Four hundred consecutive patients referred for diagnosis of hyperventilation syndrome were studied to assess the utility of self-reported complaints for making primary diagnoses of the syndrome. One-half of the subjects were unequivocally diagnosed as positive for the disorder, the other half as negative. This diagnosis was determined by the presence or absence of two criteria with established validity: (a) recognition of major presenting symptoms during a period of voluntary hyperventilation, and (b) slow return of end-tidal CO2 levels to pre-hyperventilation baseline values after the voluntary period of overbreathing. Analyses focused on differences in presenting symptoms between those patients with and those without the syndrome. Results revealed many significant differences in frequency of specific complaints between groups. However, there was much overlap between groups with regard to all complaints. A discriminant analysis of the complaint items led to a correct classification of 66 per cent of the subjects. Our findings thus indicate that the risks of misclassification of hyperventilation syndrome are relatively large when diagnosis is solely based on presenting complaints. Consequently, reported symptoms characteristic of the disorder should be used as preliminary indications requiring further evaluation.