Around the world, many species are confined to “Sky Islands,” with different populations in isolated patches of montane habitat. How does this pattern arise? One scenario is that montane species were widespread in lowlands when climates were cooler, and were isolated by local extinction caused by warming conditions. This scenario implies that many montane species may be highly susceptible to anthropogenic warming. Here, we test this scenario in a montane lizard (Sceloporus jarrovii) from the Madrean Sky Islands of southeastern Arizona. We combined data from field surveys, climate, population genomics, and physiology. Overall, our results support the hypothesis that this species’ current distribution is explained by local extinction caused by past climate change. However, our results for this species differ from simple expectations in several ways: (1) their absence at lower elevations is related to warm winter temperatures, not hot summer temperatures; (2) they appear to exclude a low-elevation congener from higher elevations, not the converse; (3) they are apparently absent from many climatically suitable but low mountain ranges, seemingly “pushed off the top” by climates even warmer than those today; (4) despite the potential for dispersal among ranges during recent glacial periods (~18,000 years ago), populations in different ranges diverged ~4.5–0.5 million years ago and remained largely distinct; and (5) body temperatures are inversely related to climatic temperatures among sites. These results may have implications for many other Sky Island systems. More broadly, we suggest that Sky Island species may be relevant for predicting responses to future warming.

Introgression is now commonly reported in studies across the Tree of Life, aided by recent advancements in data collection and analysis. Nevertheless, researchers working with non‐model species lacking reference genomes may be stymied by a mismatch between available resources and methodological demands. In this study, we demonstrate a fast and simple approach for inferring introgression using RADseq data, and apply it to a case study involving spiny lizards (Sceloporus) from northeastern México. First, we find evidence for recurrent mtDNA introgression between the two focal species based on patterns of mito‐nuclear discordance. We then test for nuclear introgression by exhaustively applying the “five‐taxon” D‐statistic (DFOIL) to all relevant individuals sampled for RADseq data. In our case, this exhaustive approach (dubbed “ExDFOIL”) entails testing up to ~250,000 unique four‐taxon combinations of individuals across species. To facilitate use of this ExDFOIL approach, we provide scripts for many relevant tasks, including the selection of appropriate four‐taxon combinations, execution of DFOIL tests in parallel, and visualization of introgression results in phylogenetic and geographic space. Using ExDFOIL, we find evidence for ancient introgression between the focal species. Furthermore, we reveal geographic variation in patterns of introgression that is consistent with patterns of mito‐nuclear discordance and with recurrent introgression. Overall, our study demonstrates that the combination of DFOIL and RADseq data can effectively detect introgression under a variety of sampling conditions (for individuals, populations, and loci). Importantly, we also find evidence that batch‐specific error and linkage in RADseq data may mislead inferences of introgression under certain conditions.

Targeted sequence capture is becoming a widespread tool for generating large phylogenomic data sets to address difficult phylogenetic problems. However, this methodology often generates data sets in which increasing the number of taxa and loci increases amounts of missing data. Thus, a fundamental (but still unresolved) question is whether sampling should be designed to maximize sampling of taxa or genes, or to minimize the inclusion of missing data cells. Here, we explore this question for an ancient, rapid radiation of lizards, the pleurodont iguanians. Pleurodonts include many well-known clades (e.g., anoles, basilisks, iguanas, and spiny lizards) but relationships among families have proven difficult to resolve strongly and consistently using traditional sequencing approaches. We generated up to 4921 ultraconserved elements with sampling strategies including 16, 29, and 44 taxa, from 1179 to approximately 2.4 million characters per matrix and approximately 30% to 60% total missing data. We then compared mean branch support for interfamilial relationships under these 15 different sampling strategies for both concatenated (maximum likelihood) and species tree (NJst) approaches (after showing that mean branch support appears to be related to accuracy). We found that both approaches had the highest support when including loci with up to 50% missing taxa (matrices with ∼40–55% missing data overall). Thus, our results show that simply excluding all missing data may be highly problematic as the primary guiding principle for the inclusion or exclusion of taxa and genes. The optimal strategy was somewhat different for each approach, a pattern that has not been shown previously. For concatenated analyses, branch support was maximized when including many taxa (44) but fewer characters (1.1 million). For species-tree analyses, branch support was maximized with minimal taxon sampling (16) but many loci (4789 of 4921). We also show that the choice of these sampling strategies can be critically important for phylogenomic analyses, since some strategies lead to demonstrably incorrect inferences (using the same method) that have strong statistical support. Our preferred estimate provides strong support for most interfamilial relationships in this important but phylogenetically challenging group.