45Ca efflux and tissue Ca content were examined in goldfish ventricles under conditions known to affect cellular Ca movements. 2. EGTA or Ca-EGTA was added to the washout solutions in sufficient concentration (10 mM) to avoid retardation of the apparent tissue 45Ca efflux by extracellular 45Ca binding or backflux. 3. After a variable initial increase, the cellular Ca content usually stabilizes within 60 min when ventricles are immersed in Li- or K-substituted saline containing 1.8 mM Ca0 (under these conditions the internal Ca2+ concentration is below 10(-5) M). 4. 45Ca efflux is maximally activated by external concentrations of Ca2+ as low as 10(-6) M, in both Na-containing and Na-free saline. 5. 45Ca efflux decreases in Na-free solutions. It is reactivated by Na-saline. The effect of different external Na concentration on 45Ca efflux is comparable at external Ca2+ concentrations between 10(-6) M and 2 x 10(-3) M. 6. Reactivation of Ca efflux after Na0 readmission is inhibited by metabolic poisoning, or in the presence of 10 mM-caffeine. Loading with 45Ca at very low external Ca2+ concentration prevents the inhibition of Ca efflux in Na-free medium. 7. Caffeine (10 mM) produces contractions of about equla size when K-depolarized preparations are immersed in either Na- or Li-saline. At the same time there is a similar increase in 45Ca efflux in absence of Na0 and in its presence. 8. In the virtual absence of Ca2+0 (10(-5) M-Ca, 10(-2) M-EGTA) and Na+0, the residual 45Ca efflux is reversibly inhibited by cyanide (2 mM). 9. The results are roughly compatible with the general concept of ATP-dependent Na-Ca exchange in internal Ca2+ homeostasis. However, this hypothesis should probably be modified to account for the fact that under physiological concentrations Na+0 and Ca2+0 do not compete for activating 45Ca efflux. Metabolic products may be involved in Na0- and Ca0-dependent Ca efflux. It is therefore not excluded that a Na-independent active mechanism co-operates with Na-Ca exchange in Ca extrusion.