Laboratory stressors are employed to assess the 'typical' cardiovascular reactivity to stress of a given subject. It is believed that this may assist in the diagnosis of hypertension and predict future development of blood pressure (BP) elevation. However, the internal consistency of the data obtained by laboratory stressing manoeuvres and their ability to reflect the effects of everyday stressful events occurring over a 24-h period have never been clearly established. We studied a group of ambulant normotensive and essential hypertensive in-patients who were undergoing prolonged intra-arterial blood pressure monitoring (Oxford method) to determine firstly, whether four routine laboratory stressors provided a homogeneous evaluation of subjects' reactivity and secondly, whether cardiovascular reactivity to these stressors reflected a tendency of BP to vary over a 24-h period. A significant correlation (P less than 0.01) was found between the pressor responses to mental arithmetic and the mirror drawing test and between the pressor responses to the cold pressor test and hand-grip. These correlations, however, were not close and responses to the 'mental' and 'physical' stressors were unrelated. The elevation in BP induced by laboratory stressors also did not correlate with the increase in BP accompanying stress elicited outside laboratory conditions (i.e. doctor's visit). The pressor effects of both the laboratory and the out-of-laboratory stressors showed little or no correlation with the 24-h absolute or per cent blood pressure variabilities, although for the stressors involving a mental challenge, a significant though not close correlation with daytime blood pressure variability was found. These data underline the limitations of laboratory tests in assessing cardiovascular reactivity to stress.(ABSTRACT TRUNCATED AT 250 WORDS)