Several lines of evidence indicate that the antinociception produced by intrathecal administration of the alpha(2)-adrenoceptor agonists dexmedetomidine or ST-91 is mediated by different subtypes of the alpha(2)-adrenoceptor. We recently provided additional pharmacologic evidence for this idea, as well as for differences in the function of these receptors between Harlan and Sasco rats, two widely-used outbred substrains of Sprague-Dawley rat. The present study used isobolographic analysis to further characterize the receptors at which intrathecally administered ST-91 and dexmedetomidine act in these two substrains. The rationale for these studies derives from the assumption that if dexmedetomidine and ST-91 act as agonists at the same receptor then they should interact in an additive manner. However, if they interact in a supra-additive manner, then they must act at different subtypes of the alpha(2)-adrenoceptor. In the tail-flick test, the dose-effect relationship for a 1:3 mixture of dexmedetomidine and ST-91 was shifted significantly to the left of the theoretical dose-additive line in both Harlan and Sasco Sprague-Dawley rats. A similar finding was made in the hot-plate test despite the fact that the dose-response characteristics of the agonists were different in this test. Thus, in Harlan rats, in which ST-91 is a full agonist and dexmedetomidine is essentially inactive, the dose-effect relationship for the mixture of dexmedetomidine and ST-91 was shifted far to the left of the dose-additive line. Similarly, in Sasco rats, in which ST-91 is a partial agonist and dexmedetomidine is inactive, co-administration of the two agonists also shifted the dose-response relationship to the left of the dose-additive line. The consistent finding that these two alpha(2)-adrenoceptor agonists interact in a supra-additive manner provides strong evidence that dexmedetomidine and ST-91 produce antinociception by acting at different alpha(2)-adrenoceptor subtypes in the spinal cord. This conclusion is consistent with the earlier proposal that dexmedetomidine acts predominantly at alpha(2A)-adrenoceptors whereas ST-91 acts predominantly at non-alpha(2A)-adrenoceptors. Recent anatomical evidence indicates that these non-alpha(2A) adrenoceptors may be of the alpha(2C) type. The synergistic combination of an alpha(2A)- and an alpha(2C)-adrenoceptor agonist may provide a unique and highly effective drug combination for the treatment of pain without the sedation produced by an equianalgesic dose of a single alpha(2)-adrenoceptor agonist.