There is a positive association between cigarette smoking and coronary heart-disease (C.H.D.). In non-fatal myocardial infarction a dose-response relation persists even after the effects of additional variables have been controlled for. The relation between cigarette consumption and deaths from C.H.D. was investigated in a matched-pair case/control study. The overall simple matched-pair risk ratio (R.R.) between current smokers and non-smokers was 1.9 (95% confidence limits 1.5-2.4). For smokers of fewer than 20 cigarettes per day, the R.R. was 1.2; at a level of 21-40 cigarettes per day, the R.R. was 2.3; and for smokers of 41+ cigarettes per day, the R.R. was 4.0. A similar relation was found after adjustment for additional variables. These results suggest that the heaviest smokers could halve their risk of death from C.H.D. by reducing their tobacco consumption to an intermediate level; and that benefit of a similar order would be experienced by smokers of 21-40 cigarettes per day who cut down to less than one pack (20 cigarettes) daily.