An adequate description of the pressure distribution exerted by the fluid flow on pharyngeal walls is a first requirement to enhance the understanding, modelling and, consequently, the prediction of airway collapse during obstructive sleep apnoea. From a fluid mechanical point of view, several flow assumptions can be formulated to reduce the governing flow equations. The relevance of some major flow assumptions and the accuracy of the resulting flow description with respect to obstructive sleep apnoea was investigated on a rigid geometrical replica of the pharynx. Special attention was given to the influence of geometrical asymmetry and to the position of the flow separation point. An in vitro experimental and theoretical study of steady pharyngeal fluid flow is presented for different constriction heights and upstream pressures. Pressure and velocity distributions along a rigid in vitro replica of the oro-pharyngeal cavity were compared with different flow predictions based on various assumptions. Fluid flow models were tested for volume flow rates ranging from 5 to 120 1 min(-1) and for minimum apertures between 1.45 and 3.00 mm. Two-dimensional flow models were required and predicted experimental results with an accuracy of 15%. Flow theories classically used in the case of a Starling resistor provided poor agreement.