Experimental tests of the Kaiser effect, the stress-history dependence of acoustic emission production, show that interactions between principal stresses cannot be ignored as is commonly done when trying to use the Kaiser effect to determine in situ stress. Experimental results obtained under multiaxial stress states are explained in terms of a qualitative model. The results show that the commonly-used technique of loading uniaxially along various directions to determine stress history must be reevaluated as it cannot be justified in terms of the laboratory experiments. One possible resolution of the conflict between laboratory and field results is that the Kaiser effect phenomenon observed in cores retrieved from the earth is not the same phenomenon as is observed in rock loaded under laboratory conditions. ¿ American Geophysical Union 1993