5. Analysis of the environmental component of genotype x environment interaction in crop adaptation evaluation.
Field Crops Res., 28, 1991, pp. 71-84
In this paper, emphasis is given to the methodology of analysing and interpreting the environmental component of genotype (G) x environmental (E) interaction analysis using seed yield data from the on-farm evaluation of six soybean lines grown in 19 environments throughout the major soybean-growing areas in Thailand.
The relative performance of plant genotypes or lines is commonly found to vary in different environments, due to the interaction of genetic and non-genetic factors. This genotype by environment (G X E) interaction confounds comparisons of genotypes with the environments used for plant yield evaluation, and complicates the selection of lines for release as commercial varieties, recommendations of cultivars for particular environments or the definition of future breeding objectives. Effective plant improvement depends on an understanding of G X E interaction.
Various methods have been used for detecting and characterizing G X E interactions.
A basic objective of the regression approach is to identify systematic variation in performance in G X E matrices, but it is only informative where G X E interactions are linearly associated with an environmental index - often not the case in crop variety trials.
Another technique used for the investigation of systematic response or pattern in G X E matrices is cluster analysis, or numerical classification.
This method has been seen as useful in summarizing patterns of genotypic performance and environmental productivity.
There has been less attention given to the methodology used to improve or simplify the interpretation of the differential response of genotypes in different environments and the environmental factors which are causing differential genotype responses in different environments.
The use of genotype yield deviations from environmental mean yield as a measure of the G X E effect in pattern analysis, and their representation in bar graphs has proved very effective in separating differential soybean line responses in different environments in Thailand. The two methods enhanced the description of the way each of the line responses differed across the environments in which they were grown.
In the soybean farm trials, biotic factors had little influence on crop yield, due to the low incidence of diseases and insect pests, and the soybeans were irrigated in the dry season. In these circumstances, differences in environment mean seed yield should be mainly related to the fertility status of the location. The classification of environments based on mean seed yield showed no particular geographic distribution, which indicated that differences in mean yield were location-specific and probably due to differences in such things as soil fertility or the moisture environment.
It has been clearly demonstrated that recommendations can be made to farmers on the best line to grow in different regions, viz., CM60 in the north and northeast and 7608 in the lower north, west and central region. It is also apparent that the promising soybean line 7608 should be released as a new cultivar because of its widespread superiority to all other cultivars in the southern regions. In addition, regional recommendations can be made for a second "back-up" cultivar, where it is felt that there may be insufficient seed supplies of the premier cultivar available to farmers for planting.
1095 92 - 4/138
Asia, India, review, semiarid tropics, ICRISAT, cropping systems, climatic analysis
[Ukrainian] [English] [Russian]