Cross-sectional and short-term prospective studies in humans support the concept that low energy availability, and not other factors associated with exercise, causes the development of exercise-induced reproductive dysfunction. To rigorously test this hypothesis, we performed a longitudinal study, examining the role of low energy availability on both the development and the reversal of exercise-induced amenorrhea, using a monkey model (Macaca fascicularis). Eight adult female monkeys developed amenorrhea (defined as absence of menses for at least 100 d, with low and unchanging concentrations of LH, FSH, E2, and P4) after gradually increasing their daily exercise to 12.3 +/- 0.9 km/d of running over a 7- to 24-month period. Food intake remained constant during exercise training. To test whether amenorrhea is caused by low energy availability, four of the eight amenorrheic monkeys were provided with supplemental calories (138-181% of calorie intake during amenorrhea) while they maintained their daily training. All four monkeys exhibited increased reproductive hormone levels and reestablished ovulatory cycles, with recovery times for circulating gonadotropin levels ranging from 12-57 d from the initiation of supplemental feeding. The rapidity of recovery within the reproductive axis in a given monkey was directly related to the amount of energy that was consumed during the period of supplemental feeding (r = -0.97; P < 0.05). Repeated measurements of plasma T3 concentrations, a marker of cellular energy availability, revealed a tight correlation between the changes in reproductive function and T3 levels, such that T3 significantly decreased (27%) with the induction and significantly increased (18%) with the reversal of amenorrhea (P < 0.05). These data provide strong evidence that low energy availability plays a causal role in the development of exercise-induced amenorrhea.