The elementary events of excitation-contraction coupling in heart
muscle are Ca$^2+$ sparks, which arise from one or more ryanodine
receptors in the sarcoplasmic reticulum (SR). Here a simple numerical
model is constructed to explore Ca$^2+$ spark formation, detection,
and interpretation in cardiac myocytes. This model includes Ca$^2+$
release, cytosolic diffusion, resequestration by SR Ca$^2+$-ATPases,
and the association and dissociation of Ca$^2+$ with endogenous
Ca$^2+$-binding sites and a diffusible indicator dye (fluo-3).
Simulations in a homogeneous, isotropic cytosol reproduce the brightness
and the time course of a typical cardiac Ca$^2+$ spark, but underestimate
its spatial size (approximately 1.1 micron vs. approximately 2.0
micron). Back-calculating Ca$^2+$i by assuming equilibrium
with indicator fails to provide a good estimate of the free Ca$^2+$
concentration even when using blur-free fluorescence data. A parameter
sensitivity study reveals that the mobility, kinetics, and concentration
of the indicator are essential determinants of the shape of Ca$^2+$
sparks, whereas the stationary buffers and pumps are less influential.
Using a geometrically more complex version of the model, we show
that the asymmetric shape of Ca$^2+$ sparks is better explained
by anisotropic diffusion of Ca$^2+$ ions and indicator dye rather
than by subsarcomeric inhomogeneities of the Ca$^2+$ buffer and
transport system. In addition, we examine the contribution of off-center
confocal sampling to the variance of spark statistics.