Analysis of variance designs with two or more factors in which the number of observations in each cell are not equal.