Local selection can promote phenotypic divergence despite gene flow across habitat mosaics, but adaptation itself may generate substantial barriers to genetic exchange. In plants, life-history, phenology, and mating system divergence have been proposed to promote genetic differentiation in sympatry. In this study, we investigate phenotypic and genetic variation in Mimulus guttatus (yellow monkeyflowers) across a geothermal soil mosaic in Yellowstone National Park (YNP). Plants from thermal annual and nonthermal perennial habitats were heritably differentiated for life history and mating system traits, consistent with local adaptation to the ephemeral thermal-soil growing season. However, genome-wide genetic variation primarily clustered plants by geographic region, with little variation sorting by habitat. The one exception was an extreme thermal population also isolated by a 200m geographical gap of no intermediate habitat. Individual inbreeding coefficients (F_IS) were higher (and predicted by trait variation) in annual plants and annual pairs showed greater isolation by distance at local (<1km) scales. Finally, YNP adaptation does not re-use a widespread inversion that underlies M. guttatus life-history ecotypes range-wide, suggesting a novel genetic mechanism. Overall, this work suggests that life history and mating system adaptation strong enough to shape individual mating patterns does not necessarily generate incipient speciation without geographical barriers.

This dataset was collected by first author Kory Kolis, as described in the paper and readme file. Plants are labelled by parental ID (see Table S1 in paper for more infomration on parental collection locations), and progeny number (last digit), and their parental collection sites were assigned to annual (A) vs. perennial (P) habitat categories.  Most families consist of 2 selfed full siblings, but a few families have just 1 member.

The column headers are defined in the readme file.

Missing data for a given trait set (e.g. pollen traits,  floral traits, vegetative traits) generally reflect accidents that happened during course of the experiment. Examples include the main stem breaking off, which would cause us not to collect the final vegetative data (plant height, stolon number, branch number), loss of pollen collection tubes (no pollen number or pollen viability), or obvious deformity of the first flower (in which case floral measures and pollen traits would be missing).

Empty cells in the data matrix indicate missing values.