Spontaneous transmitter release at the neuromuscular junction of the frog and rat was monitored during exposures to hyperosmotic solutions containing different sugars. Raising the osmolarity of the medium with D-glucose causes a marked, but transient, increase in the frequency of miniature end-plate potentials (MEPPs): after the initial elevation in frequency there is a subsequent decline towards the control levels, in spite of a continuous perfusion with the hyperosmotic solution. This decline occurs more rapidly in the frog. Two nonmetabolized analogues of glucose, 2-deoxy-D-glucose and 3-O-methylglucose, cause a transient hyperosmotic increase in MEPP frequency, which is very similar to the effect of D-glucose. The elevation of MEPP frequency with hyperosmotic glucose is stereospecific. Hyperosmotic solutions of L-glucose cause a sustained increase in transmitter release in the rat and frog. Insulin dramatically reduces the response of the frog nerve terminal to hyperosmotic D-glucose. Phenolphthalein, a glucose transport blocker, reduces or eliminates the secondary decline in MEPP frequency. It is suggested that the transient nature of the response to hyperosmotic solutions reflects the penetration of the hyperosmotic agent into the nerve terminal. The rate of decline of the MEPP frequency presumably indicates the rate of transport, which determines the rate of osmotic equilibration. This rate can then serve as an index of the relative permeability of the functioning presynaptic membrane to different sugars.