It is well known that ankle joint power generation (A2) is reduced in healthy older adults (OG) during gait. No general consensus exists, however, as to what fundamental compensatory actions are made at the knee and hip joints by the OG to compensate for this loss of power. The failure to control gait speed may account for this lack of agreement. This study investigated the effect of aging on lower limb joint power and work during gait. The gait patterns of eight old (OG: 66.8±5.4 yr) and 12 young adults (YG: 26.6±2.9 yr) were recorded for a range of matched speeds (1.0 m s(-1), 1.3 m s(-1), 1.6 m s(-1)). Speed did not differ between the groups. Combining speeds, the OG generated 17% less A2 power and 21% less A2 work (p<0.05). Compared to the YG, the OG generated 46% more H1 work, 30% more H2 peak power, 16% more H3 peak power, 30% more K3 peak power and 19% more K4 peak power (p<0.05). These actions by the OG were associated with less ankle plantar-flexion, more hip flexion and anterior pelvic tilt (p<0.05). The OG adopted a different gait pattern at the faster speeds by generating more H3 work than A2 work. This shows the OG rely on hip flexors to propel the leg into swing when ankle plantar-flexor function is reduced. This may partly explain how gait changes emerge with aging.