Abstract
Summary: Methods of determining the causes of year-to-year variation in yield are discussed. The types of data required for valid analysis, the need to build up a logical argument, and the use of direct experimentation to verify conclusions drawn from regression analyses, are considered.