A method (CRRS) that combines S2 and crossbred family selection in full-sib reciprocal recurrent selection (FSRRS) is proposed. The method requires four generations per cycle in single-eared maize populations. Selection is based on performance of S2 and full-sib families by applying selection index theory. Equations to estimate the coefficients included in the index are given. These estimates are functions of the genetic and phenotypic variances and covariances among and between the two kinds of families. Comparisons of FSRRS and CRRS under equivalent amount of effort show that CRRS has some advantage over FSRRS for low heritability of the trait being selected (e.g., maize yield) and when only one or two locations with two replications are involved in the selection experiment.