[BDSP. Notice produite par INIST-CNRS uJKR0x3k. Diffusion soumise à autorisation]. The contemporary labour market is widely regarded as having become more "flexible". It is proposed that such flexibility is a characteristic of employment histories which will have effects on psychosocial status, health-related behaviour, and physical health. Recent increases in flexibility are unlikely to have accumulated over sufficient portions of individual employment histories for any effect on health to be apparent, but a "preview" of these effects may be gained from study of older cohorts. This cross-sectional study is based on data collected in the early 1970s from 5399 men and 945 women in paid work, recruited from 27 workplaces in the west of Scotland. A flexible employment history was defined as one encompassing a large number of changes between jobs. Perceived psychological stress, health behaviour (cigarette smoking, alcohol consumption, physical exercise), physiology (diastolic blood pressure, body mass index, forced expiratory volume, plasma cholesterol concentration) and current health (angina, myocardial ischaemia) were assessed. Those individuals who reported having experienced frequent job change were more likely to smoke, consume greater amounts of alcohol, and perhaps to exercise less. Similar findings were observed in both males and females, and for different age and socio-economic groups. We found no suggestion that this association was due to higher levels of psychosocial stress, and the expected consequences for health were not observed. Interpretation of these findings is not straightforward due to an uncertain direction of causation, and a possible selection bias. However, the observed relationship between frequent job changing and a higher incidence of health risk behaviours, in the absence of a relationship with poorer health, invites further research.