There has been an exponential growth in interest in purinoceptors since the potent effects of purines were first reported in 1929 and purinoceptors defined in 1978. A distinction between P1 (adenosine) and P2 (ATP/ADP) purinoceptors was recognized at that time and later, A1 and A2, as well as P2x and P2y subclasses of P1 and P2 purinoceptors were also defined. However, in recent years, many new subclasses have been claimed, particularly for the receptors to nucleotides, including P2t, P2z, P2u(n) and P2D, and there is some confusion now about how to incorporate additional discoveries concerning the responses of different tissues to purines. The studies beginning to appear defining the molecular structure of P2-purinoceptor subtypes are clearly going to be important in resolving this problem, as well as the introduction of new compounds that can discriminate pharmacologically between subtypes. Thus, in this review, on the basis of this new data and after a detailed analysis of the literature, we propose that: (1) P2X(ligand-gated) and P2Y(G-protein-coupled) purinoceptor families are established; (2) four subclasses of P2X-purinoceptor can be identified (P2X1-P2X4) to date; (3) the variously named P2-purinoceptors that are G-protein-coupled should be incorporated into numbered subclasses of the P2Y family. Thus: P2Y1 represents the recently cloned P2Y receptor (clone 803) from chick brain; P2Y2 represents the recently cloned P2u (or P2n) receptor from neuroblastoma, human epithelial and rat heart cells; P2Y3 represents the recently cloned P2Y receptor (clone 103) from chick brain that resembles the former P2t receptor; P2Y4-P2Y6 represent subclasses based on agonist potencies of newly synthesised analogues; P2Y7 represents the former P2D receptor for dinucleotides. This new framework for P2 purinoceptors would be fully consistent with what is emerging for the receptors to other major transmitters, such as acetylcholine, gamma-aminobutyric acid, glutamate and serotonin, where two main receptor families have been recognised, one mediating fast receptor responses directly linked to an ion channel, the other mediating slower responses through G-proteins. We fully expect discussion on the numbering of the different receptor subtypes within the P2X and P2Y families, but believe that this new way of defining receptors for nucleotides, based on agonist potency order, transduction mechanisms and molecular structure, will give a more ordered and logical approach to accommodating new findings. Moreover, based on the extensive literature analysis that led to this proposal, we suggest that the development of selective antagonists for the different P2-purinoceptor subtypes is now highly desirable, particularly for therapeutic purposes.