Secondary MDS, or AL induced by the treatment of another primary disease, occurs at an average of 48-71 months after that treatment. A high percentage of the 2 MDS convert to AL. Survival of either is less than 1 year. A constellation of morphological abnormalities from all three cell lines produces a unique appearance. The 2 AL are difficult to classify by the FAB system. With the exception of cytogenetic analysis, the biology of 2 MDS/AL remains largely unexplored. Alterations of chromosomes 5 and 7 predominate, but other associated cytogenetical abnormalities are increasingly being recognized. A review of the development of 2 MDS/AL in a variety of primary diseases generates the following tentative conclusions: many of the commonly used alkylating agents, and the non-classical alkylating agent procarbazine, are leukaemogens; procarbazine is probably the important leukaemogen in the MOPP programme; cyclophosphamide appears to be a less potent leukaemogen than other alkylating agents; the method in which a drug is administered probably influences its leukaemogenic potential; the duration of therapy with a drug, or the total amount of drug delivered, is probably an important factor in leukaemogenesis; irradiation alone appears to be a weak leukaemogen; irradiation has little or no synergism with chemotherapy in leukaemogenesis; the older patient treated with leukaemogenic drugs is at substantial risk to develop a 2 MDS/AL; most studies show no plateau in the actuarial incidence of developing a 2 MDS/AL, despite lengthy follow-up. Benzene is the only chemical agent for which strong evidence of leukaemogenesis exists. Nonetheless, the similarities in the karyotypic alterations of leukaemic cells between those whose occupations expose them to chemical hazard and those who are exposed to cytotoxic agents lend support to the idea that more environmental leukaemogens have yet to be discovered. Aggressive therapy should be considered for a patient of any age with an adequate performance status and a diagnosis of secondary AL, especially if the karyotype in the malignant cell is predictive of a high response rate. The therapy of 2 MDS remains investigational. To mitigate the development of a leukaemic complication, maintenance therapy should be restricted to diseases in which its efficacy is established or to an investigational setting, and consideration of the leukaemogenic potential of equally effective regimens should be part of the therapeutic planning in the older patient.