1. Monitoring of [Ca2+]i and vasopressin secretion in isolated nerve endings from the rat neurohypophysis were studied to determine the relationship between the time course of vasopressin secretion and depolarization-induced changes in [Ca2+]i. 2. Membrane depolarization by increasing the extracellular [K+] led to concentration-dependent, parallel increases in the amount of vasopressin release and in peak increases in [Ca2+]i. Half-maximal activation of a change in [Ca2+]i was attained at 40 mm extracellular K+. 3. The Ca2+ chelator dimethyl-BAPTA (1,2-bis(O-aminophenoxy)ethane-N,N,N'N'-tetraacetic acid), loaded into the nerve endings, reduced K+ depolarization-evoked vasopressin release and efficiently antagonized K+-induced changes in [Ca2+]i. Moreover, dimethyl-BAPTA dramatically reduced basal [Ca2+], without a reduction in basal secretion. 4. The duration of the vasopressin secretory response was similar regardless of applied 50 mm K+ depolarizations longer than 30 s. The t1/2 of the secretory response was 45 s. Application of repetitive K+ depolarization pulses produced repetitive secretory responses of similar amplitude and duration. 5. The K+-induced changes in [Ca2+]i remained elevated throughout the duration of the depolarizing stimulus decreasing less than 30 % over 3 min. The sustained increase in [Ca2+]i resulted largely from continued enhanced Ca2+ influx, demonstrated by susceptibility to the dihydropyridine, L-type calcium channel blocker, nicardipine. 6. Vasopressin secretion could be reinitiated following its decline to a step K+ depolarization by a further step increase in K+ or by removal and readdition of extracellular [Ca2+]. Alterations in [Ca2+]i paralleled periods of secretory activity. 7. Analysis of secretory responsiveness and change in [Ca2+]i to K+ depolarization in medium of altered extracellular [Ca2+] indicates that [Ca2+]i of 20 muM is sufficient to trigger vasopressin release. K+-induced alterations in [Ca2+]i could be observed at [Ca2+]. as low as 5 muM. Although smaller in amplitude to that observed at 2.2 mm [Ca2+]. the duration of the K+-induced secretory response increased at lower [Ca2+]o. 8. Transient vasopressin secretory responses were observed to sustained levels of [Ca2+] in digitonin and streptolysin-O-permeabilized nerve endings. Secretion could be re-evoked, following its decline, by a step increase in [Ca2+] or by removal and readdition of [Ca2+]o. 9. These results show that the amount and duration of depolarization-induced vasopressin secretion from isolated nerve endings may be regulated not only by the absolute increase but also by periodic changes in [Ca2+]i.