This review synthesizes a subset of human epidemiologic and experimental animal studies that suggest that early nutrition affects susceptibility to chronic diseases in adulthood. These studies provide evidence that biological mechanisms may exist to "memorize" the metabolic effects of early nutritional environments. However, hypothesis-driven investigations of potential mechanisms have been scant. Thus, our understanding of the biology underlying metabolic imprinting is incomplete. A working definition of metabolic imprinting is proposed, emphasizing the adaptive nature and limited ontogenic window of the mechanisms putatively responsible for these relations. Five specific candidate mechanisms of metabolic imprinting are elaborated: 1) induced variations in organ structure, 2) alterations in cell number, 3) clonal selection, 4) metabolic differentiation, and 5) hepatocyte polyploidization. Last, experimental approaches for probing potential mechanisms with animal models are discussed.