Single rat ventricular myocytes, voltage-clamped at -50 to -40 mV, were depolarized in small steps in order to define the mechanisms that govern the increase in cytosolic [Ca ~+] (Ca i) and contraction, measured as a reduction in myocyte length. Small (3-5 mV), sustained (seconds) depolarizations that caused a small inward or no detectable change in current were followed after a delay by small (<2% of the resting length), steady reductions in cell length measured via a photodiode array, and small, steady increases in Ca i measured by changes in Indo-1 fluorescence. Larger (greater than -30 and less than -20 mV), sustained depolarizations produced phasic Ca 2+ currents, Ca i transients, and twitch contractions, followed by a steady current and a steady increase in Ca~ and contraction. Nitrendipine (or Cd, verapamil, or Ni) abolished the steady contraction and always produced an outward shift in steady current. The steady, nitrendipine-sensitive current and sustained increase in Ca i and contraction exhibited a similar voltage dependence over the voltage range between -40 and -20 mV. 2/zM ryanodine in the presence of intact Ca 2+ channel activity also abolished the steady increase in Cal and contraction over this voltage range. We conclude that when a sustained depolarization does not exceed about -20 mV, the resultant steady, graded contraction is due to SR Ca 2+ release graded by a steady ("window") Ca 2 § current. The existence of appreciable, sustained, graded Ca 2 § release in response to Ca ~+ current generated by arbitrarily small depolarizations is not compatible with any model of Ca2+-induced Ca ~+ release in which the releasing effect of the Ca 2+