Analysing environments for plant breeding purposes as exemplified by multivariate analyses of long term wheat yields

A previous attempt to relate long term yields from the Western Australian wheat belt to climatic factors proved only partially successful. Here, principal component analysis has been used to examine the patterns of variability created by those socio-economic factors which may have obscured any underlying relationship which existed between yield and climate. In fact, these analyses revealed the existence of variation peculiar to particular groups of years, a result which could explain why many attempts to relate crop yields directly to climatic factors have proved unsuccessful. The plant breeding implications of these genotype x environment interactions are considered.