1. Two series of experiments were performed in an attempt to elucidate the mechanism of relief of the discomfort of breath-holding.
2. In the first the effect of varying the size of a single breath at the breaking point of breath-holding on the time of a second breath-hold (BHT2) was observed in three normal subjects. The effect of two different non-ventilatory chest-wall manœuvres, performed at the breaking point, on the duration of a second breath-hold in two additional subjects was then studied.
3. The volume of an unrestricted relieving breath was always greater than 70% of vital capacity. When the size of the breath was restricted to as little as 20% of the free relieving volume, BHT2 was unchanged.
4. An inspiratory effort against an occluded airway or an isovolume movement of the rib cage and abdomen performed at the breaking point of breath-holding were as effective as a control relieving breath in allowing resumption of apnoea.
5. There seem to be two possible mechanisms of relief. First, afferent information may be generated by some consequence of diaphragmatic contraction. Secondly, changes of local pressure-volume relationships within the lung may alter the pattern of vagal afferent activity, independent of an overall change in lung volume.