Biodiversity hotspots are often found in areas with high precipitation or diverse microhabitats. For plants, geographic areas with high soil variability tend to have high species richness or functional trait diversity. Serpentine soil is thought to have driven the diversification of numerous plant lineages, and many plant taxa show phenotypic variability across soil types. Precipitation can increase the effects of soil type on trait differences across soil types, such that trait differences are greater in wetter geographic areas. However, this has been examined only for growth and vegetative traits. We used a greenhouse common garden to characterize variation in reproductive phenology for a serpentine tolerator along a strong moisture gradient. We then evaluated if trait divergence changes with precipitation, and if the divergence in reproductive traits across soil types varies with precipitation. Furthermore, we conducted a water-limitation experiment and examined if water-limitation changes trait expression or divergence in trait expression. Finally, through a set of field observations we assessed if reproductive phenology and divergence in reproductive phenology differed between wetter and drier areas. We found strong effects of native precipitation regime on flowering phenology, and some support for the hypothesis that precipitation enhances divergence between plants growing on serpentine and non-serpentine soils both in the greenhouse common garden and in the field. Experimental water limitation in the greenhouse common garden resulted in reduced floral displays, phenological delays, and an increase in divergence across soil types for some traits. These results highlight the possibility that gene flow across soil types may be lower in wetter areas. In addition, they suggest that future drought may decrease the likelihood of gene flow across soil types, as well as plant reproductive potential via smaller floral displays.

Quantifying genetic structure and levels of genetic variation are fundamentally important to predicting the ability of populations to persist in human-altered landscapes and adapt to future environmental changes. Genetic structure reflects the dispersal of individuals over generations, which can be mediated by species-level traits or environmental factors. Dispersal distances are commonly positively associated with body size and negatively associated with the amount of degraded habitat between sites, motivating investigation of these potential drivers of dispersal concomitantly. We quantified genetic structure and genetic variability within populations of seven Euglossine bee species in the genus Euglossa across fragmented landscapes. We genotyped bees at thousands of SNP loci and tested the following predictions: (1) deforested areas restrict gene flow; (2) larger species have lower genetic structure; (3) species with greater resource specialization have higher genetic structure; and (4) sites surrounded by more intact habitat have higher genetic diversity. Contrasting with previous work on bees, we found no associations of body size and genetic structure. Genetic structure was higher for species with greater resource specialization, and the amount of intact habitat between or surrounding sites was positively associated with parameters reflecting gene flow and genetic diversity. These results challenge the dominant paradigm that individuals of larger species disperse farther. They suggest that landscape and resource requirements are important factors mediating dispersal, and they motivate further work into ecological drivers of gene flow for bees.

Quantifying genetic structure and levels of genetic variation are fundamentally important to predicting the ability of populations to persist in human-altered landscapes and adapt to future environmental changes. Genetic structure reflects the dispersal of individuals over generations, which can be mediated by species-level traits or environmental factors. Dispersal distances are commonly positively associated with body size and negatively associated with the amount of degraded habitat between sites, motivating investigation of these potential drivers of dispersal concomitantly. We quantified genetic structure and genetic variability within populations of ten bee species in the tribe Euglossini across fragmented landscapes. We genotyped bees at thousands of SNP loci and tested the following predictions: (1) larger species disperse farther; (2) species with greater resource specialization disperse farther; (3) deforested areas restrict dispersal; and (4) sites surrounded by more intact habitat have higher genetic diversity. Body size was a strong predictor of genetic structure, but, surprisingly, larger species showed higher genetic structure than smaller species. The way that deforestation affected dispersal varied with body size, such that larger species dispersed less far in areas with more forest. There was no effect of geographic distance on dispersal, and sites with more intact habitat had higher genetic diversity. These results challenge the dominant paradigm that individuals of larger species disperse farther, motivating further work into ecological drivers of dispersal for bees.

Understanding how urbanization alters functional interactions among pollinators and plants is critically important given increasing anthropogenic land use and declines in pollinator populations. Pollinators often exhibit short-term specialization, and visit plants of the same species during one foraging trip. This facilitates plant receipt of conspecific pollen -- pollen on a pollinator that is the same species as the plant on which the pollinator was foraging. Conspecific pollen receipt facilitates plant reproductive success and is thus important to plant and pollinator persistence. We investigated how urbanization affects short term specialization of insect pollinators by examining pollen loads on insects' bodies and identifying the number and species of pollen grains on insects caught in urban habitat fragments and natural areas. We then assessed possible drivers of differences between urban and natural areas, including frequency dependence in foraging, species richness and diversity of the plant and pollinator communities, floral abundance, and the presence of invasive plant species. Pollinators were more specialized in urban fragments than in natural areas, despite no differences in the species richness of plant communities across site types. These differences were likely driven by higher specialization of common pollinators, which were more abundant in urban sites. Pollinators were also more specialized when foraging on invasive plants across sites, and floral abundance of invasive plants was higher in urban sites. Our findings reveal strong effects of urbanization on pollinator fidelity to individual plant species and have implications for the maintenance of plant species diversity in small habitat fragments. The higher fidelity of pollinators to invasive plants suggests that native species may receive fewer visits by pollinators. Therefore, native plant species diversity may decline in urban sites without continued augmentation of urban flora or removal of invasive species.