1. Skeletal muscles from rat, mouse and frog were incubated under different conditions and the amounts of acetylcholine (ACh) extractable from the tissue and released into the medium were determined by mass fragmentography. In some experiments measurements were made of the amounts of ACh ('bound' ACh) surviving in a muscle homogenate to which an excess of acetylcholinesterase had been added. In other experiments the membrane potentials, end‐plate potentials (e.p.p.s), and miniature end‐plate potentials (m.e.p.p.s) were studied. 2. During incubation in Ringer medium the ACh content of the rat hemidiaphragm usually did not change, but after inhibition of cholinesterase by soman the ACh content rose gradually from about 100 to 150 pmol to a plateau of about 400 pmol after 4 h. A similar formation of 'surplus ACh' after cholinesterase inhibition was found in the mouse diaphragm, but not in the frog sartorius muscle. 3. Surplus ACh accumulated predominantly in the end‐plate region of the rat diaphragm. In muscles, 16‐18 h after in vivo denervation, the capacity to form surplus ACh was decreased by more than 80%. 4. The amount of ACh diffusing from the resting hemidiaphragm into the incubation medium ('resting release') varied between 0.5 and 0.9 pmol min‐1 in different experiments; it remained at the same level during accumulation of surplus ACh. It was reduced by more than 80% 16‐18 h after denervation. 5. The amplitude of m.e.p.p.s and e.p.p.s did not increase while surplus ACh was accumulating. 6. Incubation of hemidiaphragms in Ringer solution containing [3H]choline caused the formation of [3H]ACh. Additional amounts of [3H]choline were incorporated into ACh when the nerve was stimulated for 60 min. However, incubation in the presence of soman (3,3‐dimethyl‐2‐butylmethylphosphonofluoridate), in the absence of stimulation, did not cause an increase of the [3H]ACh content of the muscles. 7. From hemidiaphragms with active cholinesterase about 120 pmol ACh was lost after prolonged nerve stimulation or incubation with 50 mM‐KCl in the presence of hemicholinium‐3, and about 35 pmol remained in the tissue. In soman‐treated muscles, containing surplus ACh, about as much ACh was released by nervous stimulation as from untreated hemidiaphragms, and much more ACh remained unreleased. 8. Transection of the muscle at both sides of the end‐plate or incubation of intact muscles in the presence of 50 mM‐KCl depolarized the muscle fibres to ‐35 and ‐31 mV, respectively. Surplus ACh was partially released by 50 mM‐KCl, but not by muscle transection.(ABSTRACT TRUNCATED AT 400 WORDS)