The purpose of this study was to establish the validity of a 90-s all-out test for the estimation of maximal oxygen uptake (V.O (2max)) and submaximal aerobic ability as represented by critical power. We hypothesized that the fall in power output by the end of the 90-s all-out test (end power) would represent the exhaustion of anaerobic work capability, and as such, would correspond with the critical power. Sixteen active individuals (mean +/- SD: 30 +/- 6 years; 69.6 +/- 9.9 kg) carried out a series of tests: (i) an incremental ramp test to determine V.O (2max), (ii) three fixed-work rate trials to exhaustion to determine critical power, and (iii) two 90-s all-out tests to measure end power and peak V.O (2). End power (292 +/- 65 W) was related to (r=0.89) but was significantly higher (p<0.01) than critical power (264 +/- 50 W). The mean +/- 95 % limits of agreement (29 +/- 65 W) were too low to use these variables interchangeably. The peak V.O (2) in the 90-s trial was significantly lower than the V.O (2max) (3435 +/- 682 ml x min (-1) vs. 3929 +/- 784 ml x min (-1); p<0.01); mean +/- 95 % limits of agreement was equal to 495 +/- 440 mL x min (-1). The 90-s all-out test cannot, therefore, assess both V.O (2max) and critical power in adult performers. The duration of all-out exercise required to allow V.O (2) to attain its maximum is longer than 90 s.