The demonstration that the trace element vanadium has insulin-like properties in isolated cells and tissues and in vivo has generated considerable enthusiasm for its potential therapeutic value in human diabetes. However, the mechanisms by which vanadium induces its metabolic effects in vivo remain poorly understood, and whether vanadium directly mimics or rather enhances insulin effects is considered in this review. It is clear that vanadium treatment results in the correction of several diabetes-related abnormalities in carbohydrate and lipid metabolism, and in gene expression. However, many of these in vivo insulin-like effects can be ascribed to the reversal of defects that are secondary to hyperglycemia. The observations that the glucose-lowering effect of vanadium depends on the presence of endogenous insulin whereas metabolic homeostasis in control animals appears not to be affected, suggest that vanadium does not act completely independently in vivo, but augments tissue sensitivity to low levels of plasma insulin. Another crucial consideration is one of dose-dependency in that insulin-like effects of vanadium in isolated cells are often demonstrated at high concentrations that are not normally achieved by chronic treatment in vivo and may induce toxic side effects. In addition, vanadium appears to be selective for specific actions of insulin in some tissues while failing to influence others. As the intracellular active forms of vanadium are not precisely defined, the site(s) of action of vanadium in metabolic and signal transduction pathways is still unknown. In this review, we therefore examine the evidence for and against the concept that vanadium is truly an insulin-mimetic agent at low concentrations in vivo. In considering the effects of vanadium on carbohydrate and lipid metabolism, we conclude that vanadium acts not globally, but selectively and by enhancing, rather than by mimicking the effects of insulin in vivo.