An investigation of the efficacy of covert negative reinforcement (Ascher and Cautela, 1972) was replicated. Thirty Ss were randomly assigned to one of three groups. Ss in the experimental group were trained to imagine a noxious scene, then to shift to the image of a ringing bell. During the test phase, the word "bell" was used to reinforce over- or under-estimations of the diameters of circles. Two control groups received differential treatment relevant to the purpose of the study. Results of the present study failed to replicate those of the original study. Several possible explanations are discussed. (Author)