Fourteen random S[subscript]2 lines were used to investigate the effects of different maintenance procedures on preserving the genetic integrity of S[subscript]2 lines after several generations of reproduction. For each line the following materials were evaluated: (1) the original S[subscript]2 line; (2) the S[subscript]2 line that was maintained for six generations sampling 10 plants per generation; (3) the S[subscript]2 line that was maintained for six generations sampling 20 plants per generation; (4) for four of the S[subscript]2 lines, progenies were obtained after five generations of maintenance with mild selection; (5) the original S[subscript]2 lines and all maintained S[subscript]2 lines crossed with two unrelated S[subscript]6 line testers; and (6) the crosses of the S[subscript]8 lines with the two testers. These plant materials were evaluated in three field experiments;Significant differences were found among S[subscript]2 line groups for all traits except kernel row number and ear length in the inbred line per se experiment, while the hybrid experiments showed similar results. Group and individual S[subscript]2 line comparisons showed that the maintained S[subscript]2 lines per se changed significantly from their original S[subscript]2 counterparts for most traits. However, when evaluated as hybrids, there were fewer significant changes and for grain yield, only one S[subscript]2 line showed any significance. The decrease in yielding ability for the maintained S[subscript]2 lines per se compared to their original S[subscript]2 line counterparts, as a group, was much less than if the lines had been inbred by pure selfing. Most of the yield reduction was due to the restricted sample sizes used during their reproduction, or because of natural and/or unintentional selection. The nonsignificant change in combining ability for the maintained S[subscript]2-line hybrid progeny indicates that the yielding ability of the hybrids was not significantly altered during S[subscript]2 line reproduction;The maintenance of S[subscript]2 lines using the sample sizes and procedures presented in this study were ineffective in maintaining the genetic integrity of the original S[subscript]2 lines per se. The S[subscript]2 lines maintained using mild selection, however, did tend to stay more similar to their original S[subscript]2 line counterparts. The evaluation of the S[subscript]2 lines in testcross showed that all maintenance procedures were effective in maintaining the genetic integrity of the S[subscript]2 lines combining ability.