Aims: To extend and analyse follow up of mortality and cancer incidence among men who took part in the UK's atmospheric nuclear weapon tests and experimental programmes 40-50 years ago, with particular reference to multiple myeloma and leukaemia. Methods: A total of 21 357 servicemen and male civilians from the UK who participated in the tests and a control group of 22 333 male controls were followed over the period 1952-98. Analyses were conducted of mortality from various causes, and of mortality and incidence for 27 types of cancer. Results: Rates of mortality from all causes continued to be similar among test participants and controls with the longer follow up, with standardised mortality ratios (SMRs) of 89 and 88 respectively over the full follow up period. For all cancers, the corresponding SMRs were 93 for participants and 92 for controls. Mortality from multiple myeloma was consistent with national rates both for participants and controls, and the relative risk (RR) of myeloma incidence among participants relative to controls was 1.14 (90% Cl 0.74 to 1.74) over the full follow up period and 0.79 (90% Cl 0.45 to 1.38) during the extended period of follow up (1991-98). Over the full follow up period, leukaemia mortality among participants was consistent with national rates, while rates among controls were significantly lower, and there was a suggestion of a raised risk among test participants relative to controls (RR 1.45, 90% Cl 0.96 to 2.17); the corresponding RR for leukaemia incidence was 1.33 (90% Cl 0.97 to 1.84). After excluding chronic lymphatic leukaemia (CLL), which is not thought to be radiation inducible, the RR of leukaemia mortality increased to 1.83 (90% Cl 1.15 to 2.93), while that for incidence was little changed. Analysis of subgroups of participants with greater potential for exposure provided little evidence of increased risks, although the numbers of men involved were smaller and the statistical power was therefore less. Among other types of cancer, only for liver cancer incidence was there evidence of differences in rates between participants and controls in both the earlier and in the additional period of follow up. Mortality rates among test participants from causes other than cancer were generally similar to those among the controls. Conclusions: Overall levels of mortality and cancer incidence in UK nuclear weapons test participants have continued to be similar to those in a matched control group, and overall mortality has remained lower than expected from national rates. There was no evidence of an increased raised risk of multiple myeloma among test participants in recent years, and the suggestion in the first analysis of this study of a raised myeloma risk is likely to have been a chance finding. There was some evidence of a raised risk of leukaemia other than CLL among test participants relative to controls, particularly in the early years after the tests, although a small risk may have persisted more recently. This could be a chance finding, in view of low rates among the controls and the generally small radiation doses recorded for test participants. However, the possibility that test participation caused a small absolute risk of leukaemia other than CLL cannot be ruled out.