1. Mechanisms controlling the secretion of [(3)H]noradrenaline from the noradrenergic nerves of guinea-pig isolated vas deferens, prelabelled by incubation with [(3)H]noradrenaline, were studied using (a) different modes of (extramural or transmural) electrical nerve stimulation (a total of 300 shocks of varying strength, and a duration of 2 msec) at 1-30 Hz, or (b) depolarizing concentrations of K(+) (60-110 mm).2. The fractional rise in efflux of (3)H-labelled material (Deltat) was used to measure the secretion of [(3)H]noradrenaline.3. The dependence of [(3)H]noradrenaline secretion on the external Ca(2+) concentration (1-8 mm) was essentially hyperbolic. Double reciprocal plot analysis (1/Deltat vs. 1/Ca(2+)) of the data yields that blockade of alpha-autoinhibition (phentolamine 1 mum) does not increase the maximal secretory velocity, but does enhance the apparent affinity of the secretory mechanism for external Ca(2+). Exogenous noradrenaline has (qualitatively) opposite effects. The interaction between alpha-autoinhibition and external Ca(2+) thus shows a ;competitive' pattern, indicating that restriction of the utilization of external Ca(2+) is a major mechanism in alpha-autoinhibition of noradrenaline secretion, in this system.4. Phenoxybenzamine (10 mum) and phentolamine (1 mum) increased the secretion of [(3)H]noradrenaline evoked by depolarization with K(+) much less than that caused by electrical nerve stimulation (frequencies up to 10 Hz). Exogenous noradrenaline (1-5 mum) depressed the secretion evoked by both modes of stimulation. The results indicate that alpha-autoinhibition of [(3)H]noradrenaline secretion is mainly operative when the secretory stimulus requires conduction of nerve impulses between varicosities.5. The frequency dependence of [(3)H]noradrenaline secretion was hyperbolic, both in the presence and in the absence of alpha-autoinhibition; at each frequency the secretion (Deltat per shock) increased with the Ca(2+) concentration in the medium (0.6-8 mm). Double reciprocal plot analysis (1/Deltat vs. 1/frequency) of the data yields that the pattern of interaction between external Ca(2+) and facilitation depends on the presence or absence of alpha-autoinhibition (phentolamine 1 mum); in the former case it is ;non-competitive', in the latter ;competitive'. Similar analysis of the effect of facilitation by increasing the length of stimulus trains (from 5 to 300 pulses) at a constant frequency (5 Hz), on the Ca(2+) dependence of Deltat (1/Deltat vs. 1/Ca(2+)) in the absence of alpha-autoinhibition also yields that facilitation promotes utilization of external Ca(2+). These results apparently imply that a rise in external Ca(2+), in the presence of alpha-autoinhibition, augments the secretory response to electrical nerve stimulation mainly by promoting recruitment of active units (varicosities?), without markedly altering their ;affinity' for facilitation. In the absence of autoinhibition (when all units are already recruited?), the results seem to imply that facilitation promotes depolarization-secretion coupling in each, by more efficient utilization of external Ca(2+).6. The pattern of interaction between alpha-autoinhibition and facilitation depends on the Ca(2+)concentration in the medium. At or below the physiological level of Ca(2+) in extracellular fluid (1.2 mm) it is ;non-competitive', indicating that alpha-autoinhibition and facilitation act, at least in part, at separate targets under these conditions. At high (5.4 mm) external Ca(2+) the pattern becomes almost purely ;competitive', indicating that facilitation can, under suitable conditions, overcome all manifestations of alpha-autoinhibition.7. The secretion evoked by electrical nerve stimulation (Deltat per shock, at 1 or 10 Hz) increased with the strength of applied shocks, both when applied extra- or transmurally, in the presence or absence of alpha-autoinhibition. In the former case the rise in (Deltat per shock) vs. (current strength) was hyperbolic, in the latter it followed a biphasic pattern. Double reciprocal plot analysis (1/Deltat vs. 1/current) of the data yields a ;non-competitive' pattern of interaction between facilitation or alpha-autoinhibition, and exogenous current, when stimulation was extramural. When it was transmural the pattern is ;competitive'. The results seem to imply that hyperpolarization, or depolarization, of nerve terminals are major mechanisms whereby alpha-autoinhibition and facilitation, respectively, exert their effects on the secretory response to electrical nerve stimulation.8. Neither activation of Na(+), K(+)-ATPase, nor promotion of G(Cl) appear to be critically involved in alpha-autoinhibition. Experiments with known blockers of G(K) (tetraethylammonium, 4-aminopyridine and Rb(+)) did not give support to the notion that promotion of K(+) efflux is a mechanism whereby prejunctional alpha-adrenoceptors cause (hyperpolarization of nerve terminals and) autoinhibition of secretion. If alpha-autoinhibition does involve K(+) channels in the nerve terminal membrane, then these must be different from the (voltage-sensitive) K(+) channels blocked by the above mentioned inhibitors of K(+) efflux.9. The results are discussed in the context of a model that assumes that local control of noradrenaline secretion from noradrenergic nerves may be exerted both by control of invasion of terminals, and by control of depolarization-secretion coupling in each invaded varicosity. Under suitable conditions facilitation and alpha-autoinhibition may interact at both levels. It proposed that utilization of external Ca(2+) plays a pivotal role for both, and that restriction of invasion of nerve terminal varicosities is the main effect of alpha-autoinhibition, while promotion of depolarization-secretion coupling is the main effect of facilitation, at physiological concentrations of Ca(2+) in the medium. For the nerve the role of this dual control system is proposed to be to ensure ;rotational' activation of varicosities, and for the effector cell of noradrenergic junctions, to increase the signal/noise ratio.