1 The e ects of adenosine, the adenosine analogue, 2-chloroadenosine (2-CADO), the speci®c adenosine A 1 receptor agonist, N 6 -cyclopentyladenosine (CPA) and A 2 receptor agonist 5'-(Ncyclopropyl) carboxamidoadenosine (CPCA), were examined against seizures induced by acute administration of pentylenetetrazole (PTZ), 60 mg kg 71 , and PTZ kindled seizures, in rats. 2 Adenosine 1000 mg kg 71 , i.p., 5 min pretreatment and CPA 10 mg kg 71 i.p., 60 min pretreatment, showed signi®cant protection against acute PTZ-induced seizures while, CPCA up to 10 mg kg 71 was ine ective. The adenosine analogue 2-CADO in a dose of 5 mg kg 71 was only partially protective and on increasing the dose to 10 mg kg 71 , this protection was lost. 3 Theophylline, a non speci®c adenosine receptor antagonist at 50 mg kg 71 and the speci®c adenosine A 1 receptor antagonist, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX), at 1 mg kg 71 , if administered before the maximally protective doses of adenosine and CPA, completely reversed the protection a orded by them against PTZ seizures. While, pretreatment with the adenosine A 2 receptor antagonist, 3,7-dimethyl-1-propargylxanthine (DMPX), failed to reverse the protection. 4 Adenosine and the adenosine A 1 receptor agonist in doses that protected against seizures after acute PTZ administration, o ered only incomplete protection when tested against PTZ kindled seizures.5 The e ects of adenosine and adenosine receptor agonists on mean arterial pressure, heart rate and rectal temperature were studied, to rule out the possibility of their systemic e ects mediating the protection of PTZ seizures. All these agents produced a fall in mean arterial pressure, heart rate and hypothermia in the doses exhibiting an anticonvulsant response. While the e ect on blood pressure and heart rate was immediate i.e. seen within 5 min and, maintained throughout the observation period, the development of hypothermia lagged behind the onset of hypotension and bradycardia. However, there was no correlation between haemodynamic and hypothermic response and the anticonvulsant e ect. 6 The results indicate that the adenosine mediated anticonvulsant e ect is via stimulation of A 1 receptors. Hypotension and hypothermia do not appear to contribute to the protection observed with adenosine and the adenosine A 1 receptor agonists.