The leading edge matters too: fitness and the expression of adaptive differentiation are greatest at the high-elevation edge of a species range

This dataset contains raw and processed data from two common garden experiments testing local adaptation and fitness variation across the elevational range of Erythranthe laciniata (Phrymaceae), an annual monkeyflower endemic to the Sierra Nevada, California, USA. Experiments were conducted in 2009 and 2021 at low-, mid-, and high-elevation field gardens spanning the species’ distribution. The dataset includes survival and lifetime fitness measurements (total flower number), population and maternal line identifiers, geographic coordinates, elevation, climate data (NASA Daymet), and associated metadata. These data support analyses presented in Shay & Pennington et al. (2025), The leading edge matters too: fitness and the expression of adaptive differentiation are greatest at the high-elevation edge of a species’ range (Ecology Letters).

Methods
Seeds were collected from 23 populations across the elevational range of E. laciniata using stratified random sampling of ≥60 maternal plants per site. To minimize maternal effects, a refresher generation was grown under controlled greenhouse conditions before planting into field common gardens. In 2009, three experimental gardens were established at low (1000 m), mid (1670 m), and high (3095 m) elevations in Fresno County, CA, with ~100–60 replicate individuals per population per garden. In 2021, three gardens were again established (1000 m, 1555 m, and 2500 m), with nine focal populations represented by 15 maternal lines and three replicates each. Blocks were randomized, overwintered naturally, and censused through flowering. Survival was scored as a binary outcome (flowered or not) and lifetime fitness was measured as the total number of flowers produced per plant. Climate data (mean daily minimum and maximum temperature, precipitation) were extracted from the NASA Daymet V4 dataset (1980–2005 normals, 2008–2009, and 2020–2021 growing seasons). Genetic and geographic distances were obtained from previously published data (Sexton et al. 2016). Data processing and analyses were conducted in R (v4.3.1), including generalized linear mixed models (glmmTMB), post-hoc contrasts (emmeans), and multiple regression of distance metrics. All analysis code is provided to allow full reproducibility.