We review studies on hair dyes and lymphomas and multiple myeloma (MM). A computerized literature search for the years 1966 through 1996 was conducted. Data were extracted using a standardized form that recorded study design, study population, type of cases, comparison group, sources of data on personal exposure to hair dyes, method of data collection, type of exposure data collected, covariates, and results. This review identified 10 epidemiologic studies published in the English literature that examined personal use of hair dyes and lymphomas or MM. These studies include three evaluations of Hodgkin's disease, five of non-Hodgkin's lymphoma (NHL), two of lymphomas with type not specified, and six of MM. For Hodgkin's disease, one case-control study reported some positive associations with use of permanent hair dyes, whereas two cohort studies found no associations with ever use of hair dyes. For NHL and MM, several evaluations suggest associations with use of permanent dyes, particularly with duration, frequency, age at first use, and dark colors. However, these associations are not consistent within and between studies. For lymphomas with type not specified, one study was superseded by a more recent report with NHL specific data and a second study was limited by small numbers of exposed subjects. At this time, it is not possible to determine if the inconsistent associations between permanent hair dyes and NHL and MM reflect sampling variability or differences in methods between studies. Because an appreciable fraction of the population has potential exposure to permanent hair dyes, elucidation of such issues may be warranted with studies that include adequate numbers of exposed subjects and that elicit information on personal use of hair dyes over time.