Synthetic calcium buffers, including fluorescent calcium indicators, were microinjected into squid 'giant' presynaptic nerve terminals to investigate the calcium signal that triggers neurotransmitter secretion. Digital imaging methods, applied in conjunction with the fluorescent calcium indicator dye fura-2, reveal that transient rises in presynaptic calcium concentration are associated with action potentials. Transmitter release terminates within 1-2 ms after a train of action potentials, even though presynaptic calcium concentration remains at micromolar levels for many seconds longer. Microinjection of the calcium buffer, EGTA, into the presynaptic terminal has no effect on transmitter release evoked by single presynaptic action potentials. EGTA injection does, however, block the change in calcium concentration measured by fura-2. Therefore, the calcium signal measured by fura-2 is not responsible for triggering release. These results suggest that the rise in presynaptic calcium concentration that triggers release must be highly localized to escape detection with fura-2 imaging. Unlike EGTA, microinjection of BAPTA--a calcium buffer with an equilibrium affinity for calcium similar to that of EGTA--produces a potent, dose-dependent, and reversible block of action-potential evoked transmitter release. The superior ability of BAPTA to block transmitter release apparently is due to the more rapid calcium-binding kinetics of BAPTA compared to EGTA. Because EGTA should bind calcium within a few tens of microseconds under the conditions of our experiments, the inability of EGTA to block release indicates that transmitter release is triggered within a few tens of microseconds after the entry of calcium into the presynaptic terminal.(ABSTRACT TRUNCATED AT 250 WORDS)