A series of calculations, centered around a possible zoned critical in a ZPR-III-type machine, is presented and discussed. Criteria of selection for equivalent samples and optimum drivers are discussed. The conclusion is drawn that measurements on a combination of one or more zoned critical, and the careful computation of correction terms, can yield nearly all the experimental data which might be derived from a full-scale, all-sample critical without significant loss of accuracy. The major exception is an experimental value for critical mass.