1. The fluorescent Na+ indicator SBFI was incorporated into isolated ventricular myocytes using the acetoxymethyl (AM) ester. 2. The excitation spectrum was found to be shifted about 20 nm in the cell compared to in vitro. In the cell, an increase of [Na+] decreased fluorescence at 380 nm (F380) and had no effect at 340 nm (F340). The ratio (R = F340/F380) was used as a measure of [Na+]i. 3. In vivo calibration of SBFI for [Na+]i was obtained by equilibrating [Na+] across the plasma membrane with a divalent-free solution in the presence of gramicidin D. 4. Selective removal of the surface membrane with saponin or digitonin released only about 50% of the indicator. Following saponin treatment, cyanide or carbonylcyanide m-chlorphenylhydrazone (CCCP) increased the apparent [Na+] measured by the remaining (presumably mitochondrial) SBFI. It is suggested that mitochondrial [Na+] is normally less than cytoplasmic. 5. Attempts to examine the effects of metabolic inhibition on [Na+]i were hampered by changes of autofluorescence due to changes of [NADH]. It is shown that this effect can be corrected for using the isosbestic signal (excited at 340 nm). 6. Inhibition of both aerobic metabolism (with CN-) and glycolysis (glucose removal or iodoacetate) produced a gradual increase of [Na+]i. This began before the resting contracture developed and may (via Na(+)-Ca2+ exchange) account for some of the rise of diastolic [Ca2+]i seen in previous work. The rise of [Na+]i began at about the same time as the decrease of systolic contraction and therefore at a time when [ATP]i had begun to fall.