A mathematical model of short term cardiovascular regulation is used to investigate how heart period variability reflects the action of the autonomic regulatory mechanisms (vagal and sympathetic). The model includes the pulsating heart, the systemic (splanchnic and extrasplanchnic) and pulmonary circulation, the mechanical effect of respiration on venous return, two groups of receptors (arterial baroreceptors and lung-stretch receptors), sympathetic and vagal efferent branches, and a very-low frequency vasomotor noise. All model parameters were given on the basis of physiological data from the literature. We used data from humans whenever possible, while parameters for the regulation loops are derived from dogs' experiments. The model, with basal parameter values, produces an heart period power spectrum with two distinct peaks (the HF peak at the respiratory rate and a LF peak at about 0.1 Hz). Sensitivity analysis on the mechanism gains suggests that the HF peak is mainly affected by the vagal mechanism, whereas the LF peak is increased by a high sympathetic gain and reduced by a high vagal gain. Moreover, the LF peak depends significantly on the reactivity of resistance vessels and is affected by noise, amplified by the sympathetic control loop at its resonance frequency. The model may represent a new tool to study alterations in the heart period spectrum on the basis of quantitative physiological hypotheses.