Spatial field variations in soybean (Glycine max [L.] Merr.) performance trials affect agronomic characters and seed composition

The objective of this study was to investigate field trends in agronomic ch aracters of soybean (Glycine max [L.] Merr,) performance trials. Field expe riments were arranged as lattice designs, and for each plot within an exper iment, an environmental covariate representing the spatial trend was calcul ated using neighbour analysis. Correlations between environmental covariate s of different characters were then used to analyse the degree of similarit y in their spatial trends. Grain yield, seed protein content, and seed size were more influenced by spatial variations than time to flowering, time to maturity, and oil content. Significant correlations between environmental covariates were found, which demonstrate the similarity of trend patterns i n different characters within particular experiments. In field trials exhib iting a high degree of spatial heterogeneity, correlations between environm ental covariates of grain yield and protein content were highly positive. T his indicates that field conditions, which promote grain yield, would also enhance protein content. Moreover, the observed trend patterns considerably affected the calculation of phenotypic coefficients of correlation between grain yield and seed protein content. The results suggest that spatial ana lysis should be applied to all characters of interest, when field condition s are not homogeneous. (C) 2000 Elsevier Science B.V. All rights reserved.