Whole cell patch clamp and intracellular Ca²⁺ transients in trout atrial cardiomyocytes were used to quantify calcium release from the sarcoplasmic reticulum (SR) and examine its dependency on the Ca²⁺ trigger source. Short depolarization pulses (2-20 ms) elicited large caffeine-sensitive tail currents. The Ca²⁺ carried by the caffeine-sensitive tail current after a 2-ms depolarization was 0.56 amol Ca²⁺/pF, giving an SR Ca²⁺ release rate of 279 amol Ca²⁺ · pF¹ · s¹ or 4.3 mM/s. Depolarizing cells for 10 ms to different membrane potentials resulted in a local maximum of SR Ca²⁺ release, intracellular Ca²⁺ transient, and cell shortening at 10 mV. Although 100 µM CdCl₂ abolished this local maximum, it had no effect on SR Ca²⁺ release elicited by a depolarization to 110 or 150 mV, and the SR Ca²⁺ release was proportional to the membrane potential in the range 50 to 150 mV with 100 µM CdCl₂. Increasing the intracellular Na⁺ concentration ([Na⁺]) from 10 to 16 mM enhanced SR Ca²⁺ release but reduced cell shortening at all membrane potentials examined. In the absence of TTX, SR Ca²⁺ release was potentiated with 16 mM but not 10 mM pipette [Na⁺]. Comparison of the total sarcolemmal Ca²⁺ entry and the Ca²⁺ released from the SR gave a gain factor of 18.6 ± 7.7. Nifedipine (Nif) at 10 µM inhibited L-type Ca²⁺ current (I_Ca) and reduced the time integral of the tail current by 61%. The gain of the Nif-sensitive SR Ca²⁺ release was 16.0 ± 4.7. A 2-ms depolarization still elicited a contraction in the presence of Nif that was abolished by addition of 10 mM NiCl₂. The gain of the Nif-insensitive but NiCl₂-sensitive SR Ca²⁺ release was 14.8 ± 7.1. Thus both reverse-mode Na⁺/Ca²⁺ exchange (NCX) and I_Ca can elicit Ca²⁺ release from the SR, but I_Ca is more efficient than reverse-mode NCX in activating contraction. This difference may be due to extrusion of a larger fraction of the Ca²⁺ released from the SR by reverse-mode NCX rather than a smaller gain for NCX-induced Ca²⁺ release.