1. Recent studies have shown beneficial effects of an adenosine A(2A) receptor agonist in dt(sz) mutant hamsters, an animal model of paroxysmal dystonia, in which stress and consumption of coffee can precipitate dystonic attacks. This prompted us to examine the effects of adenosine receptor agonists and antagonists on severity of dystonia in dt(sz) hamsters in more detail. 2. The non-selective adenosine A(1)/A(2A) receptor antagonists, caffeine (10 - 20 mg kg(-1) i.p.) and theophylline (10 - 30 mg kg(-1) s.c.), worsened the dystonia in dt(sz) hamsters. 3. Aggravation of dystonia was also caused by the selective adenosine A(1)/A(2A) antagonist CGS 15943 (9-chloro2-2-furyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine) at a dose of 30 mg kg(-1) i.p. and by the adenosine A(1) antagonist DPCPX (8-cyclopentyl-1,3-dipropylxanthine; 20 - 30 mg kg(-1) i.p.), while the A(2) antagonist DMPX (3,7-dimethyl-1-propargylxanthine; 2 - 4 mg kg(-1) i.p.) and the highly selective A(2A) antagonist ZM 241385 (4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol; 2 - 5 mg kg(-1) i.p.) failed to exert any effects on dystonia. 4. In contrast to the antagonists, both the adenosine A(1) receptor agonist CPA (N(6)-cyclopentyladenosine; 0.1 - 1.0 mg kg(-1) i.p.) and the A(2A) agonist CGS 21680 (2p-(2carboxyethylphen-ethylamino-5'-N-ethylcarboxamindoadenosine; 0.1 - 2.0 mg kg(-1) i.p.) exerted a striking improvement of dystonia. 5. These data suggest that the precipitating effects of methylxanthines are, at least in part, related to their adenosine receptor antagonistic action. 6. Although adenosine receptor agonists can be regarded as interesting candidates for the therapy of paroxysmal dystonia, adverse effects may limit the therapeutic potential of adenosine A(1) agonists, while beneficial effects of the adenosine A(2A) agonist CGS 21680 were already found at well tolerated doses.