Premise: Genetic variation influences potential for evolution to rescue populations from impacts of environmental change. Most studies of genetic variation in fitness-related traits focus on either vegetative or floral traits, with few on floral scent. How vegetative and floral traits compare in potential for adaptive evolution is poorly understood.

Results: Vegetative traits SLA and WUE varied greatly among planting sites, while showing weak or no genetic variation among source populations. SLA and trichomes responded plastically to snowmelt date, and SLA exhibited within-population genetic variation. All aspects of floral morphology varied genetically among source populations, and corolla length, corolla width, and sepal width varied genetically within populations. Heritability was not detected for volatiles, due to high environmental variation, although one terpene had high evolvability and two terpenes correlated genetically with sepal width, associated with high emission from that tissue. Environmental variation across sites was weak for floral morphology and stronger for volatiles and vegetative traits. Three of 4 volatiles showed inheritance departing from additive.

Conclusions: Results indicate stronger genetic potential for evolutionary responses to selection in floral morphology compared with scent and vegetative traits, while finding potentially adaptive plasticity in some vegetative traits.

Genetic families of seeds were produced using reciprocal factorial crosses and planted into two field sites in 2007 and 2008 as seed. Traits were measured between 2009 and 2018 on these plants. Leaves were collected for measurements of specific leaf area and trichome density. A portable photosynthesis instrument (Licor 6400) was used to measure intrinsic water-use efficiency. Floral morphometrics were measured with calipers. Nectar production was measured by covering flowers to prevent access by pollinators and using microcapillary tubes to extract nectar. Floral color was measured using reflectance spectrometry. Floral scents were sampled with dynamic headspace sampling and processed using thermal desorption gas chromatograpy - mass spectrometry.

A README file is provided for each data set. Missing data is indicated by "."