The work of recent years has clarified the function of the human fetal adrenal gland to a point where a reasonable explanation of its several peculiarities appears possible. We believe that the fetal adrenal is a steroidogenic organ with a high biogenetic capacity. Its steroid products are qualitatively similar to those of the adult gland, but certain quantitative differences may be present. There is indirect evidence that relatively little aldosterone is produced or required by the fetus. The placenta serves to regulate electrolyte homeostasis in the mammalian fetus. We believe that the primary adrenotropin stimulating the fetal adrenal is ACTH from the fetal pituitary and that the enlargement of the fetal adrenal reflects the intensity of the stimulus. Since there is some evidence that in the fetus, as in the adult, adrenal steroids control a feed-back mechanism operating between the pituitary and adrenal, we believe that strong ACTH stimulation reflects an increased fetal demand for corticosteroidogenesis. This increased demand could be secondary to either defective production of steroids by the fetal adrenal or increased metabolism of steroids by the fetus. A likely, though unproved, mechanism is destruction of corticosteroids in the placenta. We believe that species differences in fetal adrenal size, and the atrophic adrenals in human anencephalic monsters, reflect variations in the intensity of fetal pituitary ACTH stimulation, and that the poorly differentiated glomerular zone common to all mammals reflects the lack of stimulation by the tropic hormone for aldosterone. There is at present no evidence that the apparently active human fetal adrenal gland plays a critical role in normal fetal development. We have attempted to indicate those numerous areas where appropriate data with which to test our interpretation of fetal adrenal physiology are not presently available. It would be surprising if our concept survives these tests entirely unmodified. We hope, however, that this interpretation may indicate profitable areas for further work, and that it is sufficiently well founded on observation and experimentation that its basic concepts will prove to be reliable.