that long periods are needed for equilibrium. In the case of erythrocyte folate, theoretically the erythrocytes have to turn over completely (approximately 90 days). Study design should allow for examination of the effects of initial status on response to maintenance or depletion and repletion

• Intakes used in repletion regimens should bracket the expected EAR intake to assess the EAR more accurately and to allow for a measure of variance. In addition, an accurate assessment of variance requires a sufficient number of subjects.

A relatively new and increasingly popular approach to determining requirements is kinetic modeling of body pools using steady-state compartmental analyses. This approach is unlikely to supplant depletion-repletion studies because it has a number of drawbacks; for example, assumptions that cannot be tested experimentally are often needed and the numbers obtained for body pool sizes are inherently imprecise. Even if accurate assessments of body pools were possible and were obtained, such information would be useful in setting a requirement only if the size of the body pool at which functional deficiency occurs could be established.

For all the B vitamins and choline, there is a serious lack of data useful for setting Estimated Average Requirements (EARs) for children, adolescents, pregnant and lactating women, and the elderly. Studies should use graded levels of nutrient intake and a combination of response indices and should consider other points raised in the preceding section. For some of the B vitamins (e.g., folate), studies should examine whether the requirement varies substantially by trimester of pregnancy. The nutrients and life stage and gender groups for which studies of requirements appear to be priorities from a public health perspective are vitamin B₁₂ requirements of the elderly and folate requirements by trimester of pregnancy. In addition, priority should be given to the identification of indicators on which to base vitamin B₆ requirements.

This short list does not imply that studies of requirements of other nutrients or age groups are not important, merely that it seems less likely that such studies will produce results that will have significant benefit on the health of the U.S. or Canadian populations. Research