Six cotton populations were advanced from the F2 to F4 generation by two methods of inbreeding: pedigree and single-seed descent (bulk). With pedigree selection, individual F3 plants of F2:3 lines were selected using a selection index based on lint weight/seed, short fiber content, and upper quartile fiber length. Selected F3:5 lines were yield-tested during 2005 and compared with a group of nonselected lines advanced to the same level of inbreeding without selection using single-seed descent. For most populations, 18 lines were tested for each inbreeding method. Inbreeding method and previous selection had no influence on line performance for most traits. In one population lint weight/seed was increased by selection. Micronaire was decreased in one population, and increased in another. Lint yield, short fiber content, fiber length, and other fiber traits were unaffected by selection. We will present data on genetic correlation among traits, and determine the impact of selection vs. nonselection on genetic variances and heritability estimates.