ing period. Analysis of variance yielded significant main effects of treatment conditions (F -7.58, d.f. 2/ 18, P < .01) and blocks of days (F -5.06, d.f. 3/27, P < .01). As the interaction was insignificant (F < 1.0 overall treatment means were compared by the Newman-Keuls procedure (3). The E trials were significantly slower (P < .05) than N or R trials. Figure 2 shows the running speeds were slower on extinction as were again slower than trials on new paper after the first three test days.

more »

... ials on reward paper displayed a more aberrant trend measured by running speed than by starting speed. The R trials were slower than E and N tri.als over the first half of testing. However, from day 7 until the end of the testing period, performance on reward paper was coincident with trials on new paper; hoth were faster than trials on extinction paper. Analysis of variance Yielded only a significant miiain effect of treatment conditions (F = 3.66. d.f. 2/ 1 8, P < .05). Again the interaction was insignificant (F = 1.74, d.f. 6/54, P > .1O) and comparison of the overall treatmenit means was made by the Newman-Keuls procedure. The E trials were sig,nificantly slower (P < .05 than N, but not R trials. These results indicate that the odor trace of a rat undergoing experimental extinction can significantly disrupt the performiance of a subsequently run animal that was continuously reinforced. This disruption has previously been termed the "pseudo-extinction" effect and was evidenced as slower starting speeds on E as compared to N and R trials and slower running speeds on E as compared to N trials. This suggests that the mere traversal of another subject is not sufficient to disrupt the SUCceeding aninmal's performance. Rather, the state of the animal laying the trace seems to be critical in the elicitation of competing behaviors within the experimental animals. The pattern of results evidenced by the two dependent variables was different. There is the possibility that the repeated testing procedure had differential effects on running than on starting times, this influencing the time course of the observed effects. Our experiment does not discriminate between qualitative and quantitative odor effects since experimentally extinguished animals were on the paper floor longer than rewarded animals. Nor does it identify the olfactory stimuli involved, particularly whether these olfactory stimuli are isolable from those 5 DECEMBER 1969 of the excretory products deposited by the ET animals. The experiment does, nonetheless, demonstrate the importance of olfactory stimuli to the pseudlo-extinction" effect. Rats can discrimiiinate odors from animals of the same species put under stress by electric shock (4). Experimental extinction is apparently a situation capable of producing the emlission of some olfactory stimulus which, when present on the paper floor of a suhsequently rIun animal, elicits some behavior which interferes with running for food reward. Such odor effects appear to he an important, potential confounding variable in studies where learning raither than the transmlission of information between conspecifics by chemical means is investigated (5). Re-suIlts froma situations involving noxious stimulation, such as electric shock or nonreward, which seem likely to increase the probability of odor emissioni, should be reevaluated becatuse of such confounding. Control for odor effects would seem desirable if interpretation of experimental outcomes is to be tinambiguous.