It has been suggested not only that sister chromatid exchange (SCE) induction might serve as a qualitative indicator of mutagenesis, but also that induced SCE frequencies are linearly related to induced mutation frequencies. The consistency of the relationship between SCEs and mutations was tested in the present work. Confluent Chinese hamster ovary (CHO) cells were exposed to ethylnitrosourea (ENU) and then held at confluency for various times prior to initiation of SCE and mutation assays. Cells held at confluency are typically thought to be arrested in their progression through the cell cycle, so that "S-dependent" processes such as fixation of mutations and formation of SCEs will not occur, while DNA repair processes might continue to operate. If repair processes reduce the number of SCE-inducing lesions during the holding period and, hence, reduce the subsequently determined SCE frequencies, then mutation frequencies should similarly be reduced if SCEs and mutations are related. It was observed, however, that induced SCE frequencies decreased exponentially with holding time, while mutation frequencies remained constant. Qualitatively similar results were obtained in log-phase cells. Cell cycle analysis demonstrates that confluent CHO cells can cycle, and ways are considered in which this might affect SCE and mutation frequencies. It is concluded that the decline in SCE frequency (with time) cannot be attributed solely to the presence of cycling cells in confluent cultures. It appears, therefore, that at least some forms of ENU-induced DNA damage that can lead to SCE were repaired and as such are distinct from those forms that are mutagenic. Thus SCEs are not necessarily related to mutations, because the two events may represent manifestations of different forms of DNA damage. Whether or not this represents a universal phenomenon that would hold true for agents other than ENU remains to be determined.