Propolis, or bee glue, which contains a complex mixture of secondary metabolites, has long been used in many countries for the management of several diseases. The purpose of this study was to evaluate, by means of several pharmacological models, the anti-hyperalgesic effect of propolis collected in the south of Brazil. The abdominal constrictions induced in mice by intraperitoneal injection of acetic acid (0.6%), kaolin (50 mg kg-1) or zymosan (40 mg kg-1) were inhibited to different extents by an extract of propolis (1-60 mg kg-1) administered intraperitoneally 30 min earlier; mean ID50 (concentrations resulting in 50% inhibition) values were 2.7, 10.8 and 10.7 mg kg-1, respectively, and maximum inhibition was 58 +/- 5, 57 +/- 10 and 51 +/- 5%, respectively. Given orally (25-200 mg kg-1, 1h previously) propolis also inhibited the abdominal constrictions induced by acetic acid (maximum inhibition 43 +/- 5%). When injected intraperitoneally (3-60 mg kg-1, 30 min previously), propolis attenuated both the neurogenic (first phase) and inflammatory (second phase) pain responses and paw oedema caused by intraplantar injection of formalin (2.5%); maximum inhibition was 32 +/- 5, 43 +/- 6 and 19 +/- 2%, respectively. Oral administration of propolis (25-200 mg kg-1, 1 h previously) inhibited both phases and reduced the oedema formation associated with the second phase of the formalin test (maximum inhibition 22 +/- 5, 33 +/- 6 and 26 +/- 3%) and extract of propolis (3-30 mg kg-1 i.p. or 25-100 mg kg-1 p.o., respectively 30 min and 1 h previously) significantly inhibited capsaicin-induced pain with maximum inhibition of 39 +/- 8 and 41 +/- 8%, respectively. When assessed in the Randall-Sellito test of pain, the extract of propolis (3-30 mg kg-1, i.p., 30 min previously) significantly reversed the hyperalgesia induced by intraplantar injection of bradykinin (3 nmol per paw) in rats (P < 0.01). In contrast with morphine the extract of propolis (< or = 100 mg kg-1, 30 min previously) was ineffective when assessed in the tail-flick and hot-plate thermal assays. Naloxone (5 mg kg-1 i.p.) reversed (P < 0.01) the effect of morphine (5 mg kg-1 s.c.) by 70 and 94% respectively in the first and second phases of the formalin test, but did not interfere with the analgesic effect of propolis (10 mg kg-1 i.p., 30 min previously). These results show that ethanolic extract of propolis, given systemically, has significant anti-hyperalgesic action when assessed in chemical, but not thermal, models of nociception in mice and rats. Its analgesic action seems to be unrelated to release or activation of the opioid system.