The relationship between the cell density utilized in the selection for 6-thioguanine-resistant mutants induced by ethyl methanesulfonate and the recovery of mutant colonies was determined in Chinese hamster ovary cells. After an adequate phenotypic expression time of 9 days post mutagen treatment, cultures were plated in selection medium at cell densities ranging from 5 X 10(4) to 2 X 10(6) cells/100 mm dish (0.9 to 36.0 X 10(3) cells/cm2). A decline in the frequency of mutant colonies was observed with increasing cell density, and analysis of the data showed that the loss was independent of the frequency of mutants in the cell population. This loss occurred in an exponential fashion, consistent with a random-target effect. Regression analysis of the data points yielded a best-fit line defined by the equation log(y) = 2.01 - 0.26(10(-6) ) x, where y = percentage recovery and x = selection cell density (cells per 100 mm dish). This study defines the cell density-dependent loss of mutants under selection conditions, and provides a basis for further study of the influence of agents on cell contact-mediated crossfeeding and possible effects on mutation induction determinations. These results also demonstrate the importance of the use of appropriate cell densities in mutant selections.