INDUCTION IN THE PIGEON RESUU’ING FROM A DIFFERENTTALREINFORCEMENT 0F HlGH RATES ' 0F RESPONDING CONTINGENCY Thesis fot the Degree of “M. A. MICHIGAN STATE UNIVERSITY HARRY JAY CAPLAN 1971 ~””“ o-Gfl’fl .. IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 038 81 5639 L1 BRARY " Michigan $13” Univetmry LW IO’I ABSTRACT INDUCTION IN THE PIGEON RESULTING FROM A DIFFERENTIAL REINFORCEMENT OF HIGH RATES OF RESPONDING CONTINGENCY BY Harry Jay Caplan Four pigeons were used to determine the effect of a differential reinforcement of high rates of responding (DRH) contingency on the behavioral interactions of a multiple schedule. This contingency resulted in the occurrence of positive induction, i.e. an increase in the response rate in the constant component as a function of an increase in the response rate in the second variable component (the DRH component). This result failed to support Premack's 1969 hypothesis, which states that an increase in response rate, resulting from a rate con- tingency, will produce contrast. Contrast is a decrease in the response rate in one component as a function of an increase in the response rate in a second component. A subsequent reduction in the response rate in the variable component failed to produce a contrast effect in the con- stant component. The reinforcement rates in the two components were held constant. A second reduction in Harry Jay Caplan the response rate in the second component, produced during extinction, resulted in a contrast effect. W/(zué 66%.». Committee Chairman Date: 72% /al/ [f / INDUCTION IN THE PIGEON RESULTING FROM A DIFFERENTIAL REINFORCEMENT OF HIGH RATES OF RESPONDING CONTINGENCY BY Harry Jay Caplan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1971 ,r? [I / I I . _ , I A I . _. ‘ ‘ ‘1 I 1' , x J To my parents Ho Ho ACKNOWLEDGMENTS My thanks to Dr. Mark E. Rilling whose help and commentary in the preparation of this thesis has been invaluable to me. Gratitude is also extended to my wife Linda for assistance in the preparation of both the written and graphic materials of this manuscript. iii List of Tables List of Figures Introduction Method . Results Discussion List of References Appendix A . TABLE OF CONTENTS iv Page vi 10 25 32 34 LIST OF TABLES Table Page 1 Mean response rates during 81, 82, and the mean number of responses during the blackout (BO) . . . . . . . . . . . . . . . . . 23 2 t-tests on related measures . . . . . . . . . . . 24 Figure Response rate bird number Response rate bird number Response rate bird number Response rate bird number Reinforcement LIST OF FIGURES as a function 85 . . . . as a function 1191 as a function 597 as a function 738 of sessions of sessions of sessions of sessions for for for for rates as a function of sessions for birds number 85, 1191, 597, and 738 vi Page 11 13 16 18 21 INTRODUCTION The ongoing behavior of organisms is in part determined by other previous behaviors, and the conse- quences of those behaviors. Behavioral interactions may be studied through the use of multiple schedules. A multiple schedule consists of two or more independent schedules presented successively to the organism, each in the presence of a discriminable, exteroceptive stimulus. When two stimuli are to be used, one is referred to as 51 and the other as 82. An interaction is a change in the rate of responding during one stimulus as a result of a manipulation during a second stimulus. In the present study, 81 is held constant, while 82 is manipu- lated. The schedule of reinforcement in the presence of each stimulus depends upon the experimental purposes of the researcher. Usually, the same response is either reinforced or extinguished in the presence of each stimulus, although different responses may be used. Reynolds (1961b) defines four possible interactions in multiple schedules as follows: Positive contrast. The response rate in one com- ponent increases as a function of a decrease in the response rate in a second component. Positive induction. The response rate in one component increases as a function of an increase in the response rate in a second component. Negative contrast. The response rate in one component decreases as a function of an increase in the response rate in a second component. Negative induction. The response rate in one component decreases as a function of a decrease in the response rate in the second component. The most commonly studied interaction is positive behavioral contrast. The following studies illustrate procedures for its production. Reynolds (1961a) examined the effect of a reduction in the reinforcement rate during 82 on the response rate during 81. When responding during 52 was extinguished, an increase in the response rate during 81 occurred. A reduction in the number of rein- forcements per unit time during 82, resulted in an in- crease in the number of responses per unit time during 81. The reinforcement rate was held constant during 81. Terrace (1963) noted that as Reynolds reduced the rate of reinforcement during 82, a decrease in the response rate during 82 followed. Terrace proceeded to demonstrate that contrast would not occur during 81 unless there was a decrease in the response rate during 82. This was accomp- lished through the use of an errorless procedure, during which a discrimination was made between two stimuli, 81 and 82, with no responses or "errors" to 82. Since no responses occurred during 82, a response reduction during 52 was precluded. The failure to produce contrast in this multiple schedule supported Terrace's claim that a response reduction during 82 was necessary to produce contrast. Terrace (1966) found that a reduction in the response rate during 82 is sufficient to produce contrast. The frequency of reinforcement during 82 was held constant. Through the use of a multiple VI-DRL (differential reinforcement of low rates of responding) schedule, a decrease in the response rate during 82 occurred. During the VI component, the first response after a variable amount of time had elapsed was reinforced. During the DRL component the animal was reinforced only if an inter-response time (IRT) of between six to eight seconds occurred. The time varied for each subject. This supported Terrace's claim that a reduction in the response rate during 82 was sufficient to produce contrast during 81. Premack (1969) stated that a reduction in the rate of responding is not the necessary condition for the pro- duction of contrast. Rather, the necessary condition for the production of contrast is an increase in the aversive- ness of 82. Premack stated that ". . . any schedule which requires the animal to meet a rate criterion may produce contrast--whether the requirement is that the animal 'slow down' relative to the VI, or 'speed up' . . . ." Premack hypothesized that DRH, as well as DRL will produce contrast. During a DRH (differential reinforce- ment of high rates of responding) contingency, reinforce- ment occurs only when a fixed number of responses are emitted before a fixed time interval has elapsed. For example, when the rate requirement is 10 responses in 4 seconds, the pigeon is reinforced only if it responds 10 times within 4 seconds. If it does not meet the re- quirement, the clock and counter are reset, and the pigeon tries again. As a rate requirement, DRH should produce contrast as DRL has done.. The specific aim of the present study is to examine several response interactions in a multiple schedule. The procedure was an ABAC design. In phase I of the experi- ment, both 81 and S2 were reinforced on a VI 30 second schedule of reinforcement. Baseline rates of responding in the presence of each stimulus were obtained for 15 sessions. In phase II of the experiment, the effects of a response rate increase in the variable component (82) on the response rate in the constant component (81) were examined for 30 sessions. 81 was reinforced on a VI 30 second schedule of reinforcement. A DRH rate requirement was added on to the VI 30 second schedule of reinforcement during 82. In phase III of the experiment, the effects of a response rate decrease in the variable component on the response rate in the constant component were examined for 25 sessions. Both 81 and 52 were reinforced on a VI 30 second schedule of reinforcement. In phase IV of the ex- periment, the effects of a second response rate decrease during the variable component on the response rate in the constant component were examined for 15 days. 81 remained on a VI 30 second schedule of reinforcement, while 82 responding was extinguished. Except for the extinction during 82 of phase IV, reinforcement rates were held constant throughout the study. METHOD Subjects: Four experimentally naive adult female white Carneaux pigeons were used. They were maintained at 80% of their free feeding weight throughout the study. Apparatus: The experimental chamber was a standard three key Lehigh Valley pigeon box. The two side keys were covered throughout the experiment. At the rear of the chamber was an amber light. Programming and data collec- tion were carried out on Standard electromagnetic equip- ment. A cumulative recorder was used to monitor the responding. Procedure: 81 was correlated with a red light. 82 was correlated with a green light and a vertical white line, as well as an amber light illuminated at the rear of the chamber. This was done in order to make the stimulus situation, as well as the stimuli, discriminable during each component. 81 and 82 were presented in a pseudo- random series, in which a stimulus was not presented more than three consecutive times. The duration of the stimulus presentations was 90 seconds, and each stimulus presenta- tion was separated by a 5 second blackout. During the blackout, all lights, including a houselight above the center key, were darkened. Pecking the center key was the reinforced response. Reinforcement consisted of a 3 second access to a mixed grain, during which the key light was darkened. Pretraining involved shaping the subjects. The reinforcement of successive approximations of key pecking resulted in the subjects pecking the key. These two sessions terminated when the subjects received approxi- mately 20 reinforcements for pecking the key in the presence of each of the two stimuli. In phase 1, sessions 1 through 15, baseline rates of responding on VI 30 second schedules of reinforcement were obtained. The baseline rates were compared to the rates in phase II of the study. As a result of the pseudo- random series, each stimulus was presented approximately 17 times per session. The sessions terminated when 34 stimulus presentations were made. In phase II, sessions 16 through 45, 51 remained on a VI 30 second schedule of reinforcement, while 82 was changed to a variety of DRH requirements. These require- ments were accomplished through the use of a tandem schedule added on to the VI 30 second schedule of 82. When a rein- forcement became available to the pigeon, a single response no longer produced the requirement. Rather, a number of responses within a fixed period of time were required, in order to raise the food magazine. The time interval began with the pigeon's first peck. This method is similar to that used by Ferster and Skinner (1957). The number of responses, and the limited time during which they were required to occur varied according to the following rules: 1. The purpose of the DRH requirements was to raise the response rates during 82 of phase II. 2. The range of the rate requirements extended from 3 responses in .75 seconds to 11 responses in 4 seconds. 3. Over all, the subjects started at low require- ments and went to high requirements. 4. On a few occasions (3-6) subjects had a lower requirement during a session than on the previous session. If the subject did not complete a requirement, the response requirement counter reset and the pigeon began the require- ment again. Reinforcements that were not earned were not "lost," but remained available until they were either earned or carried over to the next component. In phase III, sessions 46 through 70, 82 was again on a VI 30 second schedule of reinforcement, while 81 re- mained unchanged; i.e. on a VI 30 second schedule of reinforcement. Phase IV, sessions 71 through 85, involved the extinction of responding to 82. 81 remained on a VI 30 second schedule of reinforcement. Reinforcement rates were held constant throughout the study; i.e. at about 2 reinforcements per minute. During each session the reinforcement rate during 81 was used as a guideline for the 82 rates. The reinforcement rate during 82 was kept approximately equal to the rate during 81 by offering additional reinforcement opportuni- ties during 82 through the use of manual reinforcement. This procedure was used only when greater than five re- inforcements during 82 were missed. This method is similar to that used by Yarczower et al. (1968). At the end of each session the pigeons were re- turned to their home cages and fed to 80% of their free feeding weight. The following data were collected daily: 1. The number of responses during 81. 2. The number of responses during 82. 3. The duration of SI. 4. The duration of 82. 5. The number of reinforcements during 81. 6. The number of reinforcements during 82. 7. The number of responses during the 5 second blackout. RESULTS Figures 1 through 4 show the daily response rates for the four subjects during phases I through IV. During phase II a DRH contingency was in effect during 82, while 81 remained on a VI 30 second schedule of reinforcement. The response rate during 82 increased beyond the base rate of phase I. The average increase in the 82 rate was 80% (t = 8.84, p < .01). For this and all following t-tests df 3. The 82 rate increase was accompanied by an increase in the 81 response rate. The average increase in the 81 response rate was 41.9% (t = 6.94, p = .01). This increase in the SI rate is referred to as positive induction. During phase III, both SI and 52 were on V1 30 second schedules of reinforcement. A response rate de- Crease during 82 was accompanied by a response rate de- crease during 51. The average decrease in the response rate from phase II during 82 was 33.9% (t = 5.39, E = .01). The average decrease during 81 was 14.1% (t = 2.09, not significant). The decrease in the response rates in both components is referred to as negative induction. After 25 days of phase III, the effects of the DRH manipulation had not worn off totally, for the response rates had not 10 Figure 1. Response rate as a function of sessions for bird number 85. Rates during phases 1, II, III, and IV are shown for 5, 30, 5, and 5 days, respectively. 12 _ wage: .5015 $15 $5 .._> 515 mMJDomrom Q: on mm Ow. mm Om mm ON 9 : . szmmmm .« 5.5.1.... .E 1.2.3.1.... Rivera! . 2.. 1...... . .. . .21. l. BlnNIW 33d SBSNOdSBH Om -8. .oz can- 13 Figure 2. Response rate as a function of sessions for bird number 1191. Rates during phases I, II, III, and IV are shown for 5, 30, 5, and 5 days, respectively. 11+ ._.xm.._> _>.._> Imo1_> _>.._> manomxom on OK mm o..¢ mm Om mm Cu m. : szmmmm -......Nm m . I_m . M .. o m H . 3 a S d .A O «91/ r. . N .. .er . l , O? my. .. a \: .. r m S ~.x)~ o. d . m om m N n I. 00 3 . a: 948mm . 15 returned to the original phase I base rates. The response rates during phase III, with the possible exception of bird 1191, had stabilized. During phase IV, extinction was in effect during 82. This resulted in a response rate decrease during 82. This response rate decrease during 82 was accompanied by a response rate increase during 81. This is positive behavioral contrast. The average response rate decrease during 82 was 33.2% (E = 8.52, p = .01). The average response rate increase during 81 was 16.3% (t = 3.63, p = .05). Figure 5 shows the reinforcement rates for the four subjects during phases I through IV. Once the re- inforcement rates stabilized during phase I, they remained constant throughout phases I through IV. The pigeons received approximately two reinforcements per minute. There was, however, daily variation in the reinforcement rates, due to the prOperties of the variable interval schedule. During phase IV, the reinforcement rate during 51 remained at approximately 2 per minute. The reinforce- ments were no longer available during 82, resulting in a reinforcement rate of 0 reinforcements, expressed per minute. Table 1 shows the response rate means for all of the subjects and phases. For phase I, the means are based on the last 5 days of the 15 day period. For phase II, 16 Figure 3. Response rate as a function of sessions for bird number 597. Rates during phases I, II, III, and IV are shown for 5, 30, 5, and 5 days, respectively. .17 m mmsoa menS M>1M> Ian—1; _>.._> mun—Dourom mZOMmmmm W i. O M .---~m . MM . a m .. 13 3 M .3 S M or d . . 0 M o o W. N M ..... . . S W l M 0.? Mm... 3. ’0 d 3 Nu om m N n I. 3 1‘ Om M8 02 am a. 18 Figure 4. Response rate as a function of sessions for bird number 738. Rates during phases I, II, III, and IV are shown for 5, 30, 5, and 5 days, respectively. 19 v waD@E ...xm...> _> .._> Ice 1_> 5.5 mMJDerom Emw 0N1: mm Om mm ON 9 : szmmmm a. . émm o W. I_m m y... MEON k> II 0 ¢ alnNsw 83d sasmoasr-Ja < O . \O a. 0’0 . .t‘ a. We ”a. .- xkk O m . . v . _ . s (u ,5); s) . u v . . *8 w ¢ or “2‘. ... J .. _ . . . [a a. O m _. sea as 20 the means are based on all 30 days of the period. For phase III, the means are based on the last five days of the 25 day period. For phase IV, the means are based on the first 5 days of the period. Table 1 again demon- strates the response rate changes that were evidenced in figures 1 to 4. In addition, information is given con- cerning the average number of responses during the 5 second blackouts that separated stimulus presentations during all of the four phases. All subjects showed a greater number of responses during the blackouts of phase II than phase I. All subjects showed fewer responses during the blackouts of phase III than phase II. From phase III to Phase IV, subject 85 showed a decrease in blackout re- sponses, while subject 597 showed no change, and subjects 1191 and 738 showed an increase. Table 2 contains the results of the t-tests for related measures. During 81 significant response rate increases occurred in phase II as compared to phase I (induction) and phase IV as compared to phase III (con- trast). During 82 a significant response rate increase occurred in phase II as compared to phase 1. Significant response rate decreases, during 82, occurred in phase III as compared to phase II and phase IV as compared to phase III. 21 Figure 5. Reinforcement rates as a function of sessions for birds number 85, 1191, 597, and 738. Rates during phases I, II, III, IV are shown for 5, 30, 5, and 5 days, respectively. m. $50: .5315 5.5 Imo1_> _>.._> manoonm MN Oh mm ow mm Om mm ON 0. : szmmmm O _ mm. .02 omMm .2 5m .oz 85 .m: .02 0:5 BinNIW 83d SlNBWHOUOdNIEH mm .02 omzm 23 Omam ohm muoomnsm mmouow moumh :mo: ago: >H ommgm mcfihsw mouwm HHH ommgm wqfihsw moumm .cofigmasmficme xmw mH onu wo mxmw o>Hm umma map Eoyw wousmEoo who: H ommzm mcfihsw mmumm maflpsv mopmp omcommoh one .c303m .COMumasmwcmE map mo mxmp m umma map Scam monumEoo .aOMumazmflamE zap mm may mo mxmw m umma may anm wopsmEoo opoz .GOMHmasmwcmE may mo mxww om oufiuao mg“ Eopw vowsmEoo who: HH ommcm .czonm opm muoownsm Hsow Ham How >H paw .HHH .HH .H mommam a.“ o.mN o.wv w.mN o.nv o.m m.vm n.mv v.0 m.o¢ m.om 0.0 m.m~ v.mv >H N.u w.N¢ u.H¢ m.mm ¢.vm o.m 0.0m m.mm o.m m.nm m.om N.m m.H¢ m.Hv HHH N.OH w.vo n.wv o.wo «.mv 0.5 N.om N.H¢ m.wH v.Hm N.oo N.m w.mm H.v¢ HH H.m 0.0m m.vm v.Hm N.mm o.m N.Nm o.wm v.w o.mv m.uv w.H m.om v.mm H om mm Hm Nm Hm om mm Hm om Nm Hm om mm Hm ommzm mcmoz mm» whflm 5mm vhfim HmHH ppfim mm whfim .Momv psoxumMn on“ maMase momdommoh mo gonad: came esp paw .Nm .Hm wafihsw moumh omcommok dam: .H oHan 24 .H oHnmu cH mcmoa may no woman who: mumou 6:9 Ho. Nm.m HHH ommnm ou wohmeoo mm >H ommnm mcflhsv mm mo. mo.m HHH ommam ou wohmHEoo mm >H ommzm mafipsw Hm Ho. mm.m HH ommam ow @ohmHEoo mm HHH ommnm mcfipsw Nm ucmoHHchHm uo: mo.~ HH ommnm op mopwmfioo mm HHH ommam Mawusw Hm Ho. ww.m H ommam ou wohmeoo mm HH ommgm mcHgsw Nm Ho. em.o H ommsm ow pohmmsoo mm HH ommgm mcfiusv Hm nmumwv M comHMmHEou ommzm . WOHSWNOE UmumHQH so mpmou1M .N mHan DISCUSSION The occurrence of positive induction during phase II failed to support Premack's 1969 hypothesis. This hypothesis stated that an increase in aversiveness, which could be effected through a rate contingency, in one component of a multiple schedule, would produce contrast in the second component of the multiple schedule. Although positive induction occurred in the present study, 3 ex- periments have shown that_aversive stimulation (punish- ment) during 82 will produce contrast during a nonpunished Sl. Azrin (1960) studied the effects of punishment during a variable interval schedule. He used shock of varying intensities to punish responding during the interval. He observed that the rate of responding was greater following shocked responses than it was following non-shocked responses. Azrin and H012 (1961) found similar results when studying punishment during fixed interval (FI) reinforcement, for which the pigeon was reinforced for the first response after a fixed time interval elapsed. Although multiple schedules were not used in the above studies, the results indicate that 25 26 punishing responses increase the rate of subsequent nonpunished responding. Brethower and Reynolds (1962) were able to produce contrast during 52 of a multiple variable interval schedule. When responses during 82 were shocked, there was an increase in the response rate during 81. Thus it was found that the results of the earlier studies generalized to a multiple schedule; i.e. punishment during 82 produced contrast during 81. As evidence that a DRH contingency is aversive, Premack cited Fantino (1968) as having demonstrated aversive properties of rate requirement schedules through the use of a choice paradigm. When 2 stimuli were pre- sented concurrently, a response to either stimulus re- sulted in the presentation of a second stimulus. Depend- ing upon which of the initial stimuli was responded to, a rate requirement was, or was not, effected in the presence of a second stimulus. He found that the pigeons consistently preferred the initial stimulus which would not result in a response rate requirement. However, it must be noted that a preference paradigm is not a clear measure of aversiveness, but merely a measure of prefer- ence. Aversiveness and preference are not interchange- able concepts since one stimuli could be more reinforcing or aversive than a second stimulus. A better measure of aversiveness might be the elicitation of attack by the stimulus associated with the DRH contingency. Azrin et a1. 27 (1966) demonstrated that attack against a target animal can be employed as an index of aversiveness. Premack further noted Terrace's 1966 work, in which he used a multiple VI DRL schedule of reinforcement. Terrace pro- duced contrast when there was a rate requirement during 82, while maintaining constant reinforcement rates in the two components. The absence of behavioral contrast in the present study suggests that either the DRH con- tingency, a rate requirement, was not aversive or if it were, that aversiveness is not a sufficient condition for the production of behavioral contrast. In the present study the response rate contingency was effective in raising the response rate during 82, i.e. induction occurred in all of the subjects. Travers (1963) accounts for the induction phenomenon through the process of generalization, i.e. a response learned in the presence of one stimulus generalizes to another stimulus. The increased rate of responding in the presence of 82 generalized to SI, and the response rate during 51 increased. Reynolds (1961a) produced positive induction in a multiple schedule. In his procedure, the animals were exposed to a multiple extinction-extinction schedule following a history of reinforcement on other multiple schedules. When the schedule was changed to a multiple extinction VI schedule, a response increase occurred in the VI component. This response rate increase 28 in the VI was accompanied by a response rate increase in the extinction component. This interaction is positive induction. Negative induction occurred when a schedule was changed from a multiple extinction VI to a multiple extinction-extinction. A response decrease occurred during the new extinction component. This was accompanied by a response rate decrease in the original extinction component. Similar results were obtained using VR sched- ules. In the present study, the increase in the response rate during 82 of phase II was accompanied by an increase in the response rate during 81. As with Reynold's re- sults, positive induction was obtained. It is noteworthy that studies designed explicitly to test induction phe- nomena are rare. During phase III, the reduction in the rate of responding to both 82 and 81 has shown that the increased rates during phase II were not permanent, but due rather to the temporary DRH rate requirement and the temporary induction effect. Note that the reduction in the response rate during 82 did not cause an increase in the response rate in SI; i.e. contrast. This might be interpreted as evidence against Terrace' 5 1966 suggestion that a response rate reduction is alone sufficient for the production of behavioral contrast. However, this decrease during both 81 and 82 might mean that the behavioral contrast phe- nomenon does not generalize to situations in which the 29 response rates are artificially high as a result of the previous experimental manipulations. Other experiments also failed to obtain contrast in the absence of a rate requirement, even though a response rate reduction occurred during 82. Wilke (1970) used a multiple VI FVI (free variable interval) schedule to produce a decrease in the response rate during 82. During the FVI component, free reinforcements were given to the pigeon on a variable interval schedule, i.e. there were no behavioral requirements for obtaining reinforcement. The failure to obtain contrast under this paradigm supports Premack's rate requirement theory, for the rate reduction during S2 was obtained without a rate requirement such as DRL. Reynolds (1961a) found similar results when a multiple VI DRO (differential reinforcement of other be- haviors) schedule was used. Under this paradigm, a be- havior other than responding was reinforced. Again, a rate reduction during 82, in the absence of a rate re- quirement, failed to produce contrast during 81. The failure to obtain contrast in these studies may not be supportive of Premack's theory, for negative findings are always extremely difficult to interpret. During phase IV, the extinction of responding to S2 resulted in the occurrence of behavioral contrast. This suggests that extended training on a multiple schedule A \. _,n r--- does not prevent contrast from occurring, and also 30 demonstrated that the failure to obtain contrast during phases II and III was not a result of any peculiar char- acteristics of the birds themselves. The result found in this phase was essentially a replication of Reynold's 1961a finding. Responses during the 5 second blackouts between stimulus presentations varied according to the particular phase that the subject was in. An increase occurred during phase II blackout responses as compared to phase I blackout responses. The general increase in both 81 and 82 responding accounts for this increase. A decrease occurred during phase III blackout responses as compared to phase II. The general decrease in SI and 82 responding during phase III accounts for this. During phase IV, mixed results in blackout responding occurred. 81 responding increased while 82 responding decreased during this period. At present, it must be concluded that behavioral contrast occurs only under very specific circumstances. Neither response rate reductions nor rate contingencies are alone sufficient for its production. In the present study, both a contingency produced response rate increase and a response rate decrease failed to produce contrast, when reinforcement rates were maintained. A response rate decrease, did however produce contrast. Under other circumstances, however, rate contingencies and response 31 rate reductions, with maintained reinforcement rates, have produced behavioral contrast. What these paradigms, as well as certain time out and shock paradigms, have in common, remains an open question. LIST OF REFERENCES Azrin, N. H. Effects of punishment intensity during variable interval reinforcement. Journal of the Experimental Analysis of Behavior, 1960, 3, 123-143. Azrin, N. H. and H012, W. C. Punishment during fixed interval reinforcement. Journal of the Experimental Analysis of Behavior, 1961, 4, 343-347. Azrin, N. H., Hutchinson, R. R. and Hake, D. F. Extinction induced aggression. Journal of the Experimental Analysis of Behavior, 1966, 9, 191-204. Brethower, D. M. and Reynolds, G. S. A facilitative effect of punishment upon unpunished behavior. Journal of the Experimental Analysis of Behavior, 1962, 5, 191-199. Fantino, E. Effects of required rates of responding upon choice. Journal of the Experimental Analysis of Behavior, 1968, 1;) 15-22. Ferster, C. B. and Skinner, B. F. Schedules of Rein- forcement, Appleton-Century-Crofts, Inc., 1957. Premack, D. On some boundary conditions of contrast. Reinforcement and Behavior, Academic Press (Ed. Tapp), 1969, 122-145. Reynolds, G. S. Behavioral contrast. Journal of the Experimental Analysis of Behavior, 1961(5), 4, 57-71. Reynolds, G. 8. Analysis of interactions. Journal of the Experimental Analysis of Behavior, 1961(b), 4, 107-117. Terrace, H. S. Discrimination with and without "errors.” Journal of the Experimental Analysis of Behavior, 1963(a), 6, 1-27. Terrace, H. S. Discrimination learning, the peak shift, and behavioral contrast. Journal of the Experimental Analysis of Behavior, 1968, 41, 727-741. 32 33 Travers, R. M. W. Essentials of Learning, The MacMillan Company, 1963, 212. Wilke, D. M. On some determinants of behavioral contrast. Doctoral dissertation. University of Manitoba, 1970. Yarczower, M., Gollub, L. R., Dickson, J. F. Some effects of discriminative training with equated frequency of reinforcement. Journal of the Experimental Analysis of Behavior, 1968, 11, 415-424. APPENDIX APPENDIX A and Requirement Quotients during Phase II S u b j e c t N u m b e r Rate Requirement Contingencies (Responses/Seconds) 85 1191 738 Session R/S Quot R/S Quot R/S Quot R/S Quot. 16 5/7 0.7 8/6 1.3 5/6 0.8 5/7 0.8 17 6/6 1.0 8/5 1.6 6/7 0.8 6/6 1.0 18 6/5 1.2 8/4 2.0 6/7 0.8 6/5 1.2 19 7/5 1.4 8/3.5 2.3 7/7 1.0 5/3 1.7 20 4/2 2.0 8/3.5 2.3 6/5 1.2 5/3 1.7 21 7/4 1.8 9/3.75 2.4 6/5 1.2 3/1 3.0 22 8/4 2.0 9/3.5 2.6 5/3 1.7 6/3 2.0 23 8/3.5 2.3 9/3.5 2.6 5/3 1.7 7/3 2.3 24 8/3.5 2.3 9/3.5 2.6 3/1 3.0 9/4 2.3 25 9/3.5 2.6 9/3.5 2.6 6/3 2.0 9/4 2.3 26 9/3.5 2.6 9/4.5 2.0 7/3 2.3 9/4 2.3 27 9/3.5 2.6 3/1 3.0 9/4 2.3 9/4 2.3 28 9/3.5 2.6 9/4 2.3 9/4 2.3 9/4 2.3 29 9/3.5 2.6 9/4 2.3 10/4 2.5 10/4 2.5 30 3/1 3.0 10/4 2.5 10/4 2.5 10/4 2.5 31 9/4 2.3 9/4 2.3 10/4 2.5 6/2 3.0 32 9/4 2.3 9/4 2.3 10/4 2.5 6/2 3.0 33 10/4 2.5 10/4 2.5 10/4 2.5 3/.75 4.0 34 10/4 2.5 10/4 2.5 6/2 3.0 6/2 3.0 35 10/4 2.5 9/4 2.3 6/2 3.0 6/2 3.0 36 10/4 2.5 10/4 2.5 10/4 2.5 6/2 3.0 37 11/4 2.7 10/4 2.5 3/.75 4.0 6/2 3.0 38 11/4 2.7 6/2 3.0 10/4 2.5 6/2 3.0 39 11/4 2.7 9/4 2.3 10/4 2.5 8/3 2.7 40 ll/4 2.7 9/4 2.3 6/2 3.0 3/1 3.0 41 6/2 3.0 10/4 2.5 3/1 3.0 8/3 2.7 42 10/4 2.5 6/2 3.0 8/3 2.7 9/3.S 2.6 43 10/4 2.5 6/2 3.0 4/1. 3.3 4/l.5 2.7 44 10/4 2.5 7/2.2 3.2 3/1 3.0 3/1 3.0 45 10/4 2.5 9/4 2.3 9/4 2.3 8/3 2.7 34 HICHIGQN STATE UNIV. LIBRQRIES 31293103815639