Cardioventilatory coupling (CVC) is the temporal coherence of respiratory and cardiac rhythms. We have suggested that this coherence is the result of triggering of inspiratory onset by a preceding cardiovascular afferent. One implication of this triggering hypothesis is that coupling should only exist under conditions of spontaneous ventilation and not under conditions where the ventilatory period is fixed, as during intermittent positive pressure ventilation (IPPV). This study compared the degree of CVC in 20 ASA I female subjects, aged 21-50 yr, 10 of whom were breathing spontaneously and 10 were undergoing mechanical ventilation. Over 5-10 min, we recorded the timing of consecutive ECG R waves and inspiratory onsets. Coupling was demonstrated by examining these epochs for constant timing relationship between R waves and inspiration (RI intervals). Constancy of RI intervals was examined graphically (RI plot) and quantitatively using the Shannon entropy of RI interval distribution. We observed CVC in all spontaneously breathing subjects but in none of those receiving IPPV. In spontaneously breathing subjects, temporal alignment with inspiratory onset was greatest for the heart beat preceding inspiration. We conclude that although coupling has been shown to persist in the presence of electrical cardiac pacing, IPPV disrupts the coupling process, consistent with the view that in anaesthetized subjects, coupling is the triggering of inspiratory onset by a preceding heart beat and not the converse.