Many problems concerning the human requirement for vitamin A could be solved if a reliable measure of subclinical vitamin A deficiency were found. The biophotometer test has been suggested for this purpose, but work in this as well as in other laboratories has shown that the test is only fairly reliable and that its relation to stores of vitamin A has yet to be demonstrated. A review of the literature suggests that determinations of vitamin A in the blood may offer another approach to this problem. The present study was consequently undertaken. It includes the results of chemical determinations of vitamin A and of carotene in the blood, as well as biophotometer tests, of five subjects over periods of from two to four months. During this period four of the subjects lived on vitamin A depletion diets save for two short periods of supplementation; the fifth served as the control.