This paper challenges the assumption that mortality from stroke will remain constant or decline over the next few decades. A decline in stroke mortality could be brought about by changes in factors acting close to the time of death (period effect) or by risk factors determined by the generation into which a person is born (cohort effect). Age-specific death rates for stroke (1931-1985) in England and Wales were analyzed to estimate the influence of these different effects. There were significant effects for age, period, and cohort on mortality from stroke with significantly different age and period effects in each sex. The effect of age was linear, with an increasing mortality with age in both sexes. Cohort analysis demonstrated a deceleration away from the previous trend in the mortality rates associated with birth cohorts born after 1880, followed by an acceleration in the trend of mortality rates in cohorts born after 1910. These relative increases in risk for cohorts born after 1910 were offset by a deceleration in mortality associated with periods from around 1951-1954. Since cohort effects are likely to be associated with a lifetime increase in risk of stroke mortality, it is difficult to predict the extent of any long-term fall in stroke incidence.