 – 5.2, 5.2 - present). We then used the cumulative probability through time to observe at least one migration event per million years as a proxy of the degree of connectivity between land masses (Table S3).