An artificial teat containing a force transducer was constructed to study the forces exerted by the liner on the teat during milking. The objective was to develop a method of comparing the action of different liners under a variety of different milking conditions. The rigid structure of the artificial teat did not provide an accurate representation of a real teat but deformable models used in the past have not proved practical. Initial tests of the teat showed that it gave repeatable results during calibration. When placed in a liner, however, repeatable results could not be obtained if the teat–liner interface was disturbed. The forces that were measured under simulated milking conditions varied with liner wall position and with time. There was a peak and trough prior to a stable asymptotie value being attained and a noticeable time lag behind pulsation chamber pressure. This technique can be used to compare liners when a reliable method of locating the artificial teat within the liner has been developed.