Incidence of HD varies from about 0.5 per 100,000 person-years in parts of Asia to over 3 in parts of North America. In recent decades, many registries have reported slightly declining age adjusted incidence among men and women. Some lymphomas previously diagnosed as HD now would be classified as NHL, but this shift does not explain all of the decline. When analysed by age group, incidence has decreased substantially at older ages, whereas increases have been reported among young adults in some industrial countries. Less developed countries continue to show high rates in childhood. Hodgkin's disease of the nodular sclerosis subtype has increased over time, whereas HD of mixed cellularity has declined. Improved therapy for HD has led to sharply declining mortality rates, but further understanding of the role of EBV and other possible causal agents should afford opportunities for prevention. Non-Hodgkin's lymphoma stands out from most other malignancies because incidence and mortality rates have risen dramatically, steadily and almost universally during the past few decades. Incidence overall has been rising 3-4% per year. No sudden rise has occurred in specific birth cohorts or calendar year of diagnosis, although incidence rates have increased more steeply at older ages. Diagnosis of NHL has improved with time, perhaps beyond the ways considered herein, but has it improved so much more than diagnosis of other malignancies, and roughly simultaneously around the world? Although it appears that diagnostic improvements are partly responsible for the upward trend, it is likely that aetiological factors are playing an important part. Infections with HIV have started to inflate NHL incidence rates further but cannot account for the striking trend already under way for several decades. Clues should be vigorously pursued to determine the role of other known viruses, immunosuppressive states, herbicides and other chemicals in the environment, and commercial products such as hair dyes. To clarify reasons for the upward trends and to take preventive action will require a better understanding of the origins of the lymphomas through epidemiological research, including interdisciplinary approaches that can identify new viruses, host-environmental interactions and lifestyle and other exposures that alter susceptibility.