Anaerobic capacity was assessed by calculating the maximal accumulated oxygen deficit (AOD) in 7 male and 6 female children (aged 10.8 and 10.7 years, respectively) who exercised "all-out" to fatigue on a CYBEX TM (Met, 100) isokinetic cycle ergometer. Each child performed a series of submaximal steady state oxygen uptake tests and a peak oxygen uptake test. Data from submaximal and peak oxygen uptake tests were used to predict the oxygen demand for each child as they worked under "all-out" isokinetic cycling to volitional fatigue. The AOD was calculated as the difference between the predicted oxygen demand and the actual oxygen uptake for the duration of the test. In order to examine the repeatability of the testing the children performed two "all-out" tests on an isokinetic cycle ergometer. The absolute (I) and relative (ml kg1) AOD values for the males were not significantly different between trials (1.34 + or - .13 and 1.41 + or - 16 litres, and 35.6 + or - 3.7 and 37.7 + or - 4.7 ml-kg1, for trials one and two, respectively) (Mean +SEM). Similarly there were no significant differences in the AOD values for the female children between trials (1.40 + or -.13 and 1.37 + or -.16 litres and 35.6 + or - 3.8 and 34.6 + or - 3.6 for ml-kg1, for trials one and two, respectively). Over the two tests the respective peak and mean power results revealed no significant differences between the sexes (307.4 and 192.4 W for males and 274.3 and 192.6 W for females). Intraclass correlation coefficients (R) for the AOD results in litres and ml-kg in the tests were higher in the males (.96 and .95) than the females (.86 and .89), respectively. It was concluded that the isokinetic testing produced AOD results which were repeatable in this group of children, showed no sex differences and were comparable to those from children who had been tested using a constant power to exhaustion protocol.