Spatial analysis

Abstract

A central problem confronting a plant breeder when comparing genotypes in a field trial is that the yield of a genotype is markedly affected by the condition, particularly the soil moisture and fertility, of the plot in which the genotype is sown. Comparison problems increase with separation between plots. For example, Batchelor and Reed (1918), in a study on the variation of orange tree yields, noted that ‘marked soil variations occur which tend to make adjacent tree or plot yields alike’. Fisher (1935), in advocating the importance of replication and randomization in agricultural field trials, observed that ‘after choosing an area, we usually have no guidance beyond the widely verifiable fact that patches in close proximity are commonly more alike, as judged by the yield of crops, than those which are further apart’.