How biotic and abiotic factors act together to shape biological diversity is a major question in evolutionary biology. The recent availability of large datasets and development of new methodological approaches provide new tools to evaluate the predicted effects of ecological interactions and geography on lineage diversification and phenotypic evolution. Here, we use a near complete phylogenomic-scale phylogeny and a comprehensive morphological dataset comprising more than a thousand specimens to assess the role of biotic and abiotic processes in the diversification of monitor lizards (Varanidae). This charismatic group of lizards shows striking variation in species richness among its clades and multiple instances of endemic radiation in Indo-Australasia (i.e., the Indo-Australian Archipelago and Australia), one of Earth’s most biogeographically complex regions. We found heterogeneity in diversification dynamics across the family. Idiosyncratic biotic and geographic conditions appear to have driven diversification and morphological evolution in three endemic Indo-Australasian radiations. Furthermore, incumbency effects partially explain patterns in the biotic exchange between Australia and New Guinea. Our results offer insight into the dynamic history of Indo-Australasia, the evolutionary significance of competition, and the long-term consequences of incumbency effects.

Code.zip: Contains the scripts and files used to run the analyses of diversification, morphological evolution, and incumbency effects. The .zip file contains individual R scripts used to run each group of analyses and two folders. The folder AuxiliaryFiles contains: a matrix with the maximum snout-vent length (SVL) per species and the mean of each other morphological trait (AllPhenoTraits.csv), a matrix used to replace placeholder names of DAISIE (https://doi.org/10.1111/ele.12461) models with the names used in the manuscript (model_names.csv), results of simulations under the DREaD model (https://doi.org/10.1086/701125) used to infer the geographic mode of speciation (simulation_results.csv), a shapefile with polygons representing the distributions of living varanids and auxiliary files (distrib.*), the main phylogenetic tree used in downstream analyses (Tree.tre), the anagenetic and cladogenetic events tables obtained from 50 stochastic maps obtained with BioGeoBEARS (https://doi.org/10.21425/F5FBG19694) (RES_ana_events_tables.Rdata and RES_clado_events_tables.Rdata), a sample of 100 phylogenetic trees (sample_of_trees.rdata), and a raster with polygons representing the distributions of living varanids (spp_rasters.rda). The folder called AuxiliaryScripts contains: scripts needed for the DREaD analyses (see https://doi.org/10.5061/dryad.d9j09bm for details) (DREaD.R, environmentalChange.R, findSisters.R, generateSummaryStatistics.R, helperFunctions.R, nicheEvolution.R, nicheRecenter.R, rangeDispersal.R, seedSpecies.R, speciateAllopatric.R, speciateDispersal.R, speciateParapatric.R, speciateSympatric.R, and summary_statsitics_functions.R), a function to run DAISIE in parallel (search.surface.DAISIE.R), and a function to calculate support for DAISIE models (Calculate_AICs.R).

Head_shape_dorsal.csv: Unaligned coordinates of landmarks and semi-landmarks used to describe head shape in dorsal view. The landmarks/semi-landmarks were digitized from photographs using 'geomoprh 3.0.3' (https://doi.org/10.1111/2041-210X.12035). For each individual, the following is indicated: taxon (the operational taxonomic unit to which the individual belongs), voucher (the collection and unique identifier of each individual; acronyms follow Sabaj 2019, available at http://www.asih.org/), sex (male or female), and whether the individual is sexually mature or not (inferred from size and/or gonad inspection). The remaining columns indicate the location of landmarks/semi-landmarks in the horizontal and vertical axes (e.g., 1x, 1y; numbering follows top panel in Fig. S2 of the article). The matrix includes some specimens that were excluded from the final analyses due to sexual dimorphism (see article for details). Some data could not be recorded for all specimens due to their poor condition. Missing data is indicated by "n/a" and was imputed using random forest training.

Head_shape_lateral.csv: Unaligned coordinates of landmarks used to describe head shape in lateral view. The landmarks were digitized from photographs using 'geomoprh 3.0.3' (https://doi.org/10.1111/2041-210X.12035). For each individual, the following is indicated: taxon (the operational taxonomic unit to which the individual belongs), voucher (the collection and unique identifier of each individual; acronyms follow Sabaj 2019, available at http://www.asih.org/), sex (male or female), and whether the individual is sexually mature or not (inferred from size and/or gonad inspection). The remaining columns indicate the location of landmarks in the horizontal and vertical axes (e.g., 1x, 1y; numbering follows bottom panel in Fig. S2 of the article). The matrix includes some specimens that were excluded from the final analyses due to sexual dimorphism (see article for details). Some data could not be recorded for all specimens due to their poor condition. Missing data is indicated by "n/a" and was imputed using random forest training.

Linear_measurements.csv: Linear measurements. For each individual, the following is indicated: taxon (the operational taxonomic unit to which the individual belongs), voucher (the collection and unique identifier of each individual; acronyms follow Sabaj 2019, available at http://www.asih.org/), sex (male or female), and whether the individual is sexually mature or not (inferred from size and/or gonad inspection). The remaining columns indicate the following unprocessed measurements in mm: snout-vent length (SVL; along dorsal midline, between tip of snout and vent), head length (along dorsal midline, between tip of snout and anterior edge of tympanum), head width (at level of anterior edge of tympanum), head depth (at level of middle of eyes), neck length (between anterior edge of tympanum and gular fold), body length (between gular fold and vent), arm length (between base of forelimb and proximal edge of claw on finger IV), hand width (perpendicular to base of finger V), hip width (width of pelvic gridle measured at level of middle of hindlimbs), leg length (between base of hindlimb and proximal edge of claw on toe IV), foot width (perpendicular to base of toe V), tail length (between vent and tip of tail), tail width (measured at level of one third of tail length from vent), and tail depth (measured at level of one third of tail length from vent) The matrix includes some specimens that were excluded from the final analyses due to sexual dimorphism (see article for details). Some data could not be recorded for all specimens due to their poor condition. Missing data is indicated by "n/a" and was imputed using random forest training.

Ranges.zip: Contains a shapefile (Ranges.shp) and auxiliary files. The shapefile contains polygons representing the distributions of living varanid species and two closely related taxa in the families Lanthanotidae and Shinisauridae (modified from Roll et al. 2017 https://doi.org/10.1038/s41559-017-0332-2). The "Binomial" attribute indicates the taxon.

Tree.tre: Main tree used in downstream analyses. The phylogeny is a time-calibrated tree in Newick format that is primarily based on the phylogenomic analyses presented by Brenan et al. 2021 (https://doi.org/10.1093/sysbio/syaa046). We added tips based on previous studies or our own analyses of publicly available molecular data (see article for details). Units are millions of years.