During 1956-66, 337 healthy middle-aged men in London and south-east England participated in a seven-day individual weighed dietary survey. By the end of 1976, 45 of them had developed clinical coronary heart disease (CHD) which showed two main relationships with diet. Men with a high energy intake had a lower rate of disease than the rest, and, independently of this, so did men with a high intake of dietary fibre from cereals. Energy intake reflects physical activity, but the advantage of a diet high in cereal fibre cannot be explained; there was no evidence that the disease was associated with consumption of refined carbohydrates. Fewer cases of CHD developed among men with a relatively high ratio of polyunsaturated to saturated fatty acids in their diet, but the difference was not statistically significant.