To investigate the responses of urinary creatinine (Crn) and 3-methylhistidine (3-MH) to diet, Crn and 3-MH were measured in healthy college women during two metabolic studies. Twenty subjects participated in Experiment 1. Vegetarian, ovo-lacto-vegetarian, and typical American diets provided no, trace, and customary dietary levels of Crn (and its precursors) and 3-MH. Twenty-four hour urine samples collected during days 3 and 4 of each diet were analyzed for Crn and 3-MH. In Experiment 2, ten women ate a constant diet for 28 days, then switched to a different diet for an additional 28 days. Daily 24-hour urine collections were analyzed for Crn and 3-MH;For Experiment 1, ANOVA indicated daily urinary Crn was significantly higher during Diet 1 than during either Diet 2 or Diet 3. Daily urinary 3-MH during Diet 3 was significantly higher than during either Diet 1 or Diet 2; this difference reflected the dietary 3-MH changes. Excretion patterns among the subjects on the three diets varied more for Crn than for 3-MH;Both urinary Crn and 3-MH excretion were quite variable during the 56 days on the constant diets of Experiment 2. Coefficients of variation (C.V.) were calculated by week for each of the 10 subjects. The C.V. for Crn excretion were generally less than 20 percent; no subject had C.V. less than 10 percent for all eight weeks. Urinary 3-MH was approximately twice as variable as Crn with C.V. ranging from 8.7 to 41.1 percent in all but one week for one subject. The weeks with the greatest variability for Crn were different from those with the most variability for 3-MH. Both urinary Crn and 3-MH decreased over the experimental feeding. The decline in Crn excretion was more evident (p < 0.023) than that in 3-MH (p < 0.134);In conclusion, urinary 3-MH responded to dietary changes within three days but Crn apparently was not stabilized within this time. Crn excretion on a constant diet was less variable than urinary 3-MH. The extended time required for Crn excretion to stabilize in response to dietary changes may negate its usefulness as an index of body composition. The problem of variability in 3-MH excretion may be lessened by collecting multiple consecutive 24-hour urine samples to improve the estimate of an individual's urinary 3-MH.