Stability analysis of grain yield in barley (Hordeum vulgare) in the US mid-Atlantic Region

Stability of grain yield performance is an important characteristic in the selection of new crop cultivars. Information from cultivar trials, however, is seldom fully analysed for genotype by environment interactions and, therefore, information on stability of current crop cultivars is lacking. The objectives of this study were to investigate the stability of agronomic traits among genotypes of barley (Hordeum vulgate) across 17 environments (location-years) in Maryland (USA) from 1994 through 1997 and to examine the effect of locations and years of testing on grain yield performance in this region. Significant differences were observed among barley cultivars and experimental lines for grain yield, plant height, and heading date. Grain yield was positively correlated with plant height and negatively correlated with heading date. Genotype x environment interactions measured through regression analysis were significant for grain yield, heading date and plant height, with the environmental component having the largest effect. Most barley genotypes tested (90%) had regression slopes for grain yield that did not differ from 1.0, indicating good potential for yield response under improving environmental conditions. The most widely grown cultivar in the mid-Atlantic region, 'Nomini', had a regression slope that was higher than 1.0 for grain yield. This indicates that it tends to respond with increasingly higher yields under favorable conditions. In this study, the slope and the standard error of the slope were moderately correlated with grain yield. The genotype's coefficient of variation was not a good indicator of stability for this region. Grain yields of genotype entries common to all years and locations were correlated with corresponding yields at each of the locations and years to assess the relative performance of each location and year. Correlation coefficients across locations were relatively high (r=0.64) within each year of testing. Correlations between years for the same and across locations were generally much lower. The data presented here supports a testing program over more years rather than increased locations to fully characterise the performance of new cultivars.