The role of plasma membrane depolarization as a determinant of the initial phase of insulin secretion was investigated. NMRI mouse islets and beta cells were used to measure the kinetics of insulin secretion, ATP and ADP content, membrane potential and cytosolic free calcium concentration ([Ca²⁺]_i). The depolarization of metabolically intact beta cells by KCl corresponded closely to the theoretical values. In contrast to physiological (glucose) or pharmacological (tolbutamide) KATP channel block, KCl depolarization did not induce action potential spiking. The depolarization by 15 mM K⁺ (21 mV) corresponded to the plateau depolarization by 50 or 500 µM tolbutamide, that by 40 mM K⁺ (41 mV) corresponded to the action potential peaks. Nifedipine and diazoxide abolished action potentials but not KCl depolarization suggesting that the depolarizing strength of 15, but not 40 mM K⁺ corresponds to that of KATP channel closure. 40 mM K⁺ induced a massive secretory response in the presence of 5 mM glucose, whereas 15 mM K⁺, like 50 µM tolbutamide was only slightly effective, even though a marked increase in [Ca²⁺]_i was produced. Raising glucose from 5 to 10 mM in the continued presence of 15 mM K⁺ resulted in a strongly enhanced biphasic response. The depolarization pattern of this combination could be mimicked by combining basal glucose with 15 mM K⁺ and 50 µM tolbutamide, however, the secretory response to these non-nutrients was much weaker. In conclusion, the initial secretory response to nutrient secretagogues is largely influenced by signalling mechanisms which do not involve depolarization.