Myelodysplastic syndromes (MDS) are increasingly recognized as a cause of bone marrow failure, and are at least as frequent as acute myeloid leukemias. While the overall incidence is about 2-4/100,000/year, incidence figures rise steeply with age. Incidence rates of 20-30/100,000/year in persons over 70 demonstrate that MDS are among the most common hematological neoplasias in this age group. However, due to difficulties of diagnosis and classification, patient registration in population-based registers is far from complete. As a prerequisite for truly representative statistics, future revisions of disease classification systems must incorporate MDS as a separate group of disorders. The difficulties in conducting epidemiological studies also impede the identification of risk factors for the development of MDS. Current knowledge of occupational risk factors is also reviewed here. More rapid progress in our understanding of MDS may come from recent advances in methodology that have begun to shed some light on the cytogenetic and molecular aspects of leukemogenesis in general, and MDS in particular. Non-random chromosomal changes can be found in about 50% of cases at diagnosis, but they are probably late events in the evolution of MDS, reflecting the progressive genomic instability of the premalignant clone. Proto-oncogene mutations have also been suggested to be relevant to the pathogenesis of MDS, but longitudinal studies of point mutations of the N-ras proto-oncogene revealed that such events, although often associated with rapid deterioration and transformation to AML, also appear to be late events during the course of disease. Therefore, it remains a major challenge to identify those lesions that initiate the multistep development of preleukemia. As the incidence of MDS correlates strongly with age, it is reasonable to presume that age-dependent changes of the hematopoietic system may play a role in the initiation of MDS. Aging is probably associated with a compromised marrow reserve through reduction in the size of the stem cell pool. Through increased proliferative activity, the remaining stem cells may be particularly vulnerable to mutagenic insults. Immunological attack on stem cells, mitochondrial DNA mutations, and the regulatory influence of the hematopoietic microenvironment must also be considered as possibly contributing to the early stages of MDS.