It was recently hypothesized that elite breath-hold divers may display blunted ventilatory responses to hypoxia and/or hypercapnia (Ferretti et al., J. Appl. Physiol. 70: 794-802, 1991). To test this hypothesis, the following measurements were made on three elite breath-hold divers (members of the same family), and on 9 healthy untrained control subjects (C): (1) Steady-state pulmonary ventilation (VE) at rest in the supine posture while breathing room air or normoxic CO2-enriched mixtures. (2) Breath-by-breath VE changes (delta VE), with respect to baseline conditions, after 4 breaths of 100% O2, under the following conditions: normoxia (PIO2 = 146 Torr) at rest (NR); normoxic exercise (60 watt on a bicycle ergometer) (NE); hypoxia (PIO2 = 77 Torr) at rest (HR); hypoxic exercise (HE). The results were as follows: (1) In hypercapnic experiments VE (normalized per unit of body surface area) was significantly lower in the divers than in C (4.32 +/- 0.04 [mean +/- SD]L.min-1.m-2 vs. 5.31 +/- 0.62 at FICO2 = 1.5%; 5.21 +/- 0.17 vs. 7.72 +/- 1.39 at FICO2 = 3%; 8.86 +/- 0.76 vs. 13.14 +/- 2.27 at FICO2 = 5%), as well as than in subjects described by previous authors as being characterized by 'low CO2 sensitivity'. (2) The 100% O2-breathing maneuvers did not induce significant delta VE both in NR and in HR, whereas peak delta VE were -6.73 +/- 1.38 L.min-1 (divers) vs. -5.24 +/- 3.10 (C) in NE, and -17.39 +/- 4.92 (divers) vs. -17.52 +/- 6.32 (C) in HE (no significant differences). It is concluded that the divers, compared to C, had a blunted ventilatory response to hypercapnia, but not to hypoxia. The former may represent an adaptive or genetically inherited phenomenon.