This experiment investigated use of the resistance to reinforcement procedure to measure the inhibitory properties of a negative discriminative stimulus (S-) following discrimination training. One group of rats initially learned a problem with Cue Light 1 as the positive discriminative stimulus (S + ) and Cue Light 2 as S-; a second group had equal training with house light on as S + and dark as S-. During subsequent testing both groups were reinforced for responding to both dark and Cue Light 2. Thus, inhibitory properties of each S- could be evaluated by both within- and between-subject comparisons, with responding to a novel stimulus as the baseline in both cases. Between- and within-subject comparisons corresponded perfectly with both showing the dark S- to be a strong inhibitor, but neither showing the Cue Light 2 S- to have inhibitory properties. A possible explanation of the latter result is that discrimination learning involves two stages. In the first, excitatory properties of S- are lost, and only then does further training result in the development of true inhibition. The light-dark problem was learned quickly, and training continued well beyond the point where responding to S-was very low. Thus, it is not surprising that strong inhibition was demonstrated. With the more difficult problem, subjects achieved good discrimination performance only near the end of training and responding to S- was still relatively high. Thus, training may not have been sufficient to get beyond the stage of reducing excitation and into the development of true inhibition.