The relationship between presynaptic calcium transients ([Ca2+]t) and transmitter release evoked by a single stimulus was both investigated experimentally and modeled at a mammalian central synapse, the CA3 to CA1 pyramidal cell synapse in guinea pig hippocampal slices. In the present study, we compared the low-affinity calcium indicator furaptra with the higher-affinity indicator fura-2. The 10-90% rise time of the furaptra transient was 2.4 ms compared to 7.8 ms with fura-2; the half-decay time (tau 1/2) was 30 ms for furaptra, compared to 238 ms for fura-2. The half-width of the calcium influx was 1.8 ms with furaptra, which provides an upper limit to the duration of the calcium current (ICa) evoked by an action potential. Modeling the decay time course of the furaptra transients led to the conclusion that the predominant endogenous calcium buffer in these terminals must have relatively slow kinetics (kon < 10(5)/M.s), although the presence of small amounts of fast buffers cannot be excluded. The relationship between the [Ca2+]t measured with furaptra and the postsynaptic response was the same as previously observed with fura-2: the postsynaptic response was proportional to about the fourth power (m approximately 4) of the amplitude of either [Ca2+]t or calcium influx. Thus, although fura-2 may be locally saturated by the high local [Ca2+] responsible for transmitter release, the volume-averaged fura-2 signal accurately reflects changes in this local concentration. The result that both indicators gave similar values for the power m constrains the amplitude of calcium influx in our model: Ica < 1 pA for 1 ms.