The following schedule of reinforcement as a fundamental determinant of steady state contrast in multiple schedules

Choosing a future from a murky past: A generalization-based model of behavior

Remembering the past appears critical in allowing organisms to detect order in an environment, and hence to behave in accordance with likely future events. Yet the shortcomings of remembering and perceiving typically mean that the remembered past differs from the actual past, and hence that behavior does not perfectly track the structure of the environment. Here, we outline how the process of generalization might be used to understand differences between what an organism does, and the structure of the past and potential structure of the environment. We explore how different sources of generalization - both from within the same stimulus situation, and from different stimulus situations - might be modeled quantitatively, and how predictions made by this modeling approach are supported by research. Finally, we discuss how generalization from multiple stimulus situations, longer-term experience, and from stimulus situations in the past that are not identical to the stimulus situation in the present, might contribute to our understanding of how an organism's experience translates into behavior.

Stimulus control depends on the subjective value of the outcome

Stimuli that provide information about likely future reinforcers tend to shift behavior, provided a reliable relation between the stimulus and the reinforcer can be discriminated. Stimuli that are apparently more reliable exert greater control over behavior. We asked how the subjective value (measured in terms of preference) of reinforcers associated with stimuli influences stimulus control. Five pigeons worked on a concurrent chains procedure in which half of all trials ended in a smaller reinforcer sooner, and the other half in a larger reinforcer later. In Signaled trials, the color and flash duration on the keys in the initial link signaled the outcome of the trial. In Conflicting probe trials, the color and the flash duration signaled conflicting information about the outcome of the trial. Choice in Signaled trials shifted toward the signaled outcome, but was never exclusive. In Conflicting probe trials, control was divided idiosyncratically between the 2 stimulus dimensions, but still favored the outcome with the higher subjective value. Thus, stimulus control depends not only on the perceived reliability of stimuli, but also on the subjective value of the outcome.

The effects of chlordiazepoxide and d-amphetamine during a three-component multiple schedule

Multiple schedules have been used in behavioral pharmacology research to show that a drug's effect on behavior can be a function of the schedule of reinforcement that supports that behavior. However, less research has examined whether the context of the schedule of reinforcement in a multiple schedule can change the drug's effect on behavior. We examined the effects of acute chlordiazepoxide and d-amphetamine injections on the behavior of two groups of pigeons trained on a three-component multiple schedule with identical schedules of reinforcement in the first and last components. For one group of pigeons reinforcement was unavailable during the middle component (decreased-middle-component). For the second group reinforcement rate was higher during the middle component than during the first or third components (increased-middle-component). In the decreased-middle-component group, chlordiazepoxide (3.2-32 mg/kg) decreased third-component response rates less than it decreased responding in the first component. Conversely, in the increased-middle-component group, chlordiazepoxide (3.2-10 mg/kg) decreased third-component response rates more than in the first component. In both groups, d-amphetamine did not differentially affect response rates across components. These results are consistent with previous research showing that drugs can differentially affect responding to two different schedules of reinforcement during the same session, and suggest that pharmacological preparations may be helpful in elucidating the mechanisms that control multiple schedule interactions.

Contrast and the justification of effort

When humans are asked to evaluate rewards or outcomes that follow unpleasant (e.g., high-effort) events, they often assign higher value to that reward. This phenomenon has been referred to as cognitive dissonance or justification of effort. There is now evidence that a similar phenomenon can be found in nonhuman animals. When demonstrated in animals, however, it has been attributed to contrast between the unpleasant high effort and the conditioned stimulus for food. In the present experiment, we asked whether an analogous effect could be found in humans under conditions similar to those found in animals. Adult humans were trained to discriminate between shapes that followed a high-effort versus a low-effort response. In test, participants were found to prefer shapes that followed the high-effort response in training. These results suggest the possibility that contrast effects of the sort extensively studied in animals may play a role in cognitive dissonance and other related phenomena in humans.

Session duration and the VI response function: Within-session prospective and retrospective effects

Two experiments examined the effects of session duration on responding during simple variable-interval schedules. In Experiment 1, rats were exposed to a series of simple variable-interval schedules differing in both session duration (10 min or 30 min) and scheduled reinforcement rate (7.5 s, 15 s, 30 s, and 480 s). The functions relating response rate to reinforcement rate were predominantly monotonic for the short (10-min) sessions but were predominantly bitonic for the long (30-min) sessions, when data from the entire session were considered. Examination of responding within sessions suggested that differences in the whole-session data were produced by a combination of prospective processes (i.e., processes based on events scheduled to occur later in the session) and retrospective processes (i.e., processes based on events that had already occurred in the session). In Experiment 2, rats were exposed to a modified discrimination procedure in which pellet flavor (standard or banana) predicted session duration (10 min or 30 min). All rats came to respond faster during the short (10-min) sessions than during the first 10 min of the long sessions. As in Experiment 1, the results seemed to reflect the simultaneous operation of both prospective and retrospective processes. The results shed light on the recent controversy over the form of the variable-interval response function by identifying one variable (session duration) and two types of processes (prospective and retrospective) that influence responding on these schedules.

Does contingent reinforcement strengthen operant behavior?

In Experiment 1, pigeons were trained to peck keys with equal food-reinforcement schedules in components that ended with either noncontingent or contingent transitions to a third component with a five-fold richer schedule. Response rates were higher in the initial component with contingent transitions, but resistance to prefeeding or extinction was not consistently greater. Experiment 2 also included noncontingent or contingent transitions to a signaled period of nonreinforcement. There was no effect of the contingency on transitions to nonreinforcement, but the difference in response rates maintained by contingent versus noncontingent transitions to the richer schedule was replicated. In addition, response rates were higher in components that preceded nonreinforcement than in components that preceded the richer schedule. However, resistance to extinction was greater for noncontingent transitions to the richer schedule than to nonreinforcement, implicating stimulus-reinforcer relations in the determination of resistance to change. Resistance to change was also somewhat greater for noncontingent than for contingent transitions to the richer schedule. The latter result, together with the results of Experiment 1 and related research, suggests that response-contingent reinforcement does not increase resistance to change.

Dissociation of value and response strength

Four pigeons were exposed to multiple schedules and later to concurrent-chains schedules, with terminal links that had previously been multiple-schedule components. For 2 birds, the terminal-link schedules arranged an inverse relationship between response rate and reinforcement rate; for the other 2 birds a direct corresponding relationship was arranged. Those response rates were further modified by differentially reinforcing either longer or shorter interresponse times, relative to the current means. Although the birds' initial-link responses indicated preferences for terminal links with higher rates of reinforcement, in half the cases the birds responded during the terminal links in such a way as to produce lower rates of reinforcement, rates their initial-link behavior indicated they did not prefer. That outcome is inconsistent with maximization theory, but consistent with a strengthening analysis of behavior on single-key schedules.

Reinforcement schedules: Retroactive and proactive effects of reinforcers inserted into fixed-interval performances

The responding maintained by a reinforcer depends on the relation of the reinforcer not merely to the response that produces it but also to other preceding responses. Early responses in a sequence that ends in a reinforcing consequence make smaller contributions to later response rates than more recent ones, by virtue of the longer delays that separate them from the reinforcer. This study shows that the relation between a response and a later reinforcer contributes to responding only if no other reinforcers intervene; in other words, each reinforcer blocks responses that precede it from the effects of later reinforcers. Pigeons' pecks were maintained by fixed-interval (FI) schedules of food reinforcement. When FI 60-s (short) and FI 75-s (long) schedules began simultaneously within constant 150-s cycles, long FIs did not affect short-FI performances, but short FIs eliminated the first 60 s of long-FI performances. Removing either short-FI reinforcers or short-FI stimuli showed that short-FI reinforcers and not short-FI stimuli blocked the first 60 s of the long-FI performance from the retroactive effects of the long-FI reinforcer. With FI 15-s and FI 75-s schedules, the short-FI reinforcer was followed by reduced long-FI responding, but a schedule that prevented discrimination based on time since a reinforcer eliminated this proactive effect of the short-FI reinforcer. In other words, the retroactive effects were reinforcer effects whereas the proactive effects were discriminative effects. Quantitative descriptions of variable-interval performances, in which reinforcer effects may operate in the absence of temporal discriminative effects, can be derived from these relations.

Fixed-ratio pausing: Joint effects of past reinforcer magnitude and stimuli correlated with upcoming magnitude

Pigeons responded on fixed-ratio schedules ending in small or large reinforcers (grain presentations of different duration) interspersed within each session. In mixed-schedule conditions, the response key was lit with a single color throughout the session, and pausing was directly related to the past reinforcer (longer pauses after large reinforcers than after small ones). In multiple-schedule conditions, different colors accompanied the ratios ending in small and large reinforcers, and pausing was affected by the upcoming reinforcer as well as the past one. Pauses were shorter before large reinforcers than before small ones, but they continued to be longer after large reinforcers than after small ones. The influence of the past reinforcer was modulated by the magnitude of the upcoming reinforcer; in the presence of the stimulus before the small reinforcer, the effect of the past reinforcer was enhanced relative to its effect in the stimulus before the large reinforcer. These results show that pausing between ratios is jointly determined by two competing factors: past conditions of reinforcement and stimuli correlated with upcoming conditions.

Inverse relations between preference and contrast

Pigeons were trained on a multiple schedule in which two target components with identical reinforcement schedules were followed by either the same-valued schedule or by extinction. Response rate increased in both target components but was higher in the target component followed by extinction, replicating previous findings of positive anticipatory contrast. A similar design was used to study negative contrast, in that the two target components were followed either by the same-valued schedule or by a higher valued schedule. Negative contrast occurred equally, on average, in both target components, thus failing to demonstrate negative contrast that is specifically anticipatory in nature. When the stimuli correlated with the two target components were paired in choice tests, the pattern of preference was in the opposite direction. For the positive contrast procedure, no significant preference between the two target stimuli was evident. But for the negative contrast procedure, preference favored the stimulus followed by the higher valued schedule. The results demonstrate a functional dissociation between positive and negative contrast in relation to stimulus value. More generally, the results demonstrate an inverse relation between response rate and preference and challenge existing accounts of contrast in terms of the concept of relative value.

Contrast and reallocation of extraneous reinforcers between multiple-schedule components

Four pigeons responded in components of multiple schedules in which two responses were available and reinforced with food. Pecks on the left key ("main" key) were reinforced at a constant rate in one component and at a rate that varied over conditions in the other component. When reinforcer rate was varied, behavioral contrast occurred in the constant component. On the right key ("extra" key), five variable-interval schedules and one variable-ratio schedule, presented conjointly, arranged reinforcers for responses in all conditions. These conjoint schedules were common to both multiple-schedule components-rather than unique to particular components-and reinforcers from these schedules could therefore be arranged in one component and obtained during the other component. In this way, the additional reinforcers were analogous to the "extraneous" reinforcers thought to maintain behavior other than pecking in conventional multiple schedules. Response rate on the extra key did not change systematically over conditions in the constant component, and in the varied component extra responding was inversely related to main-key reinforcement. All subjects obtained more extra-key reinforcers in whichever component arranged fewer main-key reinforcers. Consistent with the theory that reallocation of extraneous reinforcers may cause behavioral contrast, absolute reinforcer rate for the extra key in the constant component was low in conditions that produced positive contrast on the main key and high in those that produced negative contrast. Also consistent with this theory, behavioral contrast was reduced in two conditions that canceled extra-key reinforcers that had been arranged but not obtained at the end of components. Thus, a constraint on reallocation markedly reduced the extent of contrast.

Another look at contrast in multiple schedules

Recent research on multiple schedule interactions is reviewed. Contrary to formulations that view contrast as the result of elicited behavior controlled by the stimulus-reinforcer contingency (e.g., additivity theory), the major controlling variable is the relative rate of reinforcement, which cannot be reduced to some combination of stimulus-reinforcer and response-reinforcer effects. Other recent theoretical formulations are also reviewed and all are found to face serious counterevidence. The best description of the available data continues to be in terms of the "context of reinforcement," but Herrnstein's (1970) formulation of the basis of such context effects appears to be inadequate. An alternative conception is provided by Catania's concept of "inhibition by reinforcement," by which rate of responding is inversely related to the average rate of reinforcement in the situation. Such a conception is related to Gibbon's recent scalar-expectancy account of autoshaping and Fantino's delay-reduction model of conditioned reinforcement, suggesting that a common set of principles determines several diverse conditioning phenomena. However, the empirical status of such a description remains uncertain, because recent evidence shows that schedule interactions are temporally asymmetric, depending primarily upon the conditions of reinforcement that follow a schedule component.

Response-reinforcer relations and resistance to change

Behavioral momentum theory suggests that the relation between a response and a reinforcer (i.e., response-reinforcer relation) governs response rates and the relation between a stimulus and a reinforcer (i.e., stimulus-reinforcer relation) governs resistance to change. The present experiments compared the effects degrading response-reinforcer relations with response-independent or delayed reinforcers on resistance to change in conditions with equal stimulus-reinforcer relations. In Experiment 1, pigeons responded on equal variable-interval schedules of immediate reinforcement in three components of a multiple schedule. Additional response-independent reinforcers were available in one component and additional delayed reinforcers were available in another component. The results showed that resistance to disruption was greater in the components with added reinforcers than without them (i.e., better stimulus-reinforcer relations), but did not differ for the components with added response-independent and delayed reinforcement. In Experiment 2, a component presenting immediate reinforcement alternated with either a component that arranged equal rates of reinforcement with a proportion of those reinforcers being response independent or a component with a proportion of the reinforcers being delayed. Results showed that resistance to disruption tended to be either similar across components or slightly lower when response-reinforcer relations were degraded with either response-independent or delayed reinforcers. These findings suggest that degrading response-reinforcer relations can impact resistance to change, but that impact does not depend on the specific method and is small relative to the effects of the stimulus-reinforcer relation.

Within-trial contrast: pigeons prefer conditioned reinforcers that follow a relatively more rather than a less aversive event

When behavior suggests that the value of a reinforcer depends inversely on the value of the events that precede or follow it, the behavior has been described as a contrast effect. Three major forms of contrast have been studied: incentive contrast, in which a downward (or upward) shift in the magnitude of reinforcement produces a relatively stronger downward (or upward) shift in the vigor of a response; anticipatory contrast, in which a forthcoming improvement in reinforcement results in a relative reduction in consummatory response; and behavioral contrast, in which a decrease in the probability of reinforcement in one component of a multiple schedule results in an increase in responding in an unchanged component of the schedule. Here we discuss a possible fourth kind of contrast that we call within-trial contrast because within a discrete trial, the relative value of an event has an inverse effect on the relative value of the reinforcer that follows. We show that greater effort, longer delay to reinforcement, or the absence of food all result in an increase in the preference for positive discriminative stimuli that follow (relative to less effort, shorter delay, or the presence of food). We further distinguish this within-trial contrast effect from the effects of delay reduction. A general model of this form of contrast is proposed in which the value of a primary or conditioned reinforcer depends on the change in value from the value of the event that precedes it.

Capuchin monkeys, inequity aversion, and the frustration effect

Each of 4 female capuchin monkeys ("model") was paired with another female capuchin ("witness") in an adjacent cage. In Phases 1 and 3, a model could remove a grape from the experimenter's hand while the witness watched. The witness was then offered a slice of cucumber, a less preferred food. Trials alternated between subjects 50 times, defining a session. In Phases 2 and 4, both were offered cucumber. Witness rejections of cucumber were infrequent and were not dependent on whether models received grape or cucumber. When models were offered cucumber, they rejected it at higher rates than did witnesses. These results fail to support findings of Brosnan and de Waal. An account based on the frustration effect accommodates these results and those of Brosnan and de Waal.

Strict and random alternation in concurrent variable-interval schedules

Six pigeons responded on pairs of concurrent variable-interval schedules with, in different parts, four different arrangements of alternation between schedules. Following a single switching-key response, alternation was either strict or random, and the alternative presented after a switch (the postswitch alternative) was either signaled by the location of the switching key or unsignaled. Generalized-matching analyses showed little difference in behavior among the different alternation arrangements, except the usual finding of lower sensitivity of response allocation than time allocation was eliminated by arranging random alternation. Patterns of interchangeover times were similar for all arrangements except signaled random alternation. Differences in behavior preceding the different postswitch alternatives were found in the signaled random alternation procedure. Preference was biased towards the color of the signaled postswitch alternative and showed increased sensitivity when the postswitch alternative was to be the one with the higher reinforcer rate. Interchangeover times were substantially shorter when the postswitch alternative was signaled to be different from the current alternative than when it was signaled to be the same. However, when separate reinforcer ratios were calculated for the different postswitch alternatives, those effects were eliminated or greatly reduced. We suggest that, although behavior is indeed influenced by the postswitch alternative, the mechanism is indirect. That is, the distributions of reinforcers between alternatives obtained before each postswitch alternative differ when those alternatives are signaled, and those distributions are discriminated, but the same relations between choice and relative reinforcement hold irrespective of which postswitch alternative is signaled.

Behavioral contrast redux

Behavioral contrast is defined as a change in response rate during a stimulus associated with a constant reinforcement schedule, in inverse relation to the rates of reinforcement in the surrounding stimulus conditions. Contrast has at least two functionally separable components: local contrast, which occurs after component transition, and molar contrast. Local contrast contributes to molar contrast under some conditions, but not generally. Molar contrast is due primarily to anticipatory contrast. However, anticipatory contrast with respect to response rate has been shown to be inversely related to stimulus preference, which challenges the widely held view that contrast effects reflect changes in stimulus value owing to the reinforcement context. More recent data demonstrate that the inverse relation between response rate and preference with respect to anticipatory contrast is due to Pavlovian contingencies embedded in anticipatory contrast procedures. When those contingencies are weakened, anticipatory contrast and stimulus preference are positively related, thus reaffirming the view that the reinforcing effectiveness of a constant schedule is inversely related to the value of the context of reinforcement in which it occurs. The underlying basis of how the context of reinforcement controls reinforcement value remains uncertain, although clear parallels exist between contrast and the effects of contingency in both Pavlovian and operant conditioning.

"Work ethic" in pigeons: reward value is directly related to the effort or time required to obtain the reward

Stimuli associated with less effort or with shorter delays to reinforcement are generally preferred over those associated with greater effort or longer delays to reinforcement. However, the opposite appears to be true of stimuli that follow greater effort or longer delays. In training, a simple simultaneous discrimination followed a single peck to an initial stimulus (S+FR1 S-FR1) and a different simple simultaneous discrimination followed 20 pecks to the initial stimulus (S+FR20 S-FR20). On test trials, pigeons preferred S+FR20 over S+FR1 and S-FR20 over S-FR1. These data support the view that the state of the animal immediately prior to presentation of the discrimination affects the value of the reinforcement that follows it. This contrast effect is analogous to effects that when they occur in humans have been attributed to more complex cognitive and social factors.

Contrast and undermatching with regular or irregular alternation of components

Behavioral contrast and response-ratio sensitivity to reinforcement were compared in multiple schedules in which components alternated strictly or according to a pseudorandom sequence. Average component durations in the two regimes were always 60 s, and order of presentation of component alternation regimes was counterbalanced across subjects. In Part 1, the reinforcer rate in one component was reduced from 60 per hour to zero, while that in the other component was unchanged. Positive behavioral contrast occurred in the constant component in that response rates increased, but neither the reliability nor the magnitude of contrast was affected by the manner in which components alternated. Part 2 was similar, except that a number of different reinforcer rates were used in the varied component. Neither contrast nor sensitivity of response ratios to changes in reinforcer ratios depended on the regime of component alternation. Thus, the predictability in time of future reinforcement conditions, which is a feature of regular multiple scheduling, does not appear to be a determinant of multiple-schedule performance.

Shortcomings of the behavioral competition theory of contrast: Reanalysis of McLean (1992)

McLean (1992) presented significant data showing that the occurrence of behavioral contrast in a multiple schedule was correlated with shifts in the frequency of reinforcers from a second source between components of the schedule, and interpreted his results as showing that contrast was due to changes in the degree of response competition within the constant component of the multiple schedule. Reanalysis of his data shows that there was an effect of reinforcement in the alternative component of the schedule independent of the shifts in reinforcers between components. Thus, the effect of relative rate of reinforcement cannot be ascribed, at least entirely, to the mechanisms proposed by the behavioral competition theory of contrast.

Competition between stimulus-reinforcer contingencies and anticipatory contrast

Procedures used to study anticipatory contrast are conceptually similar to those used to study autoshaping, in that two target stimuli signal either higher or lower rates of reinforcement in the following components of the schedule. Despite this signal contingency, anticipatory contrast entails response rates that are higher to the target stimulus followed by the lower rate of reinforcement. To determine the relation between such effects and autoshaping, different variations of the procedure were used in which the signal contingency was presented in the absence of reinforcement in the target components themselves and in which the reinforcement schedules in the different following components were signaled by the same stimulus. Autoshaping effects of this signal contingency were demonstrated when no reinforcement was available during the target-component signals themselves. Intermediate patterns of behavior occurred when reinforcement was available during the target-component signals and when their different following schedules were correlated with the same stimulus. Attempts to isolate these signal and contrast effects functionally by using the signal-key procedure were unsuccessful. The results demonstrate that Pavlovian stimulus contingencies are in competition with the dynamics of anticipatory contrast, thus reducing its occurrence under some circumstances.

An integrative model for the study of behavioral momentum

Behavioral momentum is the product of response rate and resistance to change. The data on relative resistance to change are summarized for pigeons responding on single-key two-component multiple schedules, in the initial links of two-key multiple chained schedules, and in equivalent components of two-key serial schedules. For single-key procedures, the ratio of resistance to change in two schedule components is shown to depend on the ratio of reinforcer rates obtained in the presence of the component stimuli. For two-key procedures, the ratio of resistance to change in equivalent components is shown to depend on the ratio of reinforcer rates correlated with key locations. A model based on stimulus-reinforcer contingencies that combines the reinforcer rates in schedule components summed over key locations and reinforcer rates correlated with key locations summed over components, each expressed relative to the session average reinforcer rate, gives a good account of the data. An extension of the relative law of effect for multiple schedules fails to provide a complete account of resistance to change, but both approaches are needed for a comprehensive understanding of behavioral momentum.

Behavioral contrast: Pavlovian effects and anticipatory contrast

Two sources of behavioral contrast have been identified previously: Pavlovian stimulus-reinforcer relations and component sequence effects (anticipatory contrast). This study sought to isolate these sources of control procedurally in a four-ply multiple schedule composed of two fixed two-component sequences. Different cues were associated with the first component of each sequence, and contrast effects were studied in these target components. In Experiment 1, differential cuing of Component 2 between sequences and availability of reinforcement during target components were varied across three groups of pigeons; the stimulus-reinforcer relation between target-component cues and schedule of reinforcement in Component 2 was varied within subjects. Control by the Pavlovian relation was demonstrated under all conditions, and anticipatory contrast was not observed. In Experiment 2, target-component duration was systematically varied in the three groups of Experiment 1. Control by the Pavlovian relation was reliably obtained only when target-component behavior was unreinforced, and diminished with increases in component duration. Anticipatory contrast emerged in the two groups for which target-component reinforcement was available. These and other data indicate that Pavlovian effects in multiple schedules may be obscured when the requisite conditions for anticipatory contrast are present.

Absence of anticipatory contrast in rats trained on multiple schedules

Rats were trained on three- and four-component multiple schedules in which two of the components were correlated with identical reinforcement schedules that remained unchanged throughout training. These target components differed in terms of whether their respective following schedules were either higher or lower in value. Unlike corresponding experiments previously reported with pigeons, higher response rates occurred in the target component followed by a higher valued schedule than in the target component followed by the lower valued schedule. Overall contrast effects occurred independently of these sequential effects, but were inconsistent across subjects. The results suggest that the effects of a following schedule of reinforcement are opposite for pigeons and rats, and that one reason previous studies have often failed to show contrast effects with rats is that the effects of the following schedule in rats are in competition with contrast dynamics.

Effects of reinforcement context on choice

Two experiments investigated the effects of successive reinforcement contexts on choice. In the first, concurrent variable-interval schedules of primary reinforcement operated during the initial links of concurrent chains. The rate of this reinforcement arranged by the concurrent schedules was decreased across conditions: When it was higher than the terminal-link rate, preference for the higher frequency initial-link schedule increased relative to baseline. (During baseline, a standard concurrent-schedule procedure was in effect). When the initial-link reinforcement rate was lower than the terminal-link rate, preference converged toward indifference. In the second experiment, a chain schedule was available on a third key while a concurrent schedule was in effect on the side keys. When the terminal link of the chain schedule was produced, the side keys became inoperative. Availability of the chain schedule did not affect choice between the concurrent schedules. These results show that only when successive reinforcement contexts are produced by choice responding do those successive contexts affect choice in concurrent schedules.

Devaluation of stimuli contingent on choice: evidence for conditioned reinforcement

Pigeons were presented a concurrent-chains schedule of reinforcement that had terminal links of equal duration. The initial links of the schedule were periodically interrupted by 15-s periods during which an extinction schedule was in effect. The extinction periods were presented on either a response-contingent or a noncontingent basis. Relative response rate for the left alternative decreased when the extinction periods were accompanied by the left terminal-link stimulus. Relative response rate for the right alternative decreased when the extinction periods were accompanied by the right terminal-link stimulus. Relative response rate varied inversely with the frequency of presentation of the extinction periods but was unaffected by presence versus absence of the response contingency in the schedule of extinction-period presentation. Furthermore, relative response rate was unaffected by presentation of extinction periods accompanied by a novel stimulus. When the extinction periods were presented after reinforcement in the left terminal link instead of as interruptions of the initial links, relative response rate for the left alternative was reduced if the postreinforcement extinction period was accompanied by the terminal-link stimulus for the left chain and reduced less if the extinction period was accompanied by the terminal-link stimulus for the right chain. The results demonstrate that the correlation between the terminal-link stimulus and extinction influenced the relative response rate in the initial link.