ABSTRACT
One of the objectives of an experimental investigation is the empirical determination of the functional relationship that may exist between a response variable, Y, and a set of control (or input) variables denoted by x1, x2, . . . , xk. The response Y is assumed to have a continuous distribution, and the xi’s are nonstochastic variables whose settings can be controlled, or determined, by the experimenter. These settings are measured on a continuous scale. For example, the yield, Y, of peanuts, in pounds per acre, is influenced by two control variables, x1 and x2, representing the amounts of two different fertilizers.
