A double blind study of partial thiamin restriction was undertaken in 19 volunteer male medical students for a period of 4 to 5 weeks. Each day the subjects followed a diet supplying 500 µg thiamin and took 1 capsule randomly assigned to contain 5 mg thiamin hydrochloride or thiamin placebo. This capsule also supplied physiological doses of other vitamins. At the end of the study all were repleted with thiamin hydrochloride. Ten thiamin-depleted subjects were correctly identified by low urinary thiamin excretion, decreasing erythrocyte transketolase activity, and elevation of the thiamin pyrophosphate effect. The remaining nine subjects were controls. The degree of thiamin depletion induced varied between individuals and had no measureable ill effect on health as assessed by subjective feelings, physical findings, psychological testing, nerve conduction studies, and work performance. Subclinical thiamin deficiency is defined from this study as the presence of a low urinary thiamin level (27 µg thiamin/g creatinine) together with a thiamin pyrophosphate effect above 14.2% and below 35.4%. However, when urinary thiamin excretion is unknown, the data indicate that a thiamin pyrophosphate effect above 9% and below 41.6% is likely but not certain to represent subclinical thiamin deficiency. These guidelines, developed under experimental conditions, are a reference framework for assessment of thiamin status in population groups but should be applied with caution to individuals and to people with disease.