The effect of zolpidem on operant behavior and its relation to pharmacokinetics after intravenous and subcutaneous administration: concentration-effect relations.

We characterized the effects of i.v. and s.c. zolpidem (1-8 mg/kg) under a differential reinforcement of low-rate schedule (i.e. DRL 45 s) in 3-hour sessions. Both behavioral and pharmacokinetic-pharmacodynamic analyses were used with the intent to compare the effects of zolpidem with those of benzodiazepines reported previously under the same behavioral paradigm. Zolpidem increased the shorter-response [inter-response times (IRTs)<45 s] rate and decreased the reinforcement rate in a dose- and time-related fashion. The behavioral profiles of zolpidem were mainly similar to those of benzodiazepines, except zolpidem produced far fewer shorter IRT responses. Pharmacokinetically, zolpidem decays biexponentially with distributional and terminal elimination half-lives of 5.2 and 42 min, respectively. The absorption rate constant and absolute bioavailability for s.c. zolpidem were 0.083/min and of 84.1%, respectively. The pharmacodynamic parameters for the reinforcement rate, an index of timing performance, were determined by integration of behavioral and pharmacokinetic profiles in a between-subject design using the effect-linked inhibitory sigmoidal E(max) model. The pharmacokinetic-pharmacodynamic analysis revealed that the potency of zolpidem (concentration required to produce 50% maximal effects, IC(50)) in disrupting the timing performance was 0.129 microg/ml. The pharmacodynamic estimates of zolpidem were compared to our previous results for benzodiazepines.

Arteriovenous serum cocaine concentration difference after intravenous bolus injection and constant-rate infusions: relation to pharmacodynamic estimates in rats.

We characterized the pharmacokinetics of cocaine using both arterial and venous serum data after a bolus dose (2 mg/kg) and two constant-rate infusions (12.24 and 24.48 microg/min) for 2 h in rats. A published behavioral effect was used to investigate the effects of arteriovenous serum concentration differences on pharmacodynamic estimates for the 2 mg/kg dose. Significant temporal arteriovenous serum cocaine and benzoylecgonine (the major metabolite) concentration differences existed after cocaine administrations. The AUCs for arterial serum data were greater than the AUCs for venous data, indicating that cocaine was metabolized more extensively in the venous sampling site. Cocaine's behavioral effect could be directly related to serum concentrations with no hysteresis observed between the effects and arterial or venous serum concentrations. The pharmacodynamic estimates derived from arterial serum data approximated those from the venous data due to the most decline of cocaine's effect occurred in the elimination phase during which serum cocaine concentrations were not significantly different between the two sampling sites.

Simultaneous pharmacokinetic modeling of cocaine and its metabolites, norcocaine and benzoylecgonine, after intravenous and oral administration in rats.

To accurately assess the mechanism of involvement of the active metabolite norcocaine in the effects of oral cocaine, it is essential to determine the rate and extent of the formation of norcocaine. Although this study was designed specifically for this aim, it was also of interest to characterize the metabolite kinetics of benzoylecgonine for comparative purpose. We first characterized the pharmacokinetics of cocaine, norcocaine, and benzoylecgonine by the i.v. route of administration; all three drugs decayed biexponentially. These pharmacokinetic estimates were then used for determination of the formation of norcocaine and benzoylecgonine after i.v. and p.o. (20-40 mg/kg) cocaine administration. Although t(1/2alpha), and t(1/2beta) were similar across the three compounds, the values of volume of distribution in the central compartment and clearance for benzoylecgonine were much smaller than those of cocaine and norcocaine. Norcocaine was not detected following i.v. cocaine; however, serum norcocaine concentrations were as high as those of oral cocaine. Both routes of cocaine administration produced benzoylecgonine. A pharmacokinetic model for the metabolite kinetics was proposed by sequentially adding the models that most adequately described the formation of each metabolite to the model of cocaine. For oral cocaine, the absolute bioavailability was 3.48%, whereas 6.04 and 2.26% of cocaine were converted to benzoylecgonine and norcocaine, respectively, during first-pass absorption regardless of dose. Furthermore, the majority of norcocaine and 92% of benzoylecgonine were formed during the first-pass absorption, leaving 8% of benzoylecgonine produced in systemic circulation. The profile of norcocaine as a metabolite confirmed the involvement of norcocaine in cocaine's behavioral effects.

Negligible tolerance produced by chronic intravenous alprazolam administration: a low serum drug concentration effect.

In 3-hr sessions, the authors investigated the onset, peak, and disappearance of the effects of alprazolam on performance under a differential reinforcement of low rate 45-s schedule in rats. Alprazolam was administered chronically as a daily bolus dose (2 mg/kg) via the intravenous route. Alprazolam decreased the reinforcement rate and increased the shorter response (nonreinforced) rate in a dose- and time-related fashion. Tolerance did not develop to the decreases in reinforcement rate; tolerance to increases in shorter response rate was negligible, occurring only at the low-concentration range. Clinically, an optimal dose regimen should be designed to avoid the tolerance development that occurs in the low serum benzodiazepine concentration range.

Spontaneous activity as a contingency-controlled behavior within an operant context: alprazolam concentration-effect relations after subcutaneous administration in rats.

RATIONALE: Environmental factors affect serum drug concentration-effect relations. For example, after midazolam administration, longer pre-session delays imposed in experimental chambers produced differential concentration-effect relations compared to those of shorter delays. OBJECTIVES: To evaluate the extent to which serum concentrations determine alprazolam's effects on spontaneous activity in the presence and absence of a differential reinforcement of low rate (DRL 45-s) contingency using pharmacokinetic-pharmacodynamic analysis. Serum concentrations reported here were simulated from our published pharmacokinetic parameters for alprazolam. METHODS: One group (n=8) was used to investigate alprazolam's effects on spontaneous activity within the DRL contingency by placing an activity platform beneath each operant chamber to monitor concurrently both spontaneous activity (large and small movements) and DRL performance (shorter-response and reinforcement rates) in 3-h sessions; a parallel group (n=7) was used without the operant context. The concentration-effect relation of the reinforcement rate was compared and contrasted with those of spontaneous activity. RESULTS: Alprazolam decreased large and small movements within the DRL contingency, which corresponded to that of reinforcement rates under the DRL 45-s schedule. In contrast, without the DRL contingency, alprazolam's effects on small movements were short-lived (i.e., 30 min) and no effects on large movements were detected. Hence, the predicted concentration-effect relations for the reinforcement rate function described those of spontaneous activity well within the operant context, but not those without the operant context. Furthermore, the latter showed no correlation between serum alprazolam concentration and large movements; a significant, but low negative correlation for small movements was observed. CONCLUSIONS: The duration of alprazolam's action was dependent on not only dose size but also the behavioral measure examined. By imposing the DRL contingency, spontaneous activity behaves as an ideal pharmacodynamic measure (i.e., continuous, sensitive, and objective).

Two procedures establishing preference for oral cocaine and lidocaine solutions which do not use an associative history with a reinforcer.

Groups of rats were exposed to daily, 3-h, fixed-time 1-min food-pellet delivery sessions, a procedure that produces overdrinking (schedule-induced polydipsia, SIP). Previous research demonstrated that rats drinking a drug or non-drug solution come to prefer that solution to water if the solution had (a) a past association with either a highly acceptable vehicle (e.g. glucose/saccharin), or (b) allowed rats to eschew drinking an unacceptable solution under SIP conditions. The present experiments show that under the solution-avoidance procedure, preference for a concurrent, alternative solution (0.19 mg/ml lidocaine or 0.24 mg/ml cocaine) occurred when a concentrated quinine solution alternative was either abruptly removed or faded. A concentrated cocaine solution, however, was minimally effective in producing a preference for 0.19 mg/ml lidocaine to water when cocaine concentration was faded. Flavor/nutrient-conditioning (conditioned reinforcement) and solution-eschewing (avoidance) procedures may throw light on the kinds of historical situations that determine the genesis of stable preferences for drugs and other substances.

Contribution of the active metabolite, norcocaine, to cocaine's effects after intravenous and oral administration in rats: pharmacodynamics.

RATIONALE: Oral cocaine is more effective than IV cocaine by pharmacokinetic and pharmacodynamic analysis. One explanation is involvement of the active metabolite, norcocaine, in cocaine's effects. OBJECTIVES: To evaluate norcocaine's contribution to oral cocaine's effects, norcocaine's effects as a parent compound were determined and compared to those of cocaine using a differential reinforcement of low rate (DRL 45-s) schedule and spontaneous activity (large and small movements) after IV and PO routes of administration. METHODS: The effects of cocaine and norcocaine on DRL performance (shorter-response and reinforcement rates) and spontaneous activity were investigated in 3-h sessions. The changes in effects across time (effect-time profiles) and dose-response curves (DRCs) were constructed to evaluate the duration of action and potency (ED50) of both drugs. RESULTS: Under the DRL 45-s schedule, effect-time profiles showed both drugs via the two routes of administration significantly increasing and decreasing shorter-response rates and reinforcement rates, respectively. However, cocaine produced greater effects on shorterresponse rates than norcocaine, while both drugs produced comparable effects on reinforcement rates. For spontaneous activity, although IV cocaine, PO cocaine, and PO norcocaine dose- and time-dependently increased spontaneous activity, cocaine's effects were more profound than those of norcocaine. Effect-time profiles revealed that the duration of drug action was a function of dose, route, and behavioral paradigm used. According to ED50 values, IV cocaine was more effective than PO cocaine; however, PO cocaine was more effective than IV cocaine as judged by ED50 values corrected for absolute oral bioavailability. CONCLUSIONS: Norcocaine's contribution to oral cocaine's effects on DRL performance is evident. Other mechanism(s), such as a greater acute tolerance to IV cocaine's effects than to PO cocaine's effects, can be excluded.

Oral cocaine pharmacokinetics and pharmacodynamics in a cumulative-dose regimen: pharmacokinetic-pharmacodynamic modeling of concurrent operant and spontaneous behavior within an operant context.

Despite wide use of cumulative-dosing procedures to evaluate dose-response relations, limited attention has been paid to investigating drug concentration-effect relations. We first characterized the pharmacokinetic (PK) parameters for i.v. (2 mg/kg) and oral cocaine (20 and 40 mg/kg) in rats. Cocaine's concentration-time profile for the escalating cumulative-dose regimen was simulated from PK parameters, dose size (1, 2, 7, 20, and 45 mg/kg by the oral route), and dosing interval (tau, 35 min) as well as validated from blood sampling at various time points. This concentration-time profile was integrated with pharmacodynamic (PD) profiles of differential reinforcement of low rate performance and spontaneous activity (large and small movements) under a differential reinforcement of low rate 45-s schedule. Effects on three behavioral measures were characterized by integrated PK-PD models using the sigmoid E(max) (for increases in shorter response rate or large movements) and inhibitory E(max) (for decreases in density of reinforcement) models. But for the intrinsic differences in baseline and efficacy values among the behavioral endpoints, one set of PD parameters (i.e., potency and Hill factors) predicted concentration-effect relations for the three behavioral indices across all five doses. Concurrent monitoring of operant and spontaneous activity behavior within an operant context provides a novel behavioral paradigm to investigate drug effects on spontaneous activity under conditions where a behavioral contingency exists. Additionally, a cumulative-dosing procedure is efficient for determining the entire dose-response relation and provides an ideal mode to study phenomena such as sensitization or tolerance by varying dose size and/or tau.

High-performance liquid chromatographic determination of cocaine and its metabolites in serum microsamples with fluorimetric detection and its application to pharmacokinetics in rats.

A sensitive, selective and simple HPLC method with fluorimetric detection is described for quantitating cocaine and its three metabolites in rat serum microsamples (50 microl). Chromatographic separation is achieved on a Hypersil BDS C18 column (100X2.1 mm, 5 microm) with an isocratic mobile phase consisting of methanol-acetonitrile-25.8 mM sodium acetate buffer, pH 2.6, containing 1.0 x 10(-4) M tetrabutylammonium phosphate (14:10:76, v/v/v). The detection limit (0.5 ng/ml) for all the compounds, using direct fluorometric detection operated at excitation and emission wavelengths of 230 and 315 nm, respectively, was approximately five-times lower than that of using a UV detector operated at 235 nm. The effects of ratio of 2-propanol to chloroform in extraction solvents on the recovery and precision for cocaine and its metabolites were systematically examined. The method was used to study the pharmacokinetics of cocaine after administration of intravenous 2 mg/kg and oral 20 mg/kg doses.

Post-injection delays in experimental chambers, but not in home cages, produce both sensitization and tolerance of operant behaviour to midazolam: relation to pharmacokinetics.

The effects of post-injection delay time and environmental context on behaviour after subcutaneous administration of 3 mg/kg midazolam were investigated under a differential reinforcement of low rate schedule (i.e. DRL 45 s) in 3 h sessions. Post-injection delays were varied (0-120 min) for two groups of rats placed in either the experimental chamber (group 1) or home cage (group 2) during the pre-session delay times. Midazolam increased shorter-response (inter-response times < 45 s) rates and decreased reinforcement rates in a time-related manner. Reinforcement rate-time profiles were also integrated with parallel pharmacokinetics. Post-injection delays in either environment yielded performances that mirrored the pharmacokinetic profile operative at the corresponding time-delay points. At higher concentrations (> 0.12 microg/ml) the pharmacokinetics of midazolam largely determined the behavioural effects in both groups, regardless of post-injection delays. However, at lower drug concentrations, longer post-injection delays (> 60 min) in the experimental chamber produced both sensitization and tolerance, as measured by greater increases in shorter-response rates and a more rapid return of the reinforcement rate, respectively. Interaction of the discriminative stimulus effects of midazolam with the context probably alters the magnitude of behavioural effects when the delay occurs in the experimental chambers, whereas no such interaction is present in group 2. The DRL schedule with post-injection delays in experimental chambers provides a useful behavioural paradigm for studying both sensitization and tolerance.

Intravenous and oral clozapine pharmacokinetics, pharmacodynamics, and concentration-effect relations: acute tolerance.

We examined the pharmacokinetics and pharmacodynamics of intravenous (1-5 mg/kg) and oral clozapine (2.5-10 mg/kg) in rats (terminal half-life=81.8 min; oral bioavailability=5.32%). Both dose- and concentration-effect relations of clozapine were characterized. Clozapine's effects were similar to those of benzodiazepines because of the similarity in effect-time profiles between the two classes of drugs. The IC(50) value increased as a function of dose; consequently, clozapine's relative potency decreased linearly with the logarithm of AUC((0-infinity)), or bioavailable dose regardless of route of administration. The IC(50) is an index for the sensitivity of behavioral performance to clozapine; relative potency provides an index for estimating the extent of acute tolerance. As IC(50) increases, relative potency decreases, and consequently, acute tolerance increases. Our results demonstrated that greater acute tolerance was observed for i.v. clozapine than for p.o. clozapine; however, clozapine exhibited a single concentration-effect relation across dose and route of administration after correcting for relative potencies.

Determination of zolpidem in serum microsamples by high-performance liquid chromatography and its application to pharmacokinetics in rats.

A single-solvent extraction step high-performance liquid chromatographic method is described for quantitating zolpidem in rat serum microsamples (50 microl). The separation used a 2.1 mm I.D. reversed-phase OD-5-100 C18 column, 5 microm particle size with an isocratic mobile phase consisting of methanol-acetonitrile-26 mM sodium acetate buffer (adjusted to pH 2.0 with 40% phosphoric acid) containing 0.26 mM tetrabutylammonium phosphate (13:10:77, v/v/v). The detection limit was 3 ng/ml for zolpidem using an ultraviolet detector operated at 240 nm. The recovery was greater than 87% with analysis performed in 12 min. The method is simple, rapid, and applicable to pharmacokinetic studies of zolpidem after administering two intravenous bolus doses (1 and 4 mg/kg) in rats.

Pharmacokinetic determinants of cocaine's differential effects on locomotor and operant behavior.

Dose-response, effect-time and concentration-effect relations of intravenous cocaine (1-4 mg/kg) were investigated on contingency-controlled [fixed-ratio (FR) 70 performance] and unconditioned (locomotor activity) behaviors. Cocaine dose-response curves exhibited decreasing rates of response under the FR 70 schedule but increasing locomotor activity in a dose-related fashion. Effect-time profiles confirmed that these changes were time-dependent and provided additional clarity by mirroring the biexponential decay of cocaine concentrations with time. The duration of action of cocaine was comparatively shorter on locomotor activity than on FR performance. We integrated effect-time profiles of the two behaviors with concentration-time profiles simulated from our previously published pharmacokinetic parameters to derive cocaine's pharmacodynamic parameters. Classical inhibitory Emax and sigmoidal Emax models were used to describe cocaine's effects on FR performance and locomotor activity, respectively. Simultaneous pharmacokinetic-pharmacodynamic modeling reveals evidence of acute tolerance to cocaine in locomotor activity, as indicated by decreasing potency with dose, but not in contingency-controlled behavior.

Cocaine pharmacodynamics after intravenous and oral administration in rats: relation to pharmacokinetics.

RATIONALE: To design optimal dose regimes for oral cocaine, it is essential to characterize pharmacokinetics (PK) of cocaine after IV and PO administration. OBJECTIVES: To investigate the absolute bioavailability of oral cocaine, its effectiveness and the relation between PK and PD in a within-subject design. METHODS: We used the effects of IV and PO cocaine on a contingency-controlled timing behavior, the differential reinforcement of low rate schedule (DRL 45-s) in 3-h sessions, as the PD measures [i.e., the shorter-response rate (srr) and the reinforcement rate (rr)]. Cocaine PK parameters were determined by simultaneous modeling of the concentration-time profiles (CTPs) after IV 2 mg/kg and PO 20 mg/kg cocaine administration. The absolute oral cocaine bioavailability was determined pharmacokinetically (F) and pharmacodynamically (Fsrr and Frr). RESULTS: IV and PO cocaine increased the shorter response rate and decreased the reinforcement rate in a dose- and time-related fashion, which mirrored the respective prototypical serum cocaine CTPs. After the absorption phase, the serum cocaine CTP of PO cocaine paralleled that of IV cocaine. The duration of action for PO cocaine was longer than that for IV cocaine owing to its larger mean residence time. The active metabolite, norcocaine, was not detected after IV but after PO cocaine administration. The value of F was 4.66% which was significantly lower than the values of Fsrr (13.67%) and Frr (32.63%). Furthermore, the concentration-effect relations for the reinforcement rate revealed that PO cocaine was more potent than IV cocaine. CONCLUSIONS: Oral cocaine is more effective behaviorally than from predictions made in terms of its PK. The differences in active metabolite profiles as well as the rate and extent of acute tolerance for IV versus PO cocaine may account for the greater potency observed for oral cocaine.

A double-peak phenomenon in the pharmacokinetics of alprazolam after oral administration.

The pharmacokinetics of alprazolam (ALP) after i.v. and p.o. administration in rats were characterized. ALP decayed biexponentially after the i.v. dose (1.25 mg/kg), but the concentration-time profiles after the p.o. doses (7 and 12.5 mg/kg) exhibited a double-peak phenomenon. The presence of two peaks was confirmed by statistical analysis of the serum concentration data of ALP, as well as by observed double peaks in the serum concentration-time profiles of the two active metabolites (alpha-hydroxyalprazolam and 4-hydroxyalprazolam). An absorption model incorporating a delay site is proposed to describe the data, and the absolute oral bioavailability is estimated to be about 30%. The two peaks were approximately 80 to 115 min apart, and there was a delay in the absorption of close to 80% of oral ALP, regardless of dose. We hypothesize that the mechanism underlying the double-peak phenomenon is due to reduction in gastric motility caused by the muscle relaxant effect of ALP. This hypothesis is supported by the observed longer delay in the appearance of the second peak at the higher p.o. dose. Enterohepatic recycling is precluded from being the underlying mechanism, because of the presence of double peaks after the p.o. doses but not after the i.v. dose. This is the first reported case of double peaks for oral ALP, and this phenomenon has not been reported for other benzodiazepines. The double-peak phenomenon caused by the hypothesized mechanism may have important therapeutic and drug interaction implications, especially because benzodiazepines are commonly coadministered with other drugs.

Within-subject variability in cocaine pharmacokinetics and pharmacodynamics after intraperitoneal compared with intravenous cocaine administration.

Performance in rats (Rattus norvegicus) was measured on a differential reinforcement of low-rate schedule (DRL 45-s) in 1.5-hr sessions after 2 mg/kg intravenous (i.v.) or 10-20 mg/kg intraperitoneal (i.p.) cocaine administration, with each dose given twice and separated by 3-5 days. For successive i.v. doses, cocaine effects were similar, with minimal within-subject variability. For i.p. cocaine, the effects were not always similar; performance was variable and sometimes remained at baseline level. These diminished effects occurred following either the 1st or 2nd i.p. injection. A parallel pharmacokinetic study of cocaine confirmed that within-subject variability existed in cocaine concentration-time profiles after i.p. cocaine, and that a low serum cocaine concentration-time profile could account for the diminished effects. The i.p. route for cocaine administration should be used with caution.

Pharmacokinetic-pharmacodynamic modeling of the psychomotor stimulant effect of cocaine after intravenous administration: timing performance deficits.

We investigated dose-response cocaine pharmacokinetic and metabolite profiles in a within-subject design after intravenous bolus cocaine administration (1-4 mg/kg) in rats under a food-limited regimen. Cocaine was rapidly distributed (T1/2beta = 1.09 min) and eliminated (T1/2alpha = 14.93 min). Norcocaine was not detected. The free fraction of cocaine was 31.3-33.1% for serum cocaine concentrations of 0.5 to 1 microg/ml. Parallel pharmacodynamics was studied using performance on a contingency-controlled timing behavior, a differential reinforcement of low rate schedule (45 s) in 3-h sessions. Cocaine increased the shorter-response rate and decreased the density of reinforcement in a dose- and time-related fashion. The increased shorter-response rate is the stimulatory effect herein reported. The changes in shorter-response rate and the density of reinforcement were directly interpretable as functions of cocaine concentrations in the respective hypothetical effect compartments by using sigmoidal Emax and inhibitory Emax models, respectively. Because the concentration at half of Emax for the shorter-response rate (EC50 = 0.467 microg/ml) was greater than that for density of reinforcement (IC50 = 0.070 microg/ml), the former began to return toward baseline sooner than the latter. Only as cocaine concentration decreased to values smaller than the EC50 did the density of reinforcement begin to return toward baseline. Thus, the density of reinforcement is an index for evaluating the deficit in timing performance. The concentration-effect plot confirmed that the intensity of the effects of cocaine depends solely on concentration regardless of the dose. These results demonstrated that the pharmacokinetic-pharmacodynamic analysis allows the identification of the stimulant action of cocaine, which in turn delineates its consequence on timing performance.

Dose independent pharmacokinetics of caffeine after intravenous administration under a chronic food-limited regimen.

Several studies have shown that caffeine follows non-linear pharmacokinetics in both rats and humans. Recent data have demonstrated that caffeine may following linear pharmacokinetics when administered orally and intraperitoneally to food-limited rats. In this study the pharmacokinetics of caffeine was analyzed following intravenous (i.v.) administration to rats under a food-limited regimen. Four rats were administered four doses of caffeine and a standard dose of the caffeine metabolites, paraxanthine, theobromine, and theophylline. Caffeine pharmacokinetic parameters were dose independent following intravenous doses ranging from 1 to 20 mg/kg. Furthermore, the caffeine area under the curve (AUC) increased linearly as a function of dose. The mean fraction of caffeine converted to paraxanthine, theobromine, and theophylline was 16%, 16%, and 7%, respectively. The linear pharmacokinetics demonstrated in the present study may be attributed to the induction of hepatic metabolism under a chronic food-limited regimen.

Establishing preference for lidocaine solution to water: comparison between a fading and an abrupt-removal procedure for withdrawing a compound vehicle.

Rats were exposed to daily, 3-h, fixed-time 1-min food-pellet delivery sessions, which is a procedure that produces overdrinking (schedule-induced polydipsia). In previous polydipsia studies, rats came to prefer solutions of drug or non-drug agents to concurrently presented water if the agents had first been offered in a glucose-saccharin vehicle that was slowly eliminated (faded), leaving a choice between a substance in water vehicle vs water. In the first experiment, a more rapid vehicle-fading procedure was used to produce a preference for 0.19mg/ml lidocaine to water. In the second experiment, the vehicle was abruptly changed to water, which also resulted in a strong preference for lidocaine solution, although the lidocaine solution volumes ingested for the final sessions were significantly less than in the first experiment. The results are consistent with a conditioned flavor/nutrient preference interpretation for the institution of the lidocaine preference in both experiments. Although flavor/nutrient conditioning can be a sufficient condition for generating a substance preference, a previous experiment showed that it was not a necessary condition.

Establishing oral preference for quinine, phencyclidine and caffeine solutions in rats.

Rats were given fixed-time, 1-min food-pellet delivery sessions, for 3 h every day, which resulted in over drinking (schedule-induced polydipsia). In previous research, groups of animals came to prefer solutions of cocaine or lidocaine to concurrently presented water, if the drugs were first offered in a glucose/saccharin vehicle, which was then gradually eliminated, so that the choice became a drug solution in water vehicle versus water. In the present experiment, the same procedure produced a preference for 0.025 mg/ml quinine solution to water, indicating that a bitter solution that was not a topical anesthetic agent could come to be preferred. After establishing preference for quinine solution, it was possible to gradually remove quinine while increasing phencyclidine (PCP) concentration to produce preference for 0.075 mg/ml PCP solution, and similarly to produce subsequent preference for 0.1 mg/ml caffeine solution, although not for all animals. For additional groups, drinking either 0.15 mg/ml PCP or 0.1 mg/ml caffeine, while gradually reducing the glucose/saccharin vehicle to water vehicle was a less successful procedure for establishing the respective drug preferences. The latter result suggests that first instituting a preference for quinine solution to water facilitated the later establishment of preferences for PCP and caffeine solutions.