OF EXPERIM:ENTS~~

populations to which experiment designs pertain. The sampling of sampling units (the elements making up a population) to form experimental units (the smallest unit to which a single treatment is applied) likewise receives no discussion. The statistician appears to assume that this is in the province of the investigator and the investigator feels that the statistician should tell him what it is all about. As a consequence, these topics are not dis­ cussed. To make a start on providing a literature to fill this void, three papers have been published. They are: Federer, W. T. (1976). Sampling, blocking, and model considerations for the completely randomized, randomized complete block, and incomplete block designs. Biometrische Zeitschrift 18(7):511-525.