The metabolic demands for women hockey players (n=12) were estimated during real match play using heart rate analysis and time-motion analysis. An individual heart rate - oxygen uptake regression equation, established at steady state workloads on the treadmill, was applied to heart rates recorded throughout the match to estimate energy expenditure. A specific energy cost was assigned to each of nine discrete activities to give an energy cost for the whole match. The mean estimated energy cost for a complete match from heart rate analysis was 3873 +/-436 kj and from time-motion analysis, 2846 +/-284 kJ. In order to check the errors in these methodologies expired air was collected continually during 15 minutes intermittent activity on a treadmill (n=16) with the heart rates and work : rest ratios similar to those established in the earlier part of the study. The error in the use of heart rate to estimate energy expenditure was 3.7 +/-5.1% and for time-motion analysis was 16.6 +/-4.8%, when compared with the measured value from the analysis of expired air. It was concluded that heart rate gave a good estimation of energy expenditure during intermittent activity at workloads similar to women's hockey. In order to gain a greater insight into the metabolic demands of women's hockey both heart rate and time-motion analysis need to be applied simultaneously. The heart rate analysis suggested that the estimated energy expenditure was similar during the first and second halves, hi contrast the time-motion analysis established that less time was spent in high intensity activity during the second half. Women's hockey is played at greater intensities than previously reported with no differences in the metabolic demands when related to specific player positions.