Walking is a complex process and the physiotherapist must focus on physical signs as well as functional and practical tests to evaluate treatment. Measurement of energy expenditure during level walking is a useful objective parameter for assessing walking as being a valuable supplement to evaluate the outcome of physiotherapy. This study had two purposes. The first purpose was to investigate whether sampling and measurement of oxygen consumption were reproducible when using two different devices during walking on a treadmill. A second purpose was to find out whether the measurements were sensitive enough to reveal differences in energy expenditure and respiratory quotient (RO) during different walking speeds. Ten healthy students (mean age 22 years; range 20-25 years) volunteered in the test-retest of a slightly adjusted Deltatrac metabolic monitor. Thirteen volunteers (mean age 45 years; range 31-57 years) participated in the test-retest of Sensormedics 2900. In the first test, the subjects walked for 10 min in order to get a steady state both at an individual comfortable speed and at a preset speed. The retest was done with the same design and within 2 weeks. Nobody experienced any discomfort during the tests. The repeatability of measuring energy expenditure and RQ was acceptable for both methods and the methods were sensitive in revealing differences in energy expenditure during different walking speeds. The RQ were, however, too low when using the Deltatrac monitor, probably as a result of low air flow. We therefore conclude that only the method using Sensormedics 2900 may be used for the evaluation of energy expenditure during walking on a treadmill.