Quantification and classification of locational effects on cotton cultivar testing programs
1990
Geng, S. | Hu, B. | Bassett, D.M.
Multiple locations are commonly used to evaluate cultivar performance in breeding programs, but the relative contribution of location to cultivar selection is rarely discussed. This paper presents a method by which locational effects can be quantified and classified. Lint yield data of upland cotton (Gossypium hirsutum L.) obtained from the U.S. Regional Cotton Variety Tests between the years 1981 and 1986 were used to quantify four characteristics of the locations which were used in testing cultivars. The four characteristics of the locations are: the ability to discriminate differences among cultivars, consistency in discrimination over years, relative potential productivity, and repeatability of the potential productivity. These characteristics which are estimated from the yield data are used to evaluate the effectiveness of the location in testing cultivars. When the estimates of the characteristics are subjected to cluster analysis, a new basis for zoning the cotton belt for the cotton testing program is revealed. The resulting regions include locations which are different from the regions derived from yield data or derived from cultivar by environmental interactions. Results of this study provide an alternative basis for possible future rezoning of the U.S. Cotton Belt for cultivar testing.
显示更多 [+] 显示较少 [-]