1. The time course for the changes in miniature end-plate potential (min epp) frequency and in epp amplitude produced by alterations in the tonicity of the Ringer at the frog neuromuscular junction was studied. The relations between the tonicity and min epp frequency as well as epp amplitude were also investigated. 2. The change in min epp frequency occurred within 1 min after the start of the change in the tonicity of the extracellular solution. Following a shift to a hypertonic solution, the min epp frequencies were often maintained at a relatively steady, elevated level, even with large (+100 mosM) changes in tonicity. In other instances the elevation was transitory like the reported data for the rat neuromuscular junction. Essentially the same results were obtained in very low Ca2+-Ringer. Unlike the rat neuromuscular junction, the final level after hours of the increased min epp frequency caused by raising the osmolarity by more than 75 mosM was well above the control level. Following the return from a hypertonic to an initial solution there was a prompt decrease in min epp frequency to about the initial level; there was no indication of the transitory depression in min epp frequency following the return from hypertonic solution that has been reported in mammals. 3. Until the osmolarity of the Ringer reached about 420 mosM, the frequency of min epp continued to rise along a line relating log (min epp frequency) to (osmolarity)0.5. When the osmolarity exceeded 460 mosM, the relation started to level off. 4. The hypothesis that the min epp frequency in a Ringer with a given increased tonicity is a fixed multiple of the frequency in normal Ringer is not in accord with the data. 5. The decrease in epp amplitude caused by markedly hypertonic solutions also came about within 1 or 2 min after the start of the change in the tonicity of the solution surrounding the nerve terminal. 6. Hypertonic solutions did not appear to affect facilitation. 7. Below 360 mosM increasing the tonicity of the Ringer had little effect on the amplitude of epp. Above this level the amplitude decreased as the tonicity increased. At a given junction an increase in tonicity in a range above 360 mosM can cause an increase in min epp frequency and a decrease in epp amplitude. 8. The results are discussed in terms of the theories proposed to account for the effects of osmolarity on synaptic function. Two theories--the water flow hypothesis (11) and the barrier of water hypothesis (2)--do not fit with the results. The two other theories--calcium elevation (1) and screening of surface charges (3, 13, 21)--fail to account for important aspects of the results and therfore cannot be accepted without substantial modifications. None of the theories devised to account for the increase in min epp frequency predicts the falloff in frequency and in evoked quantal release that occurs in highly hypertonic solutions.