The effects of nicotine on ethanol-induced conditioned taste aversions in Long–Evans rats
- First Online:
- Cite this article as:
- Rinker, J.A., Busse, G.D., Roma, P.G. et al. Psychopharmacology (2008) 197: 409. doi:10.1007/s00213-007-1050-2
- 86 Views
Overall drug acceptability is thought to be a function of the balance between its rewarding and aversive effects, the latter of which is reportedly affected by polydrug use.
Given that nicotine and alcohol are commonly co-used, the present experiments sought to assess nicotine’s impact on ethanol’s aversive effects within a conditioned taste aversion design.
Materials and methods
Experiment 1 examined various doses of nicotine (0, 0.4, 0.8, 1.2 mg/kg) to determine a behaviorally active dose, and experiment 2 examined various doses of ethanol (0, 0.5, 1.0, 1.5 g/kg) to determine a dose that produced intermediate aversions. Experiment 3 then examined the aversive effects of nicotine (0.8 mg/kg) and ethanol (1.0 g/kg) alone and in combination. Additionally, nicotine’s effects on blood alcohol concentrations (BAC) and ethanol-induced hypothermia were examined.
Nicotine and ethanol combined produced aversions significantly greater than those produced by either drug alone or the summed aversive effects of the individual compounds. These effects were unrelated to changes in BAC, but nicotine and ethanol combined produced a prolonged hypothermic effect which may contribute to the increased aversions induced by the combination.
These data demonstrate that nicotine may interact with ethanol, increasing ethanol’s aversive effects. Although the rewarding effects of concurrently administered nicotine and ethanol were not assessed, these data do indicate that the reported high incidence of nicotine and ethanol co-use is unlikely due to reductions in the aversiveness of ethanol with concurrently administered nicotine. It is more likely attributable to nicotine-related changes in ethanol’s rewarding effects.
KeywordsNicotineEthanolDrug interactionsPolydrug useDrug acceptabilityConditioned taste aversionHypothermiaBlood alcoholRatsFemale
Polydrug use is becoming increasingly more common, with nicotine and alcohol being two of the most commonly co-used psychoactive drugs (Barrett et al. 2006; Batel et al. 1995). Upwards of 80 to 90% of alcoholics reportedly smoke cigarettes (Perkins 1997; Romberger and Grant 2004). Further, according to a 1998 report by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), the heaviest drinkers are the heaviest consumers of nicotine, with more than 70% of alcoholics considered heavy smokers, compared to 10% of the general population (NIAAA 1998). Additionally, it is reported that smokers are at least 50% more likely to drink on a regular basis than non-smokers (Kozlowski et al. 1990). This consistent pattern of increased use of one in the presence of the other suggests that nicotine and alcohol may interact in ways that impact the affective properties of each drug, and in so doing, their use.
This epidemiological evidence has been supported by experimental work in both humans and animals that also suggests an interaction between nicotine and alcohol. For example, Kouri et al. (2004) reported that men pretreated with a 21-mg nicotine patch before ingesting alcoholic beverages had markedly higher subjective reports of feeling drunk, feeling the effects of alcohol and wanting to drink more than subjects given alcohol alone. Additionally, Perkins et al. (2000) found that men who smoked cigarettes after recently consuming alcohol significantly increased responding for more alcohol compared to those who did not smoke. They also found that those in the smoking group indicated greater subjective stimulatory and reduced sedative effects of alcohol compared to controls (Perkins et al. 2000; see also Rose et al. 2004).
Animal models of drug use have also revealed results in support of the human experimental literature. Using the self-administration procedure, Clark et al. (2001) demonstrated that rats delivered nicotine by micro-infusion increased the rate of acquisition and overall responding for ethanol (delivered in sucrose solution) compared to controls. This suggests that chronic nicotine infusion, which Clark et al. designed to mimic the constant blood nicotine levels maintained by smokers, may increase the rewarding effects of alcohol. Additionally, Korkosz et al. (2006b) examined the rewarding effects of nicotine and ethanol in combination using the conditioned place preference design. Combining doses that consistently condition place preferences (see Grabus et al. 2006; Risinger and Oakes 1995, 1996), Korkosz and colleagues reported that mice given an intraperitoneal (IP) injection of ethanol (1 g/kg) 5 min before receiving a subcutaneous (SC) injection of nicotine (0.3 mg/kg) showed a much stronger preference for the drug-paired side than mice given ethanol alone, but only a slightly stronger (non-significant) preference compared to those that received nicotine only (Korkosz et al. 2006b).
Although examining the rewarding effects of abused drugs gives insight into some of its motivational characteristics, many drugs that reliably produce place preferences or are readily self-administered also have aversive effects (Hunt and Amit 1987; Simpson and Riley 2005; van der Kooy et al. 1983) as indexed by the conditioned taste aversion (CTA) procedure (Garcia and Ervin 1968; Revusky and Garcia 1970; Rozin and Kalat 1971). In this procedure, an animal is given a novel tasting solution and injected with one of a number of compounds. Avoidance of that solution on subsequent exposures is thought to reflect the aversive effects induced by the compound (Riley et al. 1976; White et al. 1977; Wise et al. 1976; although for an alternative interpretation, see Grigson 1997). Although initially used to examine the aversive effects of radiation (Garcia et al. 1955; also see Freeman and Riley 2007), a wide variety of drugs have been reported to condition taste aversions, including ethanol and nicotine (see Riley and Tuck 1985). Ethanol’s effects within the taste aversion procedure are generally robust and produce dose-dependent aversions (Broadbent et al. 2002; Cailhol and Mormede 2002; Chester et al. 1998; Escarabajal et al. 2003; Quertemont 2003; Risinger and Cunningham 1998; Roma et al. 2006), while aversions induced by nicotine are generally weaker and often non-dose-dependent (Etscorn et al. 1986, 1987; Iwamoto and Williamson 1984; Kunin et al. 2001; Mucha 1997; Pescatore et al. 2005; Shoaib et al. 2000).
The balance between the rewarding and aversive effects of a drug is thought to influence its overall acceptability, and thus, its abuse potential (Lynch and Carroll 2001; Shram et al. 2006), which may be impacted by a variety of factors including species, sex, strain, drug history, and other concurrently or serially presented drugs (see Riley and Freeman 2004). Interestingly, drug combinations often weaken or potentiate the aversions induced by the individual drugs alone. For example, Braveman (1975) reported that pre-exposure to methylscopolamine attenuated LiCl- and amphetamine-induced taste aversions. Conversely, Etkind et al. (1998) found that aversions induced by cocaine (SC) and ethanol (IP) administered concurrently were significantly greater than those induced by either drug alone, as well as the summation of the aversions produced by each drug, implicating an interaction between the two substances (see also Grakalic and Riley 2002; Jones et al. 2006).
Given that conditioned taste aversions have been effective in assessing the aversive effects of drug combinations, the present experiments used the CTA design to examine the effects of nicotine on ethanol-induced taste aversions. Specifically, experiments 1 and 2 examined the relative efficacy of injections of nicotine (0, 0.4, 0.8, and 1.2 mg/kg) or ethanol (0, 0.5, 1.0, and 1.5 g/kg) in inducing aversions to saccharin. Based on these results, experiment 3 examined the effects of nicotine (0.8 mg/kg) and ethanol (1.0 g/kg), alone and in combination, in the CTA design. Specifically, experiment 3 sought to determine whether or not concurrently administered nicotine impacts ethanol-induced CTAs based on the possibility that the reported nicotine-related increases in the subjective positive effects of ethanol and increased consumption of ethanol with concurrently administered nicotine (in both human and animal literature) might be due to an attenuation in ethanol’s aversive effects.
Materials and methods
Subjects and housing
A total of 165 female (n = 39 to 42 per experiment), experimentally naïve Long–Evans rats approximately 90 days old and weighing between 200 and 300 g at the start of the experiments were used. Procedures recommended by the Guide for the Care and Use of Laboratory Animals (1996), the Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research (2003), and the Institutional Animal Care and Use Committee at American University were followed at all times. All subjects were individually housed in hanging wire-mesh cages and maintained on a 12:12 light–dark cycle (lights on at 0800 hours) and at an ambient temperature of approximately 23°C. Except where noted, food and water were available ad libitum. Animals were handled daily approximately 2 weeks before the initiation of the study to limit the effects of handling stress during conditioning and testing. Estrous cycles were not actively monitored; however, all subjects from each experiment were housed within the same room to promote estrous synchrony through olfactory and pheromonal cues (McClintock 1978; but see Schank 2001). All experimental procedures occurred between 1400 and 1800 hours.
Drugs and solutions
(−)-Nicotine hydrogen tartrate (Sigma Aldrich, St. Louis, MO, USA) was prepared as a 0.5 mg/ml solution dissolved in 0.9% saline. All doses of nicotine are expressed as the salt. A 15% (v / v) ethanol solution was prepared from a 95% ethanol stock (Sigma Aldrich) diluted in distilled water (dH2O). All drugs were administered intraperitoneally. Saccharin (sodium saccharin, Sigma) was prepared as a 1 g/l (0.1%) solution in tap water.
General CTA procedure
- Phase 1:
Habituation. After 24-h water deprivation, subjects were given 20-min access to tap water daily between 1500 and 1600 hours. This procedure was repeated until consumption stabilized, i.e., water consumption was within 2 ml of the previous day for a minimum of four consecutive days. Throughout the study, fluid was presented in graduated 50-ml Nalgene tubes and measured by subtracting the difference between the pre- and post-consumption volumes.
- Phase 2:
Conditioning. For each of the three experiments, all subjects were given 20-min access to the novel saccharin solution on conditioning trial 1. Immediately after this initial presentation, animals were rank-ordered on saccharin consumption and assigned to a treatment group (either vehicle, single drug, or drug combination) such that overall consumption was comparable among groups. Subjects received an IP injection(s) of either vehicle or drug(s) approximately 20 min after access to saccharin. The 3 days after this initial saccharin presentation were water-recovery days, during which animals were given 20-min access to tap water (no injections followed this access). This alternating procedure of conditioning and water recovery was repeated for a total of four complete cycles.
- Phase 3:
Aversion test. On the day after the last water-recovery session of the fourth cycle, all subjects were given access to the saccharin solution for 20 min in the aversion test. No injections were given on this day, except where noted in experiment 3 (see below).
Experiment 1: nicotine CTA
After saccharin consumption, rats (n = 42) were injected with one of four doses of nicotine (0, 0.4, 0.8, or 1.2 mg/kg), yielding the following groups: vehicle (saline, n = 10), nicotine 0.4 mg/kg (n = 10), nicotine 0.8 mg/kg (n = 11), and nicotine 1.2 mg/kg (n = 11).
Experiment 2: ethanol CTA
After saccharin consumption, subjects (n = 39) were injected with one of four doses of ethanol (0.0, 0.5, 1.0, or 1.5 g/kg), yielding the following groups: vehicle (dH2O, n = 9), ethanol 0.5 g/kg (n = 11), ethanol 1.0 g/kg (n = 10), and ethanol 1.5 g/kg (n = 9).
Experiment 3: nicotine–ethanol interaction
Based on the results from experiments 1 and 2, after saccharin consumption, animals (n = 42) were injected with either saline followed by dH2O (group vehicle–vehicle, n = 10), nicotine (0.8 mg/kg) followed by dH2O (nicotine–vehicle, n = 10), saline followed by ethanol (1.0 g/kg) (vehicle–ethanol, n = 11) or nicotine (0.8 mg/kg) followed by ethanol (1.0 g/kg; nicotine–ethanol, n = 11). Vehicle injection volumes were matched to either nicotine or ethanol, depending upon the group. Conditioning in experiment 3 was similar to that described above with the exception that on the fifth conditioning day, all subjects received their scheduled injections (as opposed to receiving a final aversion test). Animals were then given one water-recovery day followed on the next day by a two-bottle aversion test where animals were presented with one bottle containing saccharin and one containing tap water. Saccharin was presented first, followed 15 s later by tap water (bottle placement was counterbalanced so that the saccharin bottle was evenly placed on the right and left sides of the cages within each group). After 20-min access to the two bottles, the relative preference for saccharin was determined by dividing the amount of saccharin consumed by overall fluid intake and multiplying by 100.
In addition to assessing the impact of nicotine on ethanol’s aversive effects, experiment 3 examined the effect of nicotine on blood alcohol concentrations (BAC) after an acute ethanol challenge (Parnell et al. 2006) and on ethanol-induced hypothermia to assess its possible contribution to aversion learning (Cunningham et al. 1988).
Blood alcohol concentration
Animals from experiment 3 were maintained on ad libitum food and water for a minimum of 2 weeks before the BAC assessment. Animals were given two injections based on their group assignment from experiment 3 (vehicle–vehicle, nicotine–vehicle, vehicle–ethanol, or nicotine–ethanol). Tail blood samples were collected at 15, 60, and 180 min post-injection. Before the 15-min sampling, each animal’s tail was soaked in warm water for 60 s and wiped dry. The rat was then held in an oversized restraint tube (Plas-Labs, Lansing, MI, USA), and surgical scissors were used to trim 1 mm off the tip of the tail. For subsequent samplings, the tail was resoaked and dried, but no further incisions were made. Approximately 40–90 μl of whole blood were collected in heparinized capillary tubes (Drummond Scientific, Broomall, PA, USA) and transferred to microcentrifuge vials. Blood samples were centrifuged at 3,000 rpm, and the separated plasma was then transferred via micropipette to new vials and frozen until analysis. Undiluted plasma was assayed using the HP 6890 Series headspace gas chromatography/mass spectrometry system (Hewlett-Packard, Palo Alto, CA, USA) based on protocols developed by the Laboratory of Clinical and Translational Studies at the National Institute on Alcohol Abuse and Alcoholism.
For this assessment, 42 naïve female, Long–Evans rats, the same age as those in the preceding experiments and were maintained on ad libitum food and water before the hypothermia assessment were used. Animals were randomly assigned to one of the four conditions from experiment 3 (vehicle–vehicle, nicotine–vehicle, vehicle–ethanol, or nicotine–ethanol) and were given two injections based on these assignments. Core body temperatures were assessed with a digital thermometer (Vicks Speed-Read model V911; Kaz USA, Southborough, MA, USA) immediately before injections and again at 15, 60, and 180 min post-injection. During temperature readings, animals were held while the lubricated probe of the thermometer was inserted approximately 3 cm into the rectum for roughly 10 s and the temperatures were manually recorded.
The differences in mean saccharin consumption were analyzed for each experiment using a repeated measures analysis of variance (ANOVA) with the between-subjects variable of group and the within-subjects variable of trial. Where appropriate, pairwise comparisons were made using Tukey–Kramer post hoc tests to examine differences in mean saccharin consumption on each trial and on the aversion test. Significance levels were set at α ≤ 0.05 for all analyses.
Experiment 1: nicotine CTA
Experiment 2: ethanol CTA
Experiment 3: nicotine–ethanol CTA
Percent saccharin consumption and relative decrease from controls
% Saccharin consumption [(Sacc/dH2O + Sacc) × 100]
% Decrease in saccharin consumption from controls
Mean core body temperature and relative decrease from controls
Mean core body temperature (°C) at 15 min
Decrease from controls
Mean core body temperature (°C) at 60 min
Decrease from controls
The present series of experiments assessed the effects of nicotine on ethanol-induced conditioned taste aversions. This assessment was based on the premise that overall acceptability of a drug (or drug combination) is impacted by the relative contributions of its aversive and rewarding effects. Thus, information regarding such affective properties may be important for understanding its potential for use and abuse. Following assessments in experiments 1 and 2 of the ability of nicotine and ethanol alone to induce taste aversions, experiment 3 combined the intermediate doses of the two drugs, both of which were capable of producing significant aversions (compared to controls), to assess the aversive effects of the combination (relative to ethanol alone). As reported, the combination of 0.8 mg/kg nicotine and 1 g/kg ethanol induced a CTA that was significantly greater than that produced by nicotine or ethanol alone. During the conduct of the present study, Korkosz et al. (2006a) reported a similar interaction between ethanol and nicotine in the taste aversion procedure. Specifically, they examined the effects of a low dose of ethanol (0.25 g/kg) on nicotine’s (0.3 mg/kg) ability to induce a taste aversion. Similar to the present data, animals given several pairings of saccharin and the ethanol/nicotine combination exhibited a greater aversion to saccharin in a one-bottle test than animals injected with either drug alone. However, when animals were given a two-bottle test, those receiving the combination of ethanol and nicotine did not have a significantly greater aversion to saccharin than those receiving nicotine alone.
From the present work, as well as that from Korkosz et al. (2006a), it is clear that the combination of nicotine and ethanol can produce aversions greater than either drug alone. Although each drug appears to be able to impact the aversiveness of the other, the basis for these effects is not known. One simple explanation for how the combination of nicotine and ethanol induced greater aversions than ethanol alone is that the aversions produced by the combination were merely the result of the summation of the aversions induced by the individual drugs alone. Under such a scenario, the effects of the combination would be greater than either effect alone, but not greater than their summed effects. Although such a comparison was not described in the Korkosz et al. (2006a) report, in the present study, the effects of the combination exceeded that of the simple summation of the effects of the individual drugs. Specifically, relative to the control subjects, animals injected with nicotine and ethanol alone displayed a 10.5 and 21.6% mean reduction in saccharin consumption, respectively. If the enhanced aversions seen in the two-bottle test were the result of summed effects of nicotine and ethanol, then one would expect to see, approximately, a 32.1% decrease in saccharin consumption relative to control subjects in the group administered both drugs concurrently. However, subjects injected with the combination displayed a 60.5% decrease in saccharin consumption relative to controls, suggesting that the aversion in this group was not simply the result of the summed effects of the component parts. Thus, the aversive effects of the combination of nicotine and ethanol presented in this study suggest that nicotine may be pharmacologically interacting with ethanol to impact its ability to condition aversions, although the nature of the putative interaction is not known.
One possible explanation for the interaction involves possible changes in blood alcohol levels as a function of the concurrent administration of nicotine. For example, nicotine may have impacted the bioavailability of ethanol by affecting its absorption, distribution, metabolism, or excretion (Leonard 1997). Thus, to assess whether concurrent administration of nicotine and ethanol could increase BAC and possibly account for the increased aversions, the BAC of animals given the drug combination was compared to those given ethanol or nicotine alone. As demonstrated, when nicotine and ethanol were administered IP, nicotine had no impact on plasma ethanol levels (replicating an effect recently reported by Parnell et al. 2006 when assessing the interaction of IP nicotine and ethanol on BAC). While this argues against the possibility of altered BAC being responsible for the increased aversion, it does not rule out possible changes in the volume distribution of either compound or the production of a unique metabolite (like that with cocaine and ethanol–cocaethylene) retaining its own affective properties. Furthermore, although nicotine did not impact BAC, this does not reflect its impact on brain ethanol levels which could contribute to any behavioral differences seen among groups. Although it is possible that ethanol-induced aversions are partially mediated by central mechanisms, it should be noted that much of the literature suggests that such aversions are mediated via peripheral mechanisms. For example, studies have shown that peripheral, but not central, administration of ethanol and its primary metabolite, acetaldehyde, produce CTAs (Amit et al. 1977; Brown et al. 1978). This suggests that any differences in brain level would unlikely be mediating the effect seen in the group given the drug combination. Additionally, Hisaoka and Levy (1985) showed that acute nicotine pretreatment actually reduced brain and cerebrospinal fluid concentrations of ethanol.
Instead of the abovementioned pharmacokinetic interaction, it is possible that nicotine is directly impacting the aversiveness of ethanol. Attempting to determine how nicotine may be affecting ethanol to produce stronger aversions is somewhat difficult given that the nature of aversion learning with ethanol is not known. A variety of explanations for ethanol-induced aversions have been suggested (see Baker and Cannon 1982; Cunningham et al. 1988; Elkins et al. 2000, 2003; Orr et al. 1993); however, there is no consensus as to their mediation. One possible explanation for ethanol-induced aversions presented by Cunningham and his colleagues involves ethanol’s hypothermic effects (see Cunningham et al. 1988, 1992). Specifically, Cunningham and colleagues reported a negative correlation between the strength of the aversion induced by ethanol and core body temperature, i.e., as temperature decreased, aversions increased. The present data showing that the combination of nicotine and ethanol produced greater decreases in core temperature than that produced by either nicotine or ethanol (or their simple summation) support the relationship between hypothermia and ethanol-induced conditioned taste aversion learning and argue that the potentiated aversions seen with the combination may be mediated in part by the prolongation of the core temperature decreases (from 15 to 60 min) in the group receiving the drug combination.
The basis for the examination of the effects of nicotine on ethanol-induced aversions stemmed from an interest in understanding the mechanisms underlying their combined use. The present assessment of the effects of concurrent administration of nicotine and ethanol on their aversive effects may provide insight into its contribution to the overall acceptability of the drug combination. There are several caveats regarding this initial interest and the results. First, although the focus in the present report has been on how nicotine may have impacted ethanol, it is certainly possible that the increased aversions seen with the combination reflect ethanol-induced changes in the aversiveness of nicotine (or its subsequent metabolites; see Korkosz et al. 2006a; Kunin et al. 1999). There is nothing in the present design (see “Experiment 3”) that precludes this alternative account. According to this position, ethanol may be affecting nicotine blood levels or impacting nicotine’s activity at cholinergic systems thought to mediate its aversive effects (Adir et al. 1980; Bienkowski et al. 1998). Until comparable assessments are made on ethanol-induced changes in nicotine blood levels or on systems mediating nicotine-induced aversions specifically, such an account remains possible.
Secondly, although the current manuscript assumes possible changes in the aversive effects of ethanol (or nicotine) when the two are given in combination, there are other interpretations of CTAs that assume that the suppression seen in the CTA design reflects the rewarding rather than the aversive effects of the drug (see Grigson 1997; Hunt and Amit 1987). For example, Grigson has argued that the reductions in saccharin consumption after its pairing with one of a number of drugs of abuse are a result of the animal avoiding the naturally rewarding saccharin solution in anticipation of the more rewarding drug that follows (i.e., anticipatory contrast; see Grigson 1997). The present finding of greater taste aversions induced by the drug combination is consistent with this interpretation given that the combination of ethanol and nicotine has been reported in other preparations to be more rewarding than either drug alone (Clark et al. 2001; Korkosz et al. 2006b). While consistent, however, it should be noted that the general theoretical position of anticipatory contrast as an explanation of taste aversion learning has been challenged in a number of contexts. For example, subcutaneous injections of cocaine produce stronger taste aversions than intraperitoneal injections, while the opposite is true for the induction of place preferences. If reward mediated aversions, one would expect stronger aversions under the conditions that produce stronger place preferences (see Mayer and Parker 1993). Further, Broadbent et al. (2002) have shown a clear dissociation between the rewarding and aversive effects of ethanol in that mice that had a higher preference for alcohol developed less of an aversion to ethanol, a dissociation also seen with the F344 and LEW rat strains (see Riley et al. 2007; Roma et al. 2006; Suzuki et al. 1988). Finally, work on the effects of drug exposure on the subsequent acquisition of taste aversions and place preferences have shown that such pre-exposure sensitizes reward (Bienkowski et al. 1996; Shoaib et al. 1994) yet attenuates aversions (Barker and Johns 1978; Iwamoto and Williamson 1984; for a similar analysis with morphine, see Simpson and Riley 2005), again an effect counter to the reward comparison hypothesis. Thus, it remains to be determined the degree to which changes in the rewarding effects of drugs impact or mediate any changes in their ability to induce aversions.
Finally, if drug vulnerability is seen as a balance of a drug’s rewarding and aversive effects, it might have been expected that nicotine would attenuate the aversive effects of ethanol, an effect that would likely contribute to the increase in its overall acceptability reported in the literature. However, as described, the combination was in fact more aversive than either nicotine or ethanol alone, and the aversion induced by the combination was greater than the summed aversive effects of the individual drugs. Naturally, consideration of how these data fit with the existing drug reward literature is called for, and it appears that the most parsimonious explanation is that independent of the increase seen in ethanol’s aversive effects with the co-administration of nicotine, nicotine may be simultaneously potentiating ethanol’s rewarding effects and perhaps to a greater extent than the aversive effects. Although this report did not directly assess the rewarding effects of concurrent nicotine and ethanol administration, these data do indicate that the previously reported high incidence of nicotine and ethanol co-use (Clark et al. 2001; NIAAA 1998) is unlikely a function of reductions in the aversiveness of ethanol with concurrent nicotine administration and more likely due to nicotine-related changes in ethanol’s rewarding effects. Such a conclusion, however, should be cautiously accepted until other potentially aversive phenotypes are measured and concurrent assessments of changes in the rewarding effects of the drug combination are made under conditions comparable to those used in the present study. Ultimately, such assessments may converge with the present data and other comparative literature to yield further insights on the biobehavioral bases of polydrug abuse.
We owe a great deal of thanks to Dr. Markus Heilig for generously providing access to the gas chromatography system and to Erick Singley for his expert technical assistance therein. This research was supported by a grant from the Mellon Foundation to A.L.R. and by intramural funds from the National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Public Health Service, US Department of Health and Human Services.