The efficiency of a standardised inhalation test procedure was studied by examining the reproducibility of responses to histamine and metacholine. In addition, the responses to the two agents were compared. Each set of duplicate tests was carried out on a separate day within one week, and all factors known or presumed to influence responses were carefully controlled. The results were expressed as the provocative concentration of the agent causing a 20% fall in forced expired volume in one second (PC20). Responses to histamine and metacholine were highly reproducible (coefficients of determination [r2]=0.994 and 0.990 respectively). Responsiveness to histamine correlated closely with responsiveness to metacholine (r2=0.85). There was a small but significant cumulative dose effect with metacholine (p<0.01) but not with histamine. Side effects of throat irritation, flushing, and headache were more frequent with histamine than metacholine, and were dose-related. The high level of reproducibility indicates the efficiency of the test procedure. The similar severity of effects by agents with different mechanism of action suggests that the primary cause of non-specific bronchial hyperreactivity lies at the level of bronchial smooth muscle.