In Experiment 1, twelve Sprague-Dawley rats were trained in a two-choice food-reinforced drug discrimination task (10-min sessions) using the state-dependent interoceptive stimulus attributes of ethanol's (ETOH) delayed or rebound effects (EDE) versus "normal" basal homeostasis. Rats were injected with either 4 g/kg ETOH or equivalent volumes of saline (SAL) 18 hr before the sessions. Each rat was injected with an additional 1 ml/kg injection of SAL 15 min before the sessions. EDE training sessions were always followed by a "day off." SAL sessions were conducted between 36-96 hr after an EDE training session. Rats demonstrated > 90% discriminative accuracy. Test sessions showed a time-dependent, cyclic, return from the experimental "hangover" state to the "normal" state, by 48 hr. The acute (immediate) effects of ETOH and chlordiazepoxide (0.75 g/kg or 0.18 mg/kg, respectively; @15 min) did not cross-generalize with the "hangover" state. Both these low-dose ETOH and chlordiazepoxide pretreatments blocked the stimulus attributes of "hangover." All subjects responded on the EDE-appropriate lever at 5.6 mg/kg pentylenetetrazole and exhibited an increase in susceptibility to clonic seizures. In Experiment 2 blood alcohol concentration kinetics functions were quantified in three groups (n = 8/group) of age-matched cohorts to Experiment 1 subjects (2, 3, and 4 g/kg ETOH) using a head-space gas chromatographic technique. The training stimulus state associated with 4 g/kg, at 18 hr postinjection intervals, in Experiment 1, did not produce any chromatogram peaks for ETOH or any its active metabolite (acetaldehyde, acetone, nor methanol).(ABSTRACT TRUNCATED AT 250 WORDS)