1. Transmitter release at the squid giant synapse was stimulated by photolytic release of Ca2+ from the 'caged' Ca2+ compound DM-nitrophen (Kaplan & Ellis-Davies, 1988) inserted into presynaptic terminals. 2. Competing binding reactions cause the amount of Ca2+ released by DM-nitrophen photolysis to depend on the concentrations of DM-nitrophen, total Ca2+, Mg+, ATP and native cytoplasmic Ca2+ buffer. Measurements of presynaptic [Ca2+] changes by co-injection of the fluorescent indicator dye Fura-2 show that DM-nitrophen photolysis causes a transient rise in Ca2+ followed by decay within about 150 ms to an increased steady-state level. 3. Rapid photolysis of Ca2(+)-loaded nitrophen within the presynaptic terminal was followed in less than a millisecond by depolarization of the postsynaptic membrane. As with action potential-evoked excitatory postsynaptic potentials (EPSPs), the light-evoked response was partially and reversibly blocked by 1-3 mM-kainic acid which desensitizes postsynaptic glutamate receptors. 4. Release was similar in magnitude and rate to normal action potential-mediated EPSPs. 5. The release of transmitter by photolysis of Ca2(+)-loaded DM-nitrophen was not affected by removal of Ca2+ from the saline or addition of tetrodotoxin. Photolysis of DM-nitrophen injected into presynaptic terminals without added Ca2+ did not stimulate release of transmitter nor did it interfere with normal action potential-mediated release. 6. Stimulation of presynaptic action potentials in Ca2(+)-free saline during the light-evoked response did not elicit increased release of transmitter if the ganglion was bathed in Ca2(+)-free saline, i.e. in the absence of Ca2+ influx. Increasing the intensity of the light or stimulating presynaptic action potentials in Ca2(+)-containing saline increased the release of transmitter. Therefore the failure of presynaptic voltage change to increase transmitter release resulting from release of caged Ca2+ was not due to saturation or inhibition of the release mechanism by light-released Ca2+. 7. Decreasing the temperature of the preparation increased the delay to onset of the light-evoked response and reduced its amplitude and rate of rise to an extent similar to that observed for action potential-evoked EPSPs.