Experimental design used when there are two sources of variation among experimental units, as in the presence of a bidirectional fertility gradient in field experiments.