Heritability estimates, accession evaluation, and digital imaging, in Distichlis spicata.

摘要

Conserving water in the landscape is critical to inhabiting the arid portions of the western United States. Native accessions of the inland form of saltgrass [Distichlis spicata var stricta (Torr.) Beetle] remained green, while turfgrass lines of blue grama, buffalograss, crested wheatgrass, and bermudagrass went dormant from lack of rainfall during the drought of 2000 and 2001 in Colorado. Since saltgrass is non-domesticated, this research selected plants for four traits needing improvement to make saltgrass more turf like. Resistance to leaf rust (Puccinia aristidae Tracy), short height, high shoot density, and high seed yield were traits that made up a selection index which ranked all 158 accessions collected from the Front Range of Colorado, the Great Basin, South Dakota, and Nebraska. The top 14 females and 12 males of these were topcrossed, and progenies were evaluated. Response to selection was recorded for all four traits. Realized heritability, narrow sense heritability from half-sib analysis, narrow sense heritability from parent-offspring regression, and broad sense heritability were very high for height and shoot density. Broad sense heritability and narrow sense heritability from half-sib analysis were high for seed yield, but narrow sense heritability from parent-offspring regression and realized heritability were moderate. A major gene for rust resistance was inferred. Negative heterosis measured on the midparent for height and seed yield were noted. Positive heterosis occurred for shoot density.;Accessions were grouped by their region of origin (four), and analyzed for the above four traits, as well as, days to flower, spread, gap, seed length, and a measure of the female head height. There were significant differences among regions for most traits. Arid region plants tended to have different values than humid region plants.;Digital imaging was used to measure spread of plants in one year's time. Correlations between camera scan and visual estimates of spread were high. Calibration of equipment is important in digital imaging, but digital imaging appeared accurate at medium levels of cover. The high rate of spread in progeny was postulated to be due to heterosis, and not a correlated response from selecting the four traits.;The indication of a major gene for rust resistance suggests ease of incorporating resistance, although durability may be reduced with a single gene. Collecting from the Front Range rather than the Great Basin and Central Plains, would be more effective in developing a turf variety because this area contains accessions with better values for turf traits. Digital imaging can discriminate plant spread when differences are so small they are indiscernible by visual estimates. Significant responses to selection and/or very high heritability estimates indicate breeding to change these traits will not be difficult.