To investigate the relationship between energy availability (dietary energy intake minus energy expended during exercise) and thyroid metabolism, we studied 27 untrained, regularly menstruating women who performed approximately 30 kcal.kg lean body mass (LBM)-1.day-1 of supervised ergometer exercise at 70% of aerobic capacity for 4 days in the early follicular phase. A clinical dietary product was used to set energy availability in four groups (10.8, 19.0, 25.0, 40.4 kcal.kg LBM-1.day-1). For 9 days beginning 3 days before treatments, blood was sampled once daily at 8 A.M. Initially, thyroxine (T4) and free T4 (fT4), 3,5,3'-triiodothyronine (T3) and free T3 (fT3), and reverse T3 (rT3) were in the normal range for all subjects. Repeated-measures one-way analysis of variance followed by one-sided, two-sample post hoc Fischer's least significant difference tests of changes by treatment day 4 revealed that reductions in T3 (16%, P < 0.00001) and fT3 (9%, P < 0.01) occurred abruptly between 19.0 and 25.0 kcal.kg LBM-1.day-1 and that increases in fT4 (11%, P < 0.05) and rT3 (22%, P < 0.01) occurred abruptly between 10.8 and 19.0 kcal.kg LBM-1.day-1. Changes in T4 could not be distinguished. If energy deficiency suppresses reproductive as well as thyroid function, athletic amenorrhea might be prevented or reversed by increasing energy availability through dietary reform to 25 kcal.kg LBM-1.day-1, without moderating the exercise regimen.