Some evidence against the chemical theory of neuromuscular transmission