We have analyzed chromosome breaks in 8 patients with Fanconi anemia (FA), 42 with 'idiopathic' aplastic anemia (AA), 15 first-degree relatives of FA patients, and 13 controls. Their lymphocytes were treated in culture with three concentrations of mitomycin-C (MMC). A 60-fold increase in breaks was observed in FA patients as compared to AA patients, regardless of severity of clinical signs. The MMC-stress test was standardized to clearly differentiate FA from other pancytopenias in doubtful cases. Also, the effect of storage of MMC in solution was investigated. The data on SCEs of 12 subjects tested, 10 mo apart, showed an inverse relationship between length of storage of MMC and chromosome damage. The 10-month-old solution induced only one half as many SCEs as it induced at 4 months. Further, the usefulness and power of diepoxybutane (DEB) in detection of FA heterozygotes was investigated in 12 first-degree relatives of patients with Fanconi anemia and 12 healthy controls. The mean number of chromosome breaks per mitosis by DEB stress in obligate heterozygotes was 0.08 in comparison to 0.06 in controls. Four of twelve control subjects showed proportions of breaks almost identical to or higher than those of FA heterozygotes, ie, 0.12, 0.10, 0.10, and 0.11 breaks per mitosis. The responses of healthy controls to DEB could be separated into two groups: one with mean chromosome breaks of 0.11 per mitosis, and a second with mean breaks of 0.04 per mitosis. Thus, it appears that heterozygote detection by DEB stress of cultured lymphocytes is not unequivocal.