The influence of nasal valve on acoustic rhinometry (AR) measurements was evaluated by using simple nasal cavity models. Each model consisted of a cylindrical pipe with an insert simulating the nasal valve. The AR-determined cross-sectional areas beyond the insert were consistently underestimated, and the corresponding area-distance curves showed pronounced oscillations. The area underestimation was more pronounced in models with inserts of small passage area. The experimental results are discussed in terms of theoretically calculated "sound-power reflection coefficients" for the pipe models. The reason for area underestimation is reflection of most of the incident sound power from the barrier at the front junction between the pipe and the insert. It was also demonstrated that the oscillations are due to low-frequency acoustic resonances in the portion of the pipe beyond the insert. The results suggest that AR does not provide reliable information about the cross-sectional areas of the nasal cavity posterior to a significant constriction, such as pathologies narrowing the nasal valve area. When the passage area of the nasal valve is decreased, the role of AR as a diagnostic tool for the entire nasal cavity becomes limited.