1. Gabapentin (neurontin) is a novel antiepileptic agent that binds to the alpha 2 delta subunit of voltage-dependent calcium channels. The only other compound known to possess affinity for this recognition site is the (S)-(+)-enantiomer of 3-isobutylgaba. However, the corresponding (R)-(-)-enantiomer is 10 fold weaker. The present study evaluates the activity of gabapentin and the two enantiomers of 3-isobutylgaba in formalin and carrageenan-induced inflammatory pain models. 2. In the rat formalin test, S-(+)-3-isobutylgaba (1-100 mg kg-1) and gabapentin (10-300 mg kg-1) dose-dependently inhibited the late phase of the nociceptive response with respective minimum effective doses (MED) of 10 and 30 mg kg-1, s.c. This antihyperalgesic action of gabapentin was insensitive to naloxone (0.1-10.0 mg kg-1, s.c.). In contrast, the R-(-)-enantiomer of 3-isobutylgaba (1-100 mg kg-1) produced a modest inhibition of the late phase at the highest dose of 100 mg kg-1. However, none of the compounds showed any effect during the early phase of the response. 3. The s.c. administration of either S-(+)-3-isobutylgaba (1-30 mg kg-1) or gabapentin (10-100 mg kg-1), after the development of peak carrageenan-induced thermal hyperalgesia, dose-dependently antagonized the maintenance of this response with MED of 3 and 30 mg kg-1, respectively. Similar administration of the two compounds also blocked maintenance of carrageenan-induced mechanical hyperalgesia with MED of 3 and 10 mg kg-1, respectively. In contrast, R-(-)-3-isobutylgaba failed to show any effect in the two hyperalgesia models. 4. The intrathecal administration of gabapentin dose-dependently (1-100 micrograms/animal) blocked carrageenan-induced mechanical hyperalgesia. In contrast, administration of similar doses of gabapentin into the inflamed paw was ineffective at blocking this response. 5. Unlike morphine, the repeated administration of gabapentin (100 mg kg-1 at start and culminating to 400 mg kg-1) over 6 days did not lead to the induction of tolerance to its antihyperalgesic action in the formalin test. Furthermore, the morphine tolerance did not cross generalize to gabapentin. The s.c. administration of gabapentin (10-300 mg kg-1), R-(-) (3-100 mg kg-1) or S-(+)-3-isobutylgaba (3-100 mg kg-1) failed to inhibit gastrointestinal motility, as measured by the charcoal meal test in the rat. Moreover, the three compounds (1-100 mg kg-1, s.c.) did not generalize to the morphine discriminative stimulus. Gabapentin (30-300 mg kg-1) and S-(+)-isobutylgaba (1-100 mg kg-1) showed sedative/ataxic properties only at the highest dose tested in the rota-rod apparatus. 6. Gabapentin (30-300 mg kg-1, s.c.) failed to show an antinociceptive action in transient pain models. It is concluded that gabapentin represents a novel class of antihyperalgesic agents.