Previous studies from this department have emphasized the wide variations occurring amongst healthy young adults, both in the daily intake of energy as judged by records of food consumption, and in energy expenditure recorded by indirect calorimetry, for a set task such as walking under standard conditions (Kitchin and others, 1949; Delbue and others, 1949). Although no correlation was demonstrated between energy intake or food consumption and energy expen diture, the suggestion was put forward that these variations might reflect wide spread differences in internal economy or individual muscular efficiency, and the possibility was mooted that some " can go more miles to the gallon " than others. Since walking is such a complicated muscular movement and involves no readily measurable performance of external work, it is not well suited for a study of varia tions in muscular efficiency. Two classical techniques have been used for this purpose, involving the use of either a stationary bicycle ergometer or a power-driven treadmill with variable gradient. These machines have provided valuable informa tion concerning energy expenditure, but each suffers from an important drawback. With the bicycle ergometer, there is the effect of practice and training on the efficiency with which the special movements involved in stationary bicycling are carried out; with the treadmill the initial cost is very high. A study of energy expenditure involved in performing the work necessary to lift the bodyweight up and down on to a series of stools of graded heights offers an alternative technique for assessing variations in muscle efficiency. It is well known that efficiency varies with the speed of movement. Accordingly, a series of observations have been made stepping at different speeds on to stools at different heights. The effects of stepping on to stools of standard size upon the cardio-respiratory system have already been widely used as tests of physical fitness, and are the basis of the Harvard step test (Brouha and others, 1944). Methods