Fig. 3. The three-colour (black, light grey, dark grey) curves represent the maximum speed/distance relationships calculated for constant metabolic cost for each skate model. Data for klapskates were calculated assuming 5% faster speeds. The broken line reports values for running and is shown as a comparison to the ice-skating data. Obtained from equations provided by Wilkie (Wilkie, 1980), Saltin (Saltin, 1973) and Davies (Davies, 1981), the three-colour curves are based on the assumption that the available fraction of the metabolic power used for a physical activity is inversely related to the time to exhaustion (from the left; black, 40 s–10 min; light grey, 10 min–1 h; dark grey, 1–24 h). For the calculations, the maximum metabolic power available has been set at 21.3 W kg^–1. The light grey curves are iso-duration speed/distance pairs; the open squares represent the actual records in ice-skating and the open circles show records for cross-country skiing, reported as a means of comparison. Example: the energy cost of skating on bones (1800 BC) is indicated by the thick 345 J m^–1 iso-cost line. The intersection between this iso-cost line and the light 10 min iso-time line shows that in 10 min, for an energy cost of 345 J m^–1, a skater could cover a distance of 2638 m at an average speed of 4.4 m s^–1 before exhaustion. The energy cost of modern ice-skating is only 99 J m^–1, less than one-third of the energy cost associated with skating on bones. Consequently, in 10 min, a distance of almost 10 km can be travelled at an average speed of 16 m s^–1 before exhaustion, as indicated by the intersection between the 99 J m^–1 iso-cost curve and the 10 min iso-time line.