An increasing number of studies have used acoustic rhinometry (AR) for study of pharmacological interventions on nasal cavity dimensions in dogs and cats, but there have been no attempts to validate AR in these species. This is done in the present study. Method: We compared area-distance relationships of nasal cavities from 5 decapitated dogs (3.5-41 kg) and cats (3.8-6 kg). AR was compared to magnetic resonance imaging (MR), and a fluid-displacement method (FDM) using perfluorocarbon. Results: AR measured 88% (98-79%) (mean and 95% confidence interval) of nasal cavity volume in dogs determined by FDM, and 71% (83-59%) in cats. When minimum areas were below 0.1 cm² in dogs, and 0.05 cm² in cats, AR severely underestimated nasal cavity dimensions. AR underestimation increased with the severity of the constriction and with distance. Cross-sectional areas in the deeper parts of the cavity measured 76% (99-54%) of fluid-displacement in dogs and 52% (66-39%) in cats. AR agreed well with MR, especially in the deeper part of the cavity. Discussion: MR images showed that the nasal cavities had a very complex structure not expected to be reproduced by AR. MR could not be considered a "gold standard" because definition of the cross-sectional area of the lumen depended critically on subjective choices. Fluid-displacement produced repeatable measurements and possibly offers the most adequate reference in future evaluation of AR. Conclusion: AR underestimated what we believed were the most correct cross-sectional areas determined by fluid-displacement, especially in the deeper part of the dog and cat nasal cavities. Despite these difficulties AR has been shown to be useful to describe qualitative changes in cross-sectional area.