Previous investigation has shown that the neuromuscular junction is the site of the paralytic action of botulinum toxin, since this toxin affects neither conduction nor skeletal muscle excitability. Within the neuromuscular junction, the toxin either blocks the terminal nerve fibrils or in some fashion depresses the end plate''s "transmitter" mechanism. To determine which of these hypotheses is correct, the sartorius muscle-nerve prepn. of a frog was bathed in a type A botulinum toxin soln. of 1.25 LD50/ml. and the muscle action potentials were recorded. During a period of approx. 15 min. after a previously supramaximal stimulus fails to elicit a muscle action potential, 2 such stimuli delivered in rapid succession cause a muscle action potential and a contraction. If the toxin were blocking the terminal nerve fibrils, no potential would be recorded in response to the 1st stimulus which would merely increase the excitability of the nerve on the distal side of the block. However, in response to the 1st stimulus, an end plate potential can be recorded which summates with the end plate potential elicited in response to the 2d stimulus to produce a muscle action potential. In terms of the acetyl choline theory, the 1st stimulus releases a subthreshold amt. of acetylcholine which adds to that released by the 2d stimulus and contraction follows. The toxin is, therefore, acting in some fashion on the end plate mechanism, probably preventing the release of acetylcholine, for it is known that botulinum toxin affects neither the choline-acetylating enzymes nor choline esterase.