The size of motor variability increases with fatigue in repetitive upper limb tasks, and the structure of variability differs with old age. However, the combined influences of old age and fatigue on the size and structure of movement-to-movement variability are unclear.

Eighteen young and sixteen old adults performed a fatiguing repetitive tapping task while seated using their dominant arm. Optoelectronic motion capture served to measure upper body angles via forward kinematics. Movement-to-movement variability was measured by the size at joints (standard deviation:​ SD) and by the structure of the uncontrolled manifold (variance:​ V_UCM, V_ORT;​ synergy index:​ ΔV_z) in the first and final minutes of the task for the early, middle, and late forward movement phases. Outcomes were analyzed by Age*Condition*Phase general estimating equations.

Old adults had lower humerothoracic abduction/adduction and flexion/extension SD, wrist flexion/extension SD, V_UCM, and V_ORT, mainly in the early movement phase (p < 0.014). With fatigue, humerothoracic flexion/extension SD increased in young adults only and humerothoracic abduction/adduction SD, wrist pronation/supination SD, wrist flexion/extension SD, and V_UCM increased in both groups. ΔV_z was positive and did not differ with age or fatigue (p > 0.014).

Results indicate that fatigue adjustments were mainly in the frontal plane, old age did not affect the ratio of good vs. bad variability, and motor synergy was preserved during fatigue despite less motor flexibility in old age.