Statistics and agriculture

The relationship between statistics and agricultural research is reviewed. From the end of the 18th century until the present, three main periods are identified: from the beginnings of scientific agriculture in the Age of Improvement until the first world war, which saw the use of field experiments to give information on plant nutrition and which lead to a recognition of the need to handle variability, the interwar years, dominated by R. A. Fisher, which saw the founding of experimental design as a statistical discipline, the clearer understanding of the different contributions to variability, the application of the new ideas to new areas associated with agriculture (e.g. bioassay and multivariate analysis) and outside agriculture (e.g. medicine and industry) and the development of mathematical statistics, and the post-war years, initially concerned with an elaboration and unification of earlier ideas but, more recently, with new methodology encouraged in response to new directions in agricultural research (e.g. molecular biology, modelling), new forms of measurement provided by novel instrumentation, and by computing developments (agricultural statisticians have been prominent in developing statistical software). Over the 200 years covered by the review, agricultural production of all kinds has dramatically increased and it is held that statistics has played its part, usually indirectly, by encouraging the efficient use of limited research resources and, occasionally, by discouraging unjustified research. Increased production is currently less of a priority but many agricultural research problems remain to which statisticians can and should contribute (e.g. food quality, efficiency, environmental impact, stability of production).