Acoustic rhinometry now has an established place in the rhinology laboratory as a measure of nasal geometry. We aimed to investigate several aspects of technique in order to offer some guidance on preferred procedures. We studied the effects of nosepiece seal quality, nosepiece aperture diameter, angle of inclination of the wave tube (in two planes), palate position and inter-observer variation on the nasal area-distance function. One hundred nasal cavities in adults and children were examined: 50 normal and 50 pathological. Each factor was examined intensively in 20 cavities, and reproducibility data obtained on all 100 cavities. The baseline mean coefficient of variation for nasal cavity volume (V1) was 6% and for minimum cross-sectional area (MCA) was 8%. Altering the angle of incidence of the wave tube in the axial and coronal planes caused considerable change in the traces from the anterior nasal cavity, including the I- and C-notches, and affected the MCA significantly (p < 0.01, Wilcoxon signed rank test). Using a small nosepiece aperture accentuated the I-notch, and the nosepiece in some cases became the site of the minimal area. Addition of a silicone-based sealant to the standard nosepiece caused a mean reduction of 14.3% in nasal volume, if the seal quality was suspected to be suboptimal. Nasopharyngeal volume decreased by a mean of 28.6% when the palate is raised by the modified Valsalva manoeuvre, and no difference was found between quiet oral respiration and cessation of nasal respiration. Acoustic rhinometry is sensitive to minor changes in the details of technique. We recommend using an intermediate range of angles in both planes, the addition of a sealant where the nosepiece seal is suspect, use of newer improved nosepieces and synchronizing readings with either cessation of nasal respiration or with quiet oral respiration. There is a pressing need for international agreement on such details if collaboration and clinical application of acoustic rhinometry is to flourish.