Two experiments were conducted in an attempt to provide a direct, response-independent test of the delay-reduction hypothesis of conditioned reinforcement. In both experiments, pigeons made observing responses, by pressing a treadle, for stimuli associated with the schedule component in effect . The consequences of an observing response were varied; an observing response produced: a) either the stimulus associated with the shorter component or the stimulus associated with the longer component depending on the schedule component in effect; b) the stimulus associated with the short component only; c) the stimulus associated with the long component only; or, d) neigher stimulus (no consequence). In Experiment I, naive pigeons were initially exposed to a mixed schedule with two differential reinforcement of other (ORO) behavior components; 10 seconds and 30 seconds (Phase One). In the second phase the same birds were exposed to an identical schedule, but the components were fixed time (FT) components (Phase Two). Reinforcement in both phases was six seconds access to food. In Experiment II, naive pigeons were exposed to both phases of Experiment I., but reinforcement density was altered. Each 10 second component was followed by 3 seconds of food and each 30 second component was followed by 9 seconds of food. In both experiments, differential observing behavior was maintained during the FT (Phase Two) procedure but not during the ORO (Phase One) procedure. In addition, equalizing reinforcement density (Experiment II) had the effect of altering the pattern of observing behavior but did not reverse or eliminate the preference shown for the stimulus associated with the shorter delay to reinforcement over the stimulus associated with the longer delay to reinforcement. It is suggested that some characteristic of the DRO procedure may have been responsible for the lack of differential observing. While the delay-reduction hypothesis of conditioned reinforcement was supported by the results of theFT procedure of both experiments, some amendments are required to account for the lack of differential observing during theDRO procedure. Reinforcement density appeared to have little effect upon observing behavior, but further research is advised concerning its effect upon observing response patterns.