For both leukemia and lymphoma, reported incidence rates may be substantially higher than reported mortality rates in the same populations. These differences seem to be due mainly to four factors: errors in certifying cause of death, errors in registration of incidence data, differences in definitions used in reports on cancer incidence and mortality, and deaths that were in fact not related to lymphoma or leukemia. The use of mortality data as a substitute for incidence data, or the reverse, can lead to significant errors in estimating the hazards to health associated with various forms of leukemia and lymphoma.