The general validity of calibration of supraaural headphones by the method of an inserted probe tube is examined. In particular, the possibility of the probe tube affecting the seal between the headphone and the ear is considered. Experimental proof that this can occur is obtained by the use of a small microphone located within the ear canal. Furthermore, attention is drawn to the fact that single cavity couplers such as the NBS9A make no attempt to simulate the ear's impedance and so a discrepancy between probe and coupler responses can be expected. In the light of these factors, some recommendations about the design of artificial ears are made.