We quantified the mortality benefits and attributable fractions associated with engaging in physical activity across a range of levels, including those recommended by national guidelines. Data were from the Allied Dunbar National Fitness Survey, a population-based prospective cohort comprising 1,796 male and 2,122 female participants aged 16-96 years, randomly selected from 30 English constituencies in 1990. Participants were tagged for mortality at the Office for National Statistics. Cox multivariable regression quantified the association between self-reported achievement of activity guidelines--150 min of at least moderate activity per week, equivalent here to 30 or more 20-min episodes of at least moderate activity per month--and mortality adjusting for age, sex, smoking status, social class, geographical area, anxiety/depression and interview season. There were 1,175 deaths over a median (IQR) of 22.9 (3.9) years follow-up; a mortality rate of 15.2, 95% confidence interval (CI) 14.4-16.1 per 1,000 person years. Compared with being inactive (no 20-min bouts per month), meeting activity guidelines (30+ bouts) was associated with a 25% lower mortality rate, adjusting for measured confounders. If everyone adhered to recommended-, or even low-activity levels, a substantial proportion of premature mortality might be avoided (PAF, 95% CI 20.6, 6.9-32.3 and 8.9, 4.2-13.4%, respectively). Among a representative English population, adherence to activity guidelines was associated with significantly reduced mortality. Efforts to increase population-wide activity levels could produce large public health benefits and should remain a focus of health promotion efforts.