The biological variation in maximum aerobic power (V O2max) was examined in four trained females and one trained male. An average of 8-20 repeat VO2max treadmill tests over a 2-4 wk period were performed on each subject (80 total tests). Biological variation (Si) in VO2max was computed as the standard deviation for the VO2max values for each individual, after subtracting net technological error (Se). Technological error was computed for each piece of equipment as the standard deviation of multiple trials. Results revealed that Si + Se amounted to +/- 5.6%. Biological variability accounted for 90% or more of this variability, while technological error accounted for less than 10%. In light of the magnitude of biological variation for VO2max, the necessity for securing control data when attempting to study training effects is pointed out.