Seeds sown in the fall to restore sagebrush (Artemisia spp.) steppe plant communities could experience high mortality when they germinate and seedlings freeze during the winter. Delaying germination until the risk of frost is past could increase seedling survival. We evaluated the use of abscisic acid (ABA) to delay germination of Elymus elymoides, Pseudoroegneria spicata, and Linum perenne. The following treatments were applied: uncoated seed, seed coated with ABA at 2.2, 4.4, 8.8, 13.2, and 17.6 g of active ingredient kg-1 of seed, and seed coated with no ABA. The influence of seed treatments on germination were tested at five different incubation temperatures (5-25°C). The lowest application rate of ABA had no significant influence on germination percentage but higher application rates showed a decline. All concentrations of ABA tested delayed germination, especially at low incubation temperatures. For example, the time required for 50% of the seeds to germinate at 5°C was increased with the use of the lowest ABA application rate by 56, 61, and 14 days, for E. elymoides, P. spicata, and L. perenne, respectively. Quadratic thermal accumulation regression models were developed for each species and treatment to predict progress toward germination. For the two grasses, models had sufficient accuracy (R2 = 0.61- 0.97) to predict germination timing using field seedbed temperatures. Equations for L. perenne were less accurate (R2 = 0.03-0.70). Use of these models will allow testing whether ABA will delay germination sufficiently to avoid winter frost periods and provide the basis for future field tests.

Little information is available on the extent of local adaptation for many native grass species. This is the case for squirreltail (Elymus section Sitanion), despite this group's prevalence and importance in rangeland restoration efforts. I evaluated 32 populations of E. elymoides ssp. brevifolius race C, a phylogenetic subdivision of bottlebrush squirreltail (E. elymoides) centered in the northern Intermountain West, for phenotypic variables and neutral genetic markers to measure their association with geographical origin. Phenotypic traits were measured in common field and greenhouse environments, and genetic diversity was assessed using Amplified Fragment Length Polymorphism. Three factors were extracted from the phenotypic data set using common factor analysis. Factor 1 explained 37.7% of the variation among all of the variables; it had positive factor loadings for phenology (late maturity), biomass, and leaf area index, negative loadings for leaf area and root length, and was negatively correlated with elevation (r = -0.71). Factor 2 explained 14.5% of the variation among all of the variables; it had positive factor loadings for plant height and leaf number per tiller, negative loadings for seed yield and tiller number, and was positively correlated with longitude (r = 0.54) and average annual minimum temperature (r = 0.39). Factor 3 explained 12.8% of the variation among all of the variables; it had highly positive factor loadings for specific root length and specific leaf area, negative loadings for canopy height and mass per tiller. Correlations among phenotypic, environmental, genotypic, and geographic-origin distances were positive (r = 0.723-0.900), which suggests that ecotypic variation is an important feature of this group. This information, in conjunction with previously established Level III ecoregions, was used to delineate four adaptive zones for race C.