During the 1980s, the North Central Regional Plant Introduction Station replaced open-pollinated samples of Cucumis sativus with samples regenerated by cage pollination. Enzyme polymorphisms, detected by horizontal starch gel electrophoresis, were evaluated to document genetic changes in 157 Cucumis accessions resulting from changes in seed-regeneration methods. Analysis of levels of sample homozygosity indicates that cage-pollinated samples are significantly more homozygous than the open-pollinated samples they replaced, but the two groups of samples differ little in overall allelic composition. The frequency of rare alleles also has not changed significantly, although three alleles were found in cage-pollinated samples that were not noted in their older counterparts. The results of this study may not be interpreted as conclusive support for the use of cage pollination for regeneration of Cucumis germplasm collections. But the comparison of within-accession levels of homozygosity suggests that the genetic integrity of individual accessions is maintained better with cage pollination than with less-controlled regeneration methods. Only a genetic comparison of original seed samples with regenerated samples can conclusively document the effectiveness of any regeneration system.