The spatial scale of genotype by environment interaction (GEI) for fitness in the loose-flowered gilia, Ipomopsis laxiflora (Polemoniaceae)

Genotypes within a population may respond differently to fine-scale environmental variation leading to the occurrence of genotype by environment interaction (GEI). We used 12 genotypes of an annual wildflower, Ipomopsis laxiflora, to investigate the spatial scale of GEI for fitness in natural shortgrass prairie in two field seasons. In each year, we planted ca. 100 replicate plants per genotype across 8100 m² of prairie habitat-replicates were separated by distances ranging from ca. 3 to ca. 90 m. We used a nested generalized linear model to test for GEI and to partition the effect of genotypes, a nested series of spatial scales, and interactions between genotypes and scales on relative fitness. In both years, plants were severely damaged by early-season hailstorms. Hail damage occurred across the entire experimental meadow and therefore is likely to have homogenized environmental variation at fairly large spatial scales. In 1997, very few plants were killed by hail damage, and we detected significant GEI for fitness across the largest spatial scale tested (ca. 45-90 m). This GEI is probably the product of differential tolerance of the genotypes to hail damage contingent on the environmental conditions found across sections of our experimental meadow. We found significant spatial heterogeneity in soil moisture across our experimental meadow, but this variation did not explain substantial GEI for fitness. In 1998, only 10 experimental plants survived hail damage. Patterns of GEI in nature are the result of a balance of large- (e.g., hail, fire, and severe grazing) and small-scale (e.g., soil and water nutrient availability and local competition) processes. Additional studies in natural populations are needed to evaluate the relative importance of processes occurring at landscape and local scales in generating or homogenizing GEI.