A rise in hemoglobin concentration ([Hb]) seems to be a reaction useful for survival and for performance in hypoxia as well as normoxia. But it increases blood viscosity and thus cardiac work as well as the risk of blood clotting. Enlarged cardiac output as alternative reaction only increases cardiac work. An evaluation of published data shows divergent results. In Caucasians [Hb] increases at altitude, but hemodilution does not reduce VO2. Also in the Andean population polycytemia is a regular effect but some investigations show negative correlations between health state and [Hb]. In the Arabian Highland the risk for thrombotic stroke increases with [Hb]. In Quechuas the drawback of the regular [Hb] increase for blood viscosity is compensated by low [fibrinogen]. Despite enlarged Hb mass trained highlanders in Colombia (2600 m) tend to lowered [Hb] compared to sedentary residents. The correlation between [Hb] and VO2 is weak or even lacking. Tibetans and Sherpas (4000 m) show only small increases in [Hb], Ethiopeans living at 3500 m of altitude possibly since 200000 years show no [Hb] increase and also no decrease of arterial O2 saturation. The latter is hardly explainable without a rise in Hb O2 affinity which has not been investigated until now. Possibly the Llama model of hypoxia adaptation (no [Hb] increase, small red cells, left shift of the O2 dissociation curve) is also used by human populations living since long time at altitude.