1. The temperature dependence of presynaptic processes involved in neuromuscular transmission was studied by rapidly increasing the temperature of cooled frog neuromuscular junctions by 4--10 degrees C using pulses from a neodymium laser. The temperature elevation was complete within 0.5 msec, and decayed back to control levels with a time constant of about 7--8 sec. 2. Temperature jumps completed before nerve stimulation increased the quantal content and decreased the latency of the end-plate potential (e.p.p.). The Q10 for e.p.p. quantal content in low [Ca2+] Ringer averaged about 3.9 over the range 1--18 degrees C. 3. Temperature jumps occurring during the synaptic delay (the interval between the presynaptic action potential and the onset of the e.p.p.) also increased the quantal content and decreased the latency of the e.p.p. These effects diminished as the onset of the temperature jump was moved closer to the expected onset of the e.p.p. Temperature jumps applied after the onset of the e.p.p. immediately accelerated the time course of the e.p.p. but did not significantly alter quantal content. These results demonstrate that the magnitude and timing of evoked release are influenced by temperature-sensitive processes that operate both during and shortly after the presynaptic nerve action potential, but are largely complete before the onset of release. 4. Temperature jumps were applied at various times during the interval between two nerve stimuli. The amplitude of the second e.p.p. decreased as the temperature jump was moved earlier in the interstimulus interval, suggesting that the rise in temperature following the first nerve stimulus accelerates the decay of facilitation. When the temperature jump was moved from 10 msec after to 10 msec before the onset of the first e.p.p., the amplitude of the second e.p.p. either decreased or showed no change. The fact that the second e.p.p. did not increase suggests that the temperature-sensitive processes that increase the quantal content of the conditioning e.p.p. do not greatly increase the facilitation following that e.p.p. 5. Temperature jumps immediately accelerated the time course of spontaneous miniature end-plate potentials (m.e.p.p.s) and increased their frequency. Experiments using slow temperature changes revealed that the Q10 for m.e.p.p. frequency in normal Ringer is about 10 over the range 10--20 degrees C. M.e.p.p. frequency was much less sensitive to temperature changes below about 10 degrees C. When the nerve terminal was depolarized by 20 mM-K+ in the presence of Ca2+, the Q10 for the rate of spontaneous release over the range 10--20 degrees C decreased to about 4, similar to the Q10 for e.p.p. quantal content. In the absence of extracellular Ca2+ the Q10 for m.e.p.p. frequency in 20 mM-K+ remained near 10. 6. The marked difference in Q10S for spontaneous transmitter release under different experimental conditions suggests that not all transmitter release uses identical mechanisms...