The use of binomial analysis as a tool for determining the sites of action of neuromodulators may be complicated by the non-uniformity of release probability. One of the potential sources for non-uniformity of release probability is the presence of multiple forms of synaptotagmins, the Ca²⁺ sensors responsible for triggering vesicular exocytosis. In this study we have utilized Sr²⁺, an ion whose actions may be restricted to a sub-population of synaptotagmins, in an attempt to obtain meaningful estimates of the binomial parameters p (the probability of evoked ACh release) and n (the immediate available store of ACh quanta, whereby m=np). In contrast to results in Ca²⁺ solutions, binomial analysis of Sr²⁺ dependent release reveals a dramatically reduced dependence of n on extracellular Sr²⁺ concentrations. In Sr²⁺ solutions, blockade of potassium channels with 3,4-diaminopyridine (3,4-DAP) increased m by an exclusive increase in p whilst treatment with phorbol ester increased m solely by effects on n. The cyclic AMP analogue CPT cAMP increased m by increasing both n and p. The effect of CPT cAMP on p but not n was blocked by protein kinase A inhibitors whilst the effect on n was mimicked by 8-CPT-2'-O-Me-cAMP, a selective agonist for Epac (Exchange Protein directly Activated by cAMP, otherwise known as the cAMP-sensitive guanine nucleotide-exchange protein, GEP). The results demonstrate both the utility of the binomial distribution in Sr²⁺ solutions and the dual effects of cyclic AMP on both protein kinase A (PKA) dependent and independent processes at the amphibian neuromuscular junction.