Idaho fescue (Festuca idahoensis Elmer) is a native grass species that has attracted interest for use in revegetation, reclamation and other applications. However, there is a serious shortage of commercially available seed and concern that germplasm used will not be adapted to the site. A component of adaptation is genetic variability that allows a species to occupy a greater range of environments. Field trials were conducted in southern Alberta to determine the genetic variability of phenotypic and performance characteristics among genotypes of Idaho fescue. Since grazing pressure may contribute to genetic selection and may therefore affect the variability within a population, we also compared genotypes taken from a heavily grazed paddock with those taken from a lightly grazed paddock. Forty-nine genotypes from three populations were randomly sampled in 1992, propagated vegetatively to produce plants for replicated trials and planted in four locations in 1993. Nine plant characteristics were observed from 1993 to 1997. Since the genotypes were grown in common environments, variation among them was assumed to be caused by genetic differences. All variables were affected (P<0.05) by population, test location and year, while the effect of population was also influenced by test location and year for a few variables. While the Idaho fescue plants expressed differences (P<0.05) among populations for all selected traits, they displayed considerable overlap in the range of values for all variables both within and among test locations. Therefore, while the populations may be different, individuals within populations exhibit common attributes over a large range. This suggests that sufficient genetic variability exists in all populations to allow successful establishment over a large range of environmental variability. Results from a secondary test suggest that selection pressure, induced by grazing, resulted in genotypes that were smaller in crown circumference, had less spring vigor, had shorter flowering tillers and produced less seed. However, this observation needs further validation with a more robust test.