The suitability of the Wingate Anaerobic Test (WAT40) as a laboratory measure of anaerobic capacity (AnCap) and power (AnPow) of ice hockey players was tested against the Reed Repeat Sprint Skate-RSS (1979) and the Sargeant Anaerobic Skate (SAS40). Twenty-four university and Junior A players (20.2 +/- 1.6 years), assigned by random draw, performed the three tests over a seven day period. Blood lactate taken from an unwarmed finger tip was used to assess work intensity. The AnCap (7.7 +/- 0.2 Watts X kg-1) and AnPow (10.1 +/- 0.2 Watts X kg-1) for WAT40 were significantly lower (p less than 0.05) than for RSS (AnCap 9.3 +/- 0.8 Watts X kg-1; AnPow 11.5 +/- 1.1 Watts X kg-1) and SAS40 (AnCap 9.7 +/- 0.8 Watts X kg-1; AnPow 11.9 +/- 1.8 Watts X kg-1). SAS40 was significantly higher (p less than 0.05) than RSS for both AnCap and AnPow. The RSS (r = 0.96; ME 4.5%) and SAS40 (r = 0.97; ME = 3.6%) showed excellent test-retest reliability and reproducibility for AnCap but were only fair on AnPow (RSS: r = 0.73; ME = 10.7%; SAS40: r = 0.65; ME = 18.4%). While the correlations among the tests (AnCap: SAS40 vs WAT40, r = 0.73; RSS vs WAT40, r = 0.69) were significant (p less than 0.05), the highest predictive capability estimate (r2) was only 53.3%. The correlations for blood lactates (WAT40: 10.8 +/- 1.5 mmol X l-1; SAS40: 10.7 +/- 1.9 mmol X l-1; RSS: 11.5 +/- 1.6 mmol X l-1) were not significant. Based upon the particular protocol used, the laboratory test WAT40 does not demonstrate a high relationship with on-ice measures of AnCap and AnPow in this group of ice hockey players.