A branching treadmill protocol was designed to measure functional capacity in patients with low work capacity and varying ability to walk at speeds used in traditional protocols. A comfortable walking pace is first selected (2.0 to 3.5 mph, 0.25 mph increments) and the workload is then increased every 2 minutes in 1 MET increments (a multiple of the resting oxygen uptake [1 MET = 3.5 ml O2/kg/min]) by adjusting grade. Nine trained (maximal MET = 7.6 +/- 1.6, mean +/- standard deviation) male subjects (age 59 +/- 7 years) with previous myocardial infarction and 9 trained (maximal MET = 11.7 +/- 2.5) male control subjects (age 56 +/- 8 years) completed submaximal and maximal workloads without handrail support. The measured oxygen consumption, volume of oxygen in ml/kg/min (VO2), was compared with the predicted VO2 cost of treadmill walking calculated from speed and grade. A linear regression analysis of predicted versus measured VO2 was performed. There were no significant differences between myocardial infarction and control regression lines. Therefore, a simplified prediction equation for estimated VO2 in myocardial infarction and control subjects is proposed. Overall VO2 prediction = 1.61 + 0.99 x. The main advantage of the branching protocol format is the selection of a stable, brisk walking pace compatible with age and gait, which may improve mechanical efficiency through impedance matching. The protocol is adaptable enough in design so that most patients can complete the exercise test without use of handrails, which is essential for an accurate estimate of VO2 from treadmill speed and grade.