RANDOMIZATION ALGORITHMS IN BASIC FOR EXPERIMENTAL-DESIGN

Six BASIC programs for randomization of treatments with respect to spa ce and time are presented. Program 1 is used to obtain randomization o f several treatments in an equal number of positions for any number of replicates such that identical treatments are not replicated successi vely in the same position. Program 2 randomly assigns different treatm ents as specified in a grid of any size either (a) without constraints or (b) so that similar treatments do not occur next to each other eit her horizontally or vertically, or (c) horizontally, vertically, or di agonally. Program 3 assigns from five to 100 different treatments in e qual proportions to a grid of specified size with treatments separated as in Program 2 (c) above. Program 4 randomly assigns any number of d ifferent treatments in equal proportions to a Latin square-like grid c ontaining row and column cells with a multiple number of the treatment s such that no identical treatments are replicated successively in the same row, column, or both row and column. Program 5 produces Latin sq uares of letters with unique numbered subscripts randomized. Program 6 makes Greco-Latin squares with an odd number of letters per side. The se programs will aid in randomization of treatments within positions a nd replicates when a degree of uniformity or spacing of treatments is desired in order to increase the power of statistical tests. Examples of program output are discussed for tests with bark beetle pheromone c omponents.