1. The results of previous studies have been in conflict with respect to the involvement of specific cholecystokinin (CCKA) and CCKB/gastrin receptors in guinea-pig gastric muscle. Here, in an in vitro, guinea-pig gastric muscle assay, pentagastrin (PG) and tetragastrin (TG) behaved as high potency agonists and produced symmetrical concentration-effect curves. In contrast, cholecystokinin-octapeptide (CCK-8), while also behaving as a high potency agonist, produced flat asymmetrical curves. Unlike recent data reported using this tissue (Boyle et al., 1993), the CCKA receptor-selective antagonist, devazepide (3, 10, 30 nM) produced a rightward shift of the upper region of the CCK-8 curve rendering it biphasic. The lower phase was abolished by the CCKB/gastrin receptor-selective antagonist, L-365260 (300 nM) indicating that the contractile effects of CCK-8 in this tissue are mediated by both receptor types. 2. L-365260 produced a concentration-dependent, parallel rightward displacement of PG concentration-effect curves. However, a flat Schild plot slope parameter (0.77 +/- 0.06) was obtained. Therefore, an empirical pA2 value of 8.64 +/- 0.21 was estimated from the smallest dose ratio. This value is consistent with published values characteristic of an interaction at CCKB/gastrin receptors. 3. TG (1 microM) was used to densensitize selectively the CCKB/gastrin receptors in the gastric muscle assay and thereby expose a population of receptors capable of responding to subsequent stimulation by CCK-8 but not by PG. The selectivity of TG for CCKB/gastrin- over CCKA receptors was demonstrated by its low efficacy compared to CCK-8 in the guinea-pig gallbladder assay, a tissue shown previously to contain a homogeneous population of CCKA receptors. In TG-desensitized gastric muscle, CCK-8 concentration-effect curves were symmetrical and could be displaced in a simple parallel fashion by devazepide at nanomolar concentrations consistent with an interaction at CCKA receptors (pKB approximately 10). 4. These results indicate that the guinea-pig gastric muscle contains both CCKA- and CCKB/gastrin receptors and the effects of CCK-8 are mediated via both of these receptors. Notwithstanding the complexity of the behaviour of L-365260, it was possible to obtain a reasonable description of the system using a simple 2-receptor model in which the effects of individual receptor activation were assumed to be additive. The absence of a simple competitive interaction of PG with L-365260 may indicate, for example, non-homogeneity of CCKB/gastrin receptors or lack of concentration equilibrium between the bath and the receptor biophase.