discriminative stimulus. Under one condition the response key was transilluminated and reinforcement was contingent upon the completion of ten pecks. Under a second condition the key was darkened and reinforcement was contingent on at least 1 minute of no response. These conditions were alternated, and each condition was terminated by a reinforcement. After 4 hours of training on this multiple schedule of reinforcement (3, p. 729) the duck responded during presentation of the stimulus that

more »

... panied re? inforcement for responding and did not respond during the presentation of the stimulus that accompanied rein? forcement for not responding. A response was conditioned that was incompatible with the response of fol? lowing the imprinted stimulus. A transparent key that enabled the duck to see the imprinted stimulus while responding was mounted midway along the Plexiglas wall. The imprinted stim? ulus was presented for 1 second after each response. Although this brief presentation was too short to permit the following response, a high rate of pecking was sustained. This finding suggests that the following response is not necessary for presentation of the imprinted stimulus to function as a reinforcer. To show that the pecking response was not maintained solely by the change in illumination concurrent with pres? entation of the imprinted stimulus, an attempt was made to sustain respond? ing when the reinforcement consisted only of a change in illumination. First, a stable rate of pecking was obtained by presenting the imprinted stimulus after every tenth response (Fig. 3) . At point a, the moving stimulus was re? moved from the apparatus compartment and only the change in illumina? tion was contingent upon responding. The rate of responding declined rap? idly, and no responses were emitted for several sessions. At point b the sched? ule of reinforcement was adjusted so Fig. 2. Rate of pecking by a 3-day-old duck. Each diagonal mark represents a 40-second presentation of the imprinted stimulus after every eighth response (time was not recorded on the abscissa during the presentation of the stimulus). The duck received 750 reinforcements in 12 hours. 15 Fig . 3 . Records of pecking by a 3-day-old duck. The first record (session 8) shows the rate of responding when the imprinted stimulus was presented after every tenth response. In session 9 the moving stimulus was removed at a, and every eighth re? sponse produced only a change in illumination. In session 15 illumination was contingent on each response; at b the apparatus compartment was lit for 1 minute. At c the imprinted stimulus was presented for 1 minute, and reinforcement was then contingent on each response. 1396 that each response, rather than every tenth response, produced the change in illumination. The lights in the appa? ratus compartment were then turned on for 1 minute. Figure 3 shows that this operation did not initiate respond? ing. The moving stimulus was replaced in the box and, at point c, the lights in the apparatus compartment were again turned on for 1 minute. After the reintroduction of the imprinted stimulus, the duck began to peck. The rate of responding continued to rise during the next hour. The reinforcer in these experiments comprised a complex of events including a change in illumination, presen? tation of the moving stimulus, and following by the duck. The findings re? ported above suggest that following is not a necessary component of the reinforcement but that the imprinted stim? ulus is. Another experiment showed that the imprinted stimulus must be moving in order to control the rate of responding. A stable rate of pecking was obtained by presenting the imprinted stimulus after every tenth response. When the apparatus was adjusted so that the im? printed stimulus no longer moved dur? ing presentation, the rate of responding fell rapidly to zero. The pecking re? sponse was obtained again on the fol? lowing day by reintroduction of the moving stimulus. Abstract. A procedure is described which leads to reports of hues for two black figures, one of which is shadowed. Typically, the shadowed figure is seen as blue, the other as black.