The reinforcing properties of procaine,chloroprocaine and proparacaine in rhesus monkeys

Responding was maintained under a fixed ratio 10 schedule of intravenous cocaine (six monkeys) or pentobarbital (two monkeys) delivery during a daily 3 h session. When responding was stable, intravenous doses of procaine (0.05–3.2 mg/kg), chloroprocaine (0.05–3.2 mg/kg), proparacaine (0.01–0.4 mg/kg), or saline were substituted for the cocaine or pentobarbital for six to ten sessions. Between each substitution, responding was again maintained by cocaine or pentobarbital. Procaine and chloroprocaine maintained rates of responding exceeding saline levels in all monkeys tested, with maximum rates generated by 0.2 mg/kg. Daily intake in mg/kg increased 3–10 times a dose was increased from 0.1 to 3.2 mg/kg per infusion. Within each session, there were periods of continuous responding resulting in multiple infusions, separated by intervals of no responding of varying duration. Nevertheless, the number of infusions occurring in each of the six 30 min periods was relatively constant for both drugs. Responding maintained by proparacaine was similar or slightly above that maintained by saline except at one dose (0.025 mg/kg) in one monkey. No signs of toxicity were observed with any of the drugs. These results indicate that procaine and chloroprocaine are strong positive reinforcers but that proparacaine has minimal reinforcing properties.     

Reinforcing properties of intravenous procaine in rhesus monkeys.

The lever pressing behavior of rhesus monkeys was maintained by a fixed ratio 10 schedule of intravenous cocaine (3 monkeys) or codeine (2 monkeys) injections during 2 hour sessions. Saline or various doses of procaine hydrochloride were substituted for the baseline reinforcer for 6 consecutive sessions. Each substitution was separated by 3 or more days of cocaine or codeine reinforced responding. At one or more doses, procaine substitution resulted in response rates higher than saline control in all 5 animals. High response rates (greater than 30 injections per session) were obtained in 4 of the 5 monkeys. In addition, procaine self-administration was studied in two naive monkeys given 23 hour per day access to procaine following an initial 10 days of saline contingent operant level responding. At a dose of 0.3 mg/kg/injection, both animals initiated responding for procaine reinforcement. Drug intake varied widely from day to day, however each animal took over 1200 injections per day (over 360 mg/kg) at least once during the 30 days of access. With the exception of decreased food intake, there was little evidence for behavioral toxicity from these doses. Following a second 10 days of saline self-administration, both animals were given access to 3.0 mg/kg/injection procaine. A substantially greater intake of procaine was observed which was associated with marked toxicity.     

A comparison of d-amphetamine, l-amphetamine, and methamphetamine self-administration in rhesus monkeys

Rhesus monkeys were trained to self-administer cocaine on a fixed-ratio 10 schedule of reinforcement during a daily 3 hr session. d-amphetamine, l-amphetamine, and methamphetamine, at various dosages, were substituted for the cocaine for six consecutive sessions. The animals were returned to cocaine baseline between each test series. All three drugs were self-administered at rates higher than saline control levels. d-Amphetamine and metamphetamine were equipotent in maintaining self-administration behavior and both were approximately 4 times more potent than l-amphetamine.     

Phencyclidine self-administration in the rhesus monkey

Two experiments were performed in which rhesus monkeys self-administered phencyclidine through indwelling venous catheters. In the first experiment, monkeys trained to lever press for cocaine injections, maintained higher response rates as compared to saline control rates when phencyclidine at unit doses from1.5?25.0 μg/kg were substituted for the cocaine baseline. In the second experiment, experimentally naive monkeys spontaneously initiated lever-pressing for injections of 50 μg/kg/inj phencyclidine. In both studies the animals self-administered enough drug to produce behavioral effects resembling general anesthesia.     

Discriminative and reinforcing effects of brotizolam in rhesus monkeys

The reinforcing and discriminative stimulus effects of brotizolam, a benzodiazepine-hypnotic, were evaluated in rhesus monkeys. In one experiment, separate groups of monkeys (N=3/group) were trained to discriminate pentobarbital (10 mg/kg, IG) ord-amphetamine (0.56–1.0 mg/kg, IG) from saline, in a discrete-trials avoidance/escape paradigm. Pentobarbital (5.6–10 mg/kg), diazepam (1.0–1.7 mg/kg), and brotizolam (0.3–1.7 mg/kg) resulted in 100% drug-lever responding in all three pentobarbital-trained monkeys. Ind-amphetamine-trained monkeys brotizolam administration resulted only in saline-lever responding. In another experiment, monkeys were surgically prepared with indwelling intravenous catheters and lever pressing resulted in an injection of 0.1 mg/kg/injection sodium methohexital under a fixed-ratio 10 (FR 10) schedule. Pentobarbital (0.01–0.3 mg/kg/injection) and diazepam (0.003–0.10 mg/kg/injection) maintained responding above saline control levels when substituted for methohexital. Brotizolam (0.001–0.01 mg/kg/injection) resulted in more injections received compared to saline, but fewer injections compared to pentobarbital or diazepam. Thus, results from the present experiment suggest that brotizolam would have pentobarbital-like subjective effects. However, the abuse liability of brotizolam may be lower than that for diazepam.     

Comparison of the reinforcing efficacy of cocaine and procaine in rhesus monkeys responding under a progressive-ratio schedule

Rhesus monkeys (n=5) were prepared with chronic IV catheters and trained to lever press under a PR schedule of drug injection. The schedule consisted of five components, each made up of four trials (i.e., 20 trials total). Each trial within a component had the same response requirement. The response requirement in the first component was 120/trial and doubled in successive components to a maximum of 1920 in the fifth. A trial ended with an injection or the expiration of a 12-min limited hold (LH). The inter-trial interval (ITI) was 15 or 30 min. Following an injection or expiration of the LH, all stimulus lights were extinguished and responding had no consequence for the remainder of the trial. A session ended when either all 20 injections were self-administered or the response requirement was not met within the LH for two consecutive trials. The number of injections/session and responses/session increased with dose for cocaine (0.012–0.1 mg/kg per injection) and procaine (0.12–2.0 mg/kg per injection) at both ITI values. At the 15-min ITI, responding decreased again at higher doses in some monkeys with cocaine and in all monkeys with procaine. At maximum, cocaine maintained significantly more injections and responses/session when the ITI was 30 min than when it was 15 min. In contrast, the increase in ITI did not increase the maximum maintained by procaine. Cocaine was approximately 10-fold more potent than procaine and maintained at maximum significantly more injections and responses than procaine when the ITI was 30 min but not when the ITI was 15 min. These results are consistent with previous studies demonstrating that cocaine is a more efficacious positive reinforcer than procaine. Moreover, they extend recent findings suggesting that number of injections/session provides a measure of PR performance that is amenable to statistical analysis and may, therefore, be useful in establishing reliable differences among drugs in terms of relative reinforcing efficacy. Reliable quantification of between-drug differences in reinforcing efficacy can enhance not only estimates of relative abuse liability but also pharmacological analysis of central mechanisms mediating reinforcing effects.     

Cocaine reinforced progressive ratio performance in the rhesus monkey.

A series of experiments were conducted to determine the effectiveness of a progressive ratio (PR) procedure in measuring the relative reinforcing efficacy of several intravenous doses of cocaine. In Experiment 1, utilizing much smaller increases in the ratio requirement than previously reported, the animals generally displayed increases in breaking point with increases in the cocaine unit dose up to 0.4 mg/kg/inj. The highest dose studied (0.8 mg/kg/inj.) engendered breaking points lower than the 0.4 mg/kg dose but higher than the remaining lower doses. Experiment 2 was conducted utilizing the same reinforcement schedule as in Experiment 1 but with liquid Tang as the reward. The results demonstrated that this procedure would function to discriminate reinforcing strength with a more traditional reward. Experiment 3 examined a more expedient procedure to see if results similar to those seen in Experiment 1 could be obtained in a shorter period of time. However, the shorter procedure engendered excessive intrasubject variability, suggesting that some intermediate level of baseline experience between the 5-7 days used in Experiment 1 and the 50 reinforced responses used in Experiment 3 would be necessary to obtain consistent breaking point-unit dose functions.     

Negative reinforcing properties of morphine-antagonists in naive rhesus monkeys

Rhesus monkeys were trained to press a lever to extinguish a light associated with a drug infusion scheduled to occur 30 sec after the onset of the light. Each response during the light period terminated the light for a 1-min time-out period (avoidance); a response during the infusion terminated the infusion (escape). Under these conditions the monkeys tolerated a high number of saline infusions. Saline was replaced by different unit doses of nalorphine, cyclazocine, naloxone, cocaine, codeine, pentazocine and propiramfumarate each for six successive daily 2-h sessions.Infusions of nalorphine (unit doses from 500 to 10 mcg/kg/infusion) and cyclazocine (10 to 2.5 mcg/kg/infusion) generated and maintained avoidance/escape behavior, while infusions of naloxone (100 to 5 mcg/kg/infusion), cocaine, codeine, pentazocine and propiramfumarate (all 50 mcg/kg/infusion) were tolerated by the subjects.The results show that in rhesus monkeys with no drug experience prior to the experiment the morphine-antagonists nalorphine and cyclazocine but not naloxone have negative reinforcing properties.Dedicated to Prof. Dr. Dr. W. Wirth on the occasion of his 75^th birthday.Parts of these experiments have been reported at the First International Conference on Narcotic Antagonists, Warrenton, Virginia, November 26–29, 1972.     

Effects of several benzodiazepines,alone and in combination with flumazenil,in rhesus monkeys trained to discriminate pentobarbital from saline

The purpose of the present study was to further investigate the relationship between the DS effects of PB and those of benzodiazepines (BZs) and to begin to collect pharmacological information concerning receptor mechanisms involved in this behavioral effect of BZs. Rhesus monkeys (n=3), trained to discriminate pentobarbital (PB; 10 mg/kg, IG) from saline under a discrete-trials shock avoidance procedure, were given IG diazepam (0.3–10 mg/kg), chlordiazepoxide (1.0–30 mg/kg), or etizolam (0.3–10 mg/kg) alone and in combination with flumazenil (0.01–1.7 mg/kg, IM). Flumazenil was administered 10 min prior to the administration of saline, PB or the BZs. All three BZs fully substituted for PB in all monkeys. Diazepam was the most potent with a mean ED₅₀ of 0.81 mg/kg (SEM=0.04) while chlordiazepoxide was the least potent (mean ED₅₀=5.78 mg/kg, SEM=1.22 mg/kg). The ED₅₀ for etizolam was 1.22 mg/kg (SEM=0.37 mg/kg). Pretreatment with flumazenil (0.01–1.0 mg/kg) resulted in a dose-related parallel shift to the right in the dose-response function for PB-appropriate responding in all monkeys for all three BZs. The mean (n=3) pK_B value with 0.1 mg/kg flumazenil was 6.51 (SEM=0.42) for diazepam and 6.57 (SEM=0.17) for chlordiazepoxide. This value could not be calculated for etizolam because only one monkey was tested with 0.1 mg/kg flumazenil. However, the mean pK_B for etizolam considering all monkeys and all doses of flumazenil was 6.58 (SEM=0.47). Apparent pA₂ values for flumazenil with diazepam were 6.02 for one monkey and 7.11 for another. All three BZs tended to increase average latency to respond. Apparent pK_B and pA₂ analysis may prove useful for elucidating receptor mechanisms involved in the behavioral effects of BZs.     

Stimulus control by diazepam of behavior maintained under fixed-ratio stimulus-shock termination schedules in rats

The stimulus control of behavior by diazepam (1.0 mg/kg) was investigated where responding was maintained under fixed-ratio (FR) schedules of stimulus-shock termination in rats. The size of the FR requirement was either 1, 5, 10, or 20 responses. At each FR requirement, dose-effect curves were determined for diazepam, flurazepam, pentobarbital and cyproheptadine. Diazepam-like discriminative stimuli were produced by flurazepam and pentobarbital but not by cyproheptadine. The magnitude of the FR requirement had no significant effect on the dose-effect curves for percentage of responses emitted on the diazepam-appropriate choice lever for any of the four drugs. On the other hand, the effects of these drugs on rates of responding depended on the magnitude of the FR requirement. None of the drugs altered response rates under the FR1 schedule. Diazepam tended to increase response rates under the FR5 schedule, but had no effect or decreased rates under the FR10 and FR20 schedules. Flurazepam and pentobarbital predominantly decreased rates at FR requirements of 5, 10 or 20 responses. Cyproheptadine had no significant effect on response rates at any schedule parameter. Together with previous reports, the present results indicate that the discriminative effects of diazepam are similar under schedules employing noxious (this study) or non-noxious (other reports) consequences, even though the effects on response rates of diazepam-like drugs differ depending on the schedule of reinforcement and consequent event maintaining the behavior.     

Drug dependence potential of viloxazine hydrochloride tested in rhesus monkeys

The drug dependence potential of viloxazine was tested in 5 experiments on rhesus monkeys. In gross behavioral observation of normal monkeys the acute CNS effects of the drug were found to be very weak. Decrement of spontaneous motor activity and occasional eye-closing were observed with single doses higher than 16 mg/kg IV, IM and 128 mg/kg PO, while convulsions and death occured at 64 mg/kg IV and IM. Viloxazine did not suppress the morphine and barbital withdrawal signs in monkeys that had been made physically dependent on these drugs and withdrawal. In the test for physical dependence by repeated administration of the drug at 16 mg/kg IM twice daily for 31 days in normal monkeys, no observable withdrawal sign was developed in the naloxone precipitation and natural withdrawal tests. In intravenous self-administration experiments, a weak reinforcing effect was demonstrated in some monkeys, but the effect was extremely weak. Thus, viloxazine was found to be physical dependence-free and its overall dependence potential was regarded as very low.     

The reinforcing property of ethanol in the rhesus monkey

Rhesus monkeys received intravenous injections of ethanol during daily sessions contingent on their presses on an available lever. Under the standard conditions, when each response on the lever during a 3-h period each day resulted in an i.v. injection of 0.1 g/kg ethanol, the monkeys made between 30 and 50 responses/session and developed blood ethanol levels of approximately 400 mg%. Under this and other conditions of response-contingent delivery of ethanol, a negatively accelerated pattern of self-injection within sessions was demonstrated. Variations in the dose per injection (0.05–0.2 g/kg/injection) resulted in changes in the rate of lever-pressing; the number of self-injections was inversely related to dose. Ethanol intake increased only slightly with increased dose per injection. Noncontingent administration of various doses of i.v. ethanol immediately prior to a daily session decreased the number of responses; the total amount of ethanol administered (contingent plus noncontingent), however, remained constant over a pretreatment dose range of 1 to 3 g/kg. When access time to ethanol was increased from 3 to 6 h/day, the total amount of ethanol taken increased slightly. However, the blood ethanol levels at the end of a 6-h session closely approximated those obtained following 3-h sessions, indicating that during the last 3–4 h of the 6-h sessions, the rate of ethanol intake closely matched the rate of ethanol elimination.     

Cocaine,d-amphetamine,and pentobarbital effects on responding maintained by food or cocaine in rhesus monkeys

The effects of IM injections of cocaine, d-amphetamine, and pentobarbital were studied in rhesus monkeys whose lever-press responding was maintained under a second-order fixed-interval, fixed ratio schedule of reinforcement. Within each session, fixed-interval components, ending with the IV injection of 30 g/kg cocaine (one group of monkeys) or the delivery of a 300 mg food pellet (second group of monkeys), alternated with fixed-interval components ending without an injection of cocaine or the delivery of food (extinction). Drug pretreatments generally caused comparable dose-related decreases in the overall rates of responding reinforced either by cocaine or by food. Response rates during extinction usually increased and then decreased as the dose of each drug increased. An analysis of the drug effects on response rates in different temporal segments of the fixed intervals showed that in both the reinforcement and extinction components, the normally low control rates of responding which occurred earlier in the intervals were usually increased, while higher control rates which occurred later in the intervals were increased less or decreased. Thus, the effects of these drugs were relatively independent of the reinforcing event (food or cocaine) and tended to depend more on the ongoing rate of responding under these conditions.     

The ascending limb of the cocaine dose-response curve for reinforcing effect in rhesus monkeys

Rationale. To date, the literature on the intravenous self-administration of cocaine by laboratory animals lacks a compelling demonstration of an ascending limb to the dose-response function. It has been argued that previous demonstrations of an ascending limb are confounded by the extinction process. Objective. The objective was to examine the relationship between cocaine dose and intravenous self-injection frequency at the low end of the cocaine dose range (0.03–0.00075 mg/kg per injection). Methods. Three adult rhesus monkeys were given the opportunity to self-inject cocaine on a fixed-ratio 1 schedule of reinforcement with no timeouts between injections. Single cocaine doses were presented for between 13 and 27 consecutive 2-h sessions in the order of 0.03, 0.01, 0.003, 0.0015_a, 0.00075, and 0.0015_b mg/kg per injection. Results. An ascending limb of the cocaine dose-response curve was found to exist between the doses of 0.00075 and 0.003 mg/kg per injection. Conclusions. The fact that response rate increased from 0.00075 to 0.0015_b mg/kg per injection, and remained stable at this intermediate level, negates the possibility that responding at 0.0015_b mg/kg per injection is an artifact of experimental extinction. The finding that significantly less cocaine was taken at 0.0015_b mg/kg per injection than at higher doses demonstrates that satiety was not the mechanism by which cocaine intake was regulated on the ascending limb of the dose-response curve. Electronic Publication     

Experimental studies of the abuse potential of d,l-glaucine·1.5-phosphate in rhesus monkeys

d-Glaucine is an alkaloid derived from Glaucium flavum, which is as effective as codeine as an antitussive. d,l-Glaucine·1.5 phosphate is a synthetic compound related to d-glaucine. The ability of ,l-glaucine·1.5 phosphate to maintain responding in rhesus monkeys was assessed in 2 procedures. In the first study responding was maintained under a fixed-ratio 10 schedule of codeine delivery during daily 3-hr sessions. When d,l-glaucine·1.5 posphate (0.05–0.4 mg/kg) was substituted for codeine, responding was not maintained. In the second procedure, monkeys given 23-hr/day access to 0.5–1.0 mg/kg under a fixed-ratio schedule did not self-administer d,l-glaucine·1.5 phosphate above saline levels even after a 21-day period of programmed injections. Following the period of programmed injections, there were no signs of opiate withdrawal following the administration of naloxone. These results indicate that the abuse potential of td,l-glaucine·1.5 phosphate is low relative to codeine.     

Fenfluramine and N-ethyl amphetamine: Comparison of the reinforcing and rate-decreasing actions in the rhesus monkey

N-ethyl amphetamine HCl (NEA) and fenfluramine HCl (meta-trifluoromethyl N-ethyl amphetamine) were evaluated as reinforcers in rhesus monkeys that had been previously trained to press a lever using food presentations and cocaine HCl injections as reinforcers. Each daily session consisted of episodic opportunities to obtain reinforcers under a fixed-ratio schedule of 30. A drug period was interpolated between two periods in which lever-press responding was maintained by food presentations. Compared to saline, none of the drugs altered the rate of responding in the food periods which preceded the drug sessions, indicating the absence of residual response-disrupting drug actions from previous sessions. However, NEA and fenfluramine self-injection resulted in dose-related decreases in response rates during the food periods which immediately followed the drug sessions. Cocaine HCl (30 g/kg/injection) maintained high response rates at over one response/second during the drug periods, as did the same dose of NEA. Doses of 10 and 100 g/kg/injection of NEA as well as all doses of fenfluramine HCl (10 through 300 g/kg/injection) maintained rates that were not different from those associated with saline injections. These results substantiate and extend earlier findings with fenfluramine and indicate that its failure to act as a reinforcer is attributable to its meta-trifluoromethyl group.Supported in part by USPHS Grant DA-00154.R. E. T. was an NIH Predoctoral Trainee in Pharmacology, USPHS Grant GM-00198.     

Ethanol drinking by rhesus monkeys as a function of concentration

Ethanol deliveries maintained fixed-ratio (FR) responding of three rhesus monkeys during daily 3-h sessions. At FR values of 8 or 16, ethanol concentration was varied in the sequence 0 (water), 8, 11.3, 16, 22.6, 32, 8, and 0% (w/v). As the ethanol concentration increased, number of liquid deliveries decreased, although intake of ethanol (g/kg/session) increased somewhat. Blood ethanol levels were usually greater than 200 mg% and occasionally greater than 300 mg%.These data were reported to the Committee on Problems of Drug Dependence (1976) and were part of a doctoral dissertation by J. E. Henningfield     

Parametric analysis of cocaine self-administration under a progressive-ratio schedule in rhesus monkeys

The present study was designed to investigate parameters and quantitative analysis of cocaine self-administration under a progressive-ratio (PR) schedule of reinforcement, with the goal of enhancing the resolution of PR schedules for measuring reinforcing efficacy. Six rhesus monkeys were prepared with chronic intravenous catheters and trained to self-administer cocaine under a PR schedule. The schedule consisted of five components, each made up of four trials (i.e., 20 trials total). Each trial within a component had the same response requirement. Three initial response requirements were tested: fixed-ratio (FR) 60, FR 120 and FR 240. The response requirements doubled in successive components to a maximum of FR 960, FR 1920 or FR 3840, respectively, in the fifth component. A trial ended with an injection or the expiration of a 12- or 24-min limited hold (LH). The inter-trial interval (ITI) was 15 or 30 min. Four dependent measures were assessed: break point (last FR completed), injections/session, responses/session and response rate (responses/s). For the three initial FRs, the break point, number of injections/session, responses/session and rate increased with dose of cocaine (0.013–0.1 mg/kg per injection) at both ITI/LH values. At the ITI15/LH12, responding decreased at higher doses, i.e., the dose-response functions were biphasic. In contrast, at the ITI30/LH24, responding reached an asymptote at higher doses. In general, cocaine maintained significantly higher break points, injections/session, responses/session and rate at ITI30/LH24 than at ITI15/LH12. However, at both ITI/LHs, as initial FR was increased, injections/session at the higher doses decreased while break point, total responses/session and rate did not change. A ceiling on performance, as assessed by break point, total responses/session and response rate, may have limited the number of cocaine injections an animal could take in a session. The results of this study indicate that optimal conditions for measuring the reinforcing efficacy of cocaine were obtained at the longer ITI/LH and at initial FRs above 60.     

Relationship between the discriminative stimulus effects and plasma concentrations of intramuscular cocaine in rhesus monkeys

The relationship between the discriminative stimulus effects and plasma pharmacokinetics of cocaine was evaluated in six rhesus monkeys trained to discriminate cocaine (0.4 mg/kg, IM) from saline under a FR30 schedule of food presentation. The same monkeys were tested in two procedures. In a cumulative dosing procedure, five cumulative doses of cocaine (0.013–1.3 mg/kg) were administered and discriminative stimulus effects were evaluated 10 min after the administration of each dose. Cocaine plasma concentrations were measured in separate sessions using the same doses and interdose intervals. In a single dosing procedure, the time-courses of the discriminative stimulus effects and plasma concentrations of cocaine were assessed after the administration of cocaine (0.4 mg/kg). A close correspondence between cocaine's discriminative stimulus effects and plasma concentrations was obtained in both procedures. Cocaine was virtually undetectable in plasma at doses that produced saline-appropriate responding (0.013 and 0.04 mg/kg), whereas increasing plasma concentrations were measured at doses that produced primarily cocaine-appropriate responding (0.13 mg/kg or higher). The time-course of the discriminative stimulus effects of cocaine was characterized by a rapid onset (within 1–3 min post-cocaine) and offset (within 20–60 min post-cocaine). Peak plasma levels were obtained at 10 min post-cocaine. No differences in plasma concentrations were found 10 min after the administration of a cumulative versus a single dose of cocaine 0.4 mg/kg (mean, 75.8 and 74.0 ng/ml, respectively). Cocaine plasma concentrations lasted longer than its discriminative stimulus effects. The results of the present study confirm that the cumulative dosing procedure used yields plasma concentrations of cocaine that are similar to the concentrations obtained after single cocaine dosing.In Memoriam: Xavier Lamas, MD, phD; August 26, 1995; Mt. EverestThis work was supported by in part by grants DA 02159, DA 04059, DA 07252 and KO award DA 00101 from the National Institute on Drug Abuse, NIH. Xavier Lamas was supported by a grant from the Ministry of Education and Science of Spain (Formación del Personal Investigador, Subprograma de Perfeccionamiento para Doctores y Tecnólogos). Animals used in this study were maintained in accordance with guidelines of the Committee on Care and Use of Laboratory Animal Resources, National Research Council (Department of Health, Education and Welfare, Publication No. (NIH) 85–23, revised 1985), and the McLean Hospital Institutional Animal Care and Use Committee.     

Orally delivered pentobarbital as a reinforcer for rhesus monkeys with concurrent access to water: Effects of concentration,fixed-ratio size,and liquid positions

The number of liquid deliveries and pattern of concurrent pentobarbital and water drinking were studied in three food deprived rhesus monkeys during daily 3-hr sessions. During the daily sessions, deliveries of approximately 0.6 ml of each liquid occured under fixed-ratio (FR) schedules of lip contact responses. Between sessions water was freely available. Session drinking was studied as a function of pentobarbital concentration (1.0, 1.41, 2.0, and 4.0 mg/ml) and FR size (4, 8, 16 and 32 lip contacts per delivery). The number of drug deliveries decreased with increases in drug concentration. Drug intake ranged from 21 to 52 mg/kg of body wt./3-hr session. At all concentrations and FR values tested, the number of pentobarbital deliveries substantially exceeded the number of water deliveries. The positive reinforcing effect of the pentobarbital was indicated by a consistent choice of drug over water irrespective of the side position of pentobarbital and by higher rates of drug responding. Pentobarbital drinking occured in a negatively accelerated pattern whereas water drinking did not have any consistent pattern. Marked intoxication followed bouts of drug drinking.