Claims have been made for the potential of acoustic rhinometry (AR) in the evaluation of adenoidectomy patients. Little evidence has been presented to support such claims, and evidence is accumulating that AR is inaccurate in reflecting anatomical reality in the nasopharynx. We set out to establish whether acoustic rhinometry studies could predict operative decision-making sufficiently for it to be of assistance to the clinician, despite these theoretical and practical obstacles. A total of 101 patients aged 2-13 years were examined by AR using the impulse technique. Parameters were chosen from the area-distance function to indicate nasopharyngeal volumes and areas (decongested and non-decongested). This information was compared with findings at EUA (examination under anaesthesia-obstruction categories: A-'good airway' to D-'severe obstruction'), operative decision (2 categories-'obstructive' = remove, versus 'non obstructive' = leave in situ) and parents' symptom scores. Twenty-one patients were also evaluated post-operatively. There was considerable overlap between the AR parameters in the groups classified at EUA as 'obstructive' or 'non obstructive', but this overlap diminished after decongestion. Logistic regression demonstrated that the decongested volume and area parameters were of significant predictive value with respect to operative decision (odds ratio for unit change in volume = 0.82; 95% C.I. = 0.70-0.97; p = 0.018). Parents' analogue scores for snoring and for [snoring+obstruction+ mouthbreathing] were also of significant predictive value. The presence of rhinitis diminishes the predictive value of AR. Acoustic rhinometry has potential as a pre-operative evaluation of the nasopharyngeal airway in adenoidectomy candidates, but the predictive value is low unless combined with clinical factors.