1. Miniature end-plate currents were recorded at neuromuscular junctions of toads, either in voltage-clamped fibres or with extracellular electrodes. The two methods gave similar results. 2. Two types of m.e.p.c.s, differing in their growth times (50-300 musec and 0-5-5 msec) were found. The more frequent had the shorter growth times. 3. The decay of m.e.p.c.s was exponential with a single time constant. The time constant was an exponential function of membrane potential, becoming less as the membrane was depolarized. In contrast, there was little change, or in some cases an increase, in growth times of m.e.p.c.s when the membrane was depolarized. 4. The decay time constant had a Q-10 of 3-13 plus or minus 0-22 (mean plus or minus S.E. of mean) whereas the growth time had a significantly lower Q-10 of about 1-2. The change of decay time constant with temperature followed the Arrhenius equation giving an activation energy of 18 plus or minus 1-0 kcal (mean plus or minus S.E. of mean). The amplitude of m.e.p.c.s increased with temperature and had a Q-10 of 1-5 plus or minus 0-14 (mean plus or minus S.E. of mean) in voltage-clamped fibres. 5. Neostigmine prolonged the decay phase and increased the amplitude of m.e.p.c.s but had little effect on the growth phase. The changes in m.e.p.c.s caused by membrane potential and temperature were not affected by neostigmine. 6. The results show that the growth phase and decay phase of m.e.p.c.s are governed by processes with quite different characteristics. The reaction which limits the decay phase appears first-order, is voltage sensitive and has a Q-10 of about 3, whereas the reaction underlying the growth phase does not appear first-order, is less voltage-sensitive and has a Q-10 of about 1-2. It is suggested that diffusion of transmitter across the synaptic cleft may be the rate-limiting step during the growth phase.