The nadir of JSR calcium during a simulated spark is low unless RyR closing rate is strongly regulated by lumenal calcium. (A) Distributions of the nadir of [Ca²⁺]_JSR of simulated stochastic sparks with (green) or without (red) the lumenal activation factor in Eq. 1.1. Lumenal deactivation when the JSR is depleted contributes to termination of the spark at somewhat higher [Ca²⁺]_JSR, but the nadir is quite low in both cases, which is not compatible with the idea that the spark is shut off when lumenal calcium falls to a “threshold” of ∼50%. (B) Attempt to reproduce the 50% nadir by lowering the RyR mean open time so that sparks will extinguish promptly once CICR is no longer regenerative, and reducing the opening rate constant. This required RyR sensitivity so low that only 7% of attempted sparks were detectable, and they were very stochastic and either abortive or low amplitude. (C) A well-defined 40% nadir was produced by making the RyR mean open time strongly and nonlinearly dependent on [Ca²⁺]_JSR, thereby forcing RyRs to close below the threshold.