In the region of 100 cps, the attenuation, as measured by the treshhold shift, of an ear defender with a rigid frame supporting the cups with liquid‐filled cushions is very nearly the same as when the cups are supported by a flexible frame. However, when using an artificial head, the attenuation is appreciably different for the two frames, even though sponge rubber is used to simulate the flesh. If only the spring constant of the system were important, the difference in attenuation for the two frames should be about 20 dB. The actual difference for the artificial head was about 8–15 dB. It is postulated that the lack of rigidity of the human skin may be the relevant factor, since this would allow a certain amount of flow in response to the sound pressure.