Ear lobe blood pH(a), PaCO2, PaO2, and O-2 saturation (SaO2) were measured in healthy Caucasians and Sherpas at 3400 m (Namche Bazaar, Nepal, n = 4/5), 5050 m (Pyramid Laboratory, Lobuche, Nepal, n = 20/5) and 6450 m (Camp II of Mt Everest, n = 11/7). In the investigated altitude range, pH(a) increased progressively with altitude from 7.463+/-0.005 (mean+/-SE) to 7,496+/-0.006 in Caucasians whereas it remained essentially constant (7.45-7.46) in Sherpas. At all altitudes, PaCO2 was higher in Sherpas than in Caucasians (P < 0.02). By contrast, PaO2 and SaO2 were the same in Caucasians and Sherpas at all investigated altitudes. Moreover, in Caucasians sojourning for 3 weeks at 5050 m, PaCO2 kept decreasing whereas pH(a), PaO2 and SaO2 remained constant. These data suggest that: (1) respiratory alkalosis was a common finding both in Caucasians and Sherpas; (2) at 6450 m, Sherpas were less alkalotic due to higher PaCO2 than Caucasians, possibly a consequence of a blunted ventilatory response: (3) at 6450 m. SaO2 and PaO2 were the same in Caucasians and Sherpas despite different PaCO2 values. The latter finding could be the consequence of one or more of the following adjustments in Sherpas: (1), an increased efficiency of alveolar O-2 transfer, i.e. smaller alveolar-arterial O-2 gradient; (2) a decreased (arterial - mixed venous) O-2 difference, possibly due to increased cardiac output; (3) a reduced increase of the [2.3-DPG]/[Hb] ratio: but not (4) an elevated gas exchange ratio (R). It is concluded that both physiological and biochemical variables contribute to optimize the O-2 transport at altitude. Apparently a more efficient adaptation to hypoxia allows Sherpas to limit alkalosis through a lower ventilatory drive and to maintain SaO2 at the same PaO2 by decreasing the [2.3-DPG]/[Hb] ratio.