This study, aimed at investigating some aspects of breathing control at work, was conducted on 8 heart and lung transplant recipients (HLTR) (age 33 +/- 13 years, mean +/- SD; 10 +/- 6 months post-transplantation) and on two control groups, i.e. 11 heart transplant recipients (HTR) and 11 healthy untrained subjects (C). The patients performed a series of 2 to 6 1-min exercise bouts (at 25 or 50 W, corresponding to about 50% of their VO2max) on a bicycle ergometer, followed by a 5 min 25 or 50 W constant load. C exercised both at 50 W (C1) and at 50% of their VO2max (C2). Inspiratory (VI) and expiratory (VE) ventilation, tidal volume (VT), respiratory frequency (fR), end-tidal O2 and CO2 partial pressures (PETO2 and PETCO2 and gas exchange (VO2 and VCO) were measured breath-by-breath. "Phase I" ventilatory response (ph I) was determined as the mean changes of VI, VE, VT, fR, PETO2 and PETCO2, compared to rest, during the first two respiratory cycles following exercise onset. In HLTR ph I did not significantly differ from that of C1 and C2, whereas the response was lower in HTR. VE, VO2 and VCO2 responses during "phase II" (t 1/2 on-) and "phase III" (steady state exercise) were similar in HLTR and in HTR. t 1/2 on- were longer in HLTR and in HTR compared to C1. In 3 HLTR the ventilatory pattern during the 5 min constant loads was similar to that of HTR and C, whereas 4 HLTR presented higher VT and lower fR values. It is concluded that: 1) The ventilatory response to exercise, in all its phases, is substantially preserved despite lung denervation. When slight alterations are found (i.e. the slower phase II), they are presumably of peripheral origin. 2) The normal ph I in HLTR indicates that cardiac and/or pulmonary inputs to the respiratory centers are not involved in its regulation, or that their role can be subserved by other ventilatory control mechanisms.