Many fluorescent Ca indicators, particularly those loaded as acetoxymethyl (AM)-ester derivatives, are known to become compartmentalized into organelles. This property can be exploited to measure changes in free [Ca] in subcellular compartments, including the inositol (1,4,5)-trisphosphate-sensitive store. However, quantitative measurement of free [Ca] within a particular compartment is complicated by the fact that dye may accumulate in a variety of organelles and, in many cases, by the Mg sensitivity of the indicator. Here the issue of the quantification of free [Ca] within the thapsigargin-sensitive store in BHK-21 fibroblasts using the low affinity Ca indicator, Mag-Fura-2, has been re-examined. At least 88 +/- 1.3% (SEM) of the compartmentalized dye was determined to be confined to the thapsigargin-sensitive store, with the remaining fraction accounted for by other compartments where [Ca] was below the detection limit for the dye (< 5 microM). In situ calibrations with ionophores indicated that the apparent free resting intraluminal [Ca] was 260 +/- 43 microM (SEM). Our analysis shows, however, that dye reporting from regions of low [Ca] contributes disproportionately to the Mag-Fura-2 ratio measured over the whole cell, potentially resulting in large underestimations of intraluminal [Ca] in agonist-sensitive stores. Free [Ca] in the agonist-sensitive store was calculated to be as high as 539 +/- 92 microM, assuming 12% of the Mag-Fura-2 to be in compartments where [Ca] was below 5 microM. In comparison, perturbations arising from the presence of Mg in stores are predicted to be relatively minor.