Experiments were carried out in vivo and in vitro to study the flux of Ca2+ across the sheep rumen wall. Buffer solutions with five different Ca2+ concentrations in the range 0.2-3.6 mmol l-1 were introduced into the isolated, emptied and washed reticulo-rumen for 1 h to determine Ca2+ net flux. Phosphate concentrations were either kept constant at 8 mmol l-1, or varied so as to maintain a constant phosphate/Ca2+ ratio of 2:1. Under both conditions there was a Ca2+ net secretion into the rumen at Ca2+ concentrations below approximately 1 mmol l-1 whereas Ca2+ net absorption occurred at higher Ca2+ concentrations and linearly increased with increasing Ca2+ concentrations. When phosphate concentrations were reduced to 4.3 and 0.6 mmol l-1 (Ca2+, 3 mmol l-1) Ca2+ net absorption was reduced by about 50% and completely abolished, respectively. The unidirectional (mucosal-to-serosal or serosal-to-mucosal) Ca fluxes in vitro, JCams and JCasm, were measured in Ussing chambers with computer-controlled voltage clamp. Under short-circuit current conditions (Ca2+ and phosphate, 2 mmol l-1) there was a significant net flux of Ca2+ from the mucosal to the serosal side. This net flux of Ca2+ was abolished when ouabain was added to the serosal solution, indicating that the unidirectional Ca2+ flux JCams and, hence, the net flux of Ca2+ was dependent upon active Na+ transport accomplished by the Na, K-ATPase system. Under voltage clamp conditions, the net flux of Ca2+ was greatly increased at a potential difference PD = -25 mV (serosa negative) but disappeared at PD = +25 mV (serosa positive). Analysis of the unidirectional Ca2+ fluxes JCams and JCasm under varying potential differences indicated that both Ca2+ fluxes consisted of a PD-dependent and a PD-independent component. The results show that the reticulo-rumen participates in Ca2+ absorption in sheep.