This is the first of several papers that will present ideas on statistical method in agricultural research, especially in relation to the planning and analysis of experiments, based upon experience but not found explicitly discussed in standard textbooks. Indeed, some of the contents of these papers may appear to conflict with what can be found in books: closer examination is likely to disclose lack of real disagreement with the better texts. The first paper outlines a plan for the series. Thereafter, it concentrates on relations between agricultural scientists and statisticians and on some of the issues that are too easily overlooked when one of the former, often for good reason, chooses to be his own statistician.