Nine thousand and ninety seven male members of the construction building workers trade union in Stockholm, aged between 41 and 61 years, were asked to complete a postal questionnaire between September and November 1972. The questionnaire was comprised of several psychosocial questions including life-change items for the preceding year. All subjects were followed up for a 12 to 15-month period with regard to death from all causes and hospitalization for myocardial infarction (Ml). The urban section of the sample was also followed up for other serious illnesses. The main results were as follows: (a) The life-change measurement technique did not predict near-future MI; (b) A pre-formed discord index predominantly composed of self-rated type A behaviour and dissatisfaction variables predicted a significant number of near-future MIs; (c) One individual life-change, increased responsibility at work, predicted significantly about 10 per cent of near-future Mls; (d) Elevated life-change measures for the last year indicated increased risk of near-future neurosis onset; (e) Several questions about irritability and dissatisfaction, for example hostility in queues or with slow persons and dissatisfaction with home life, predicted several kinds of illnesses; (f) Other variables, for example, having grown up as a late child in a large family, perception of financial state, mode of living, marital status and tobacco smoking, were important in illness prediction.