Although leptin was first discussed as an "adipostat" that regulated food intake in accordance with triglyceride stores, it has become clear that leptin's role is much more complex and that a great many unanswered questions persist. Human studies have not shown that serum leptin concentrations can be changed rapidly by meals. Insulin seems to have a modest but immediate effect in attenuating the morning nadir of serum leptin as well as a greater effect manifested after 4-6 h. These changes in serum leptin, like the decreases that occur with fasting, are not accompanied by corresponding changes in adipose leptin mRNA, suggesting regulation by translational control or changes in the rate of leptin degradation, secretion or clearance. There is convincing evidence that insulin increases leptin synthesis and secretion, probably through an insulin-dependent effect on glucose metabolism. This effect of insulin is possibly mediated by the hexosamine pathway. What adipocytes seem to be communicating to the brain is not how much triglyceride they contain but whether they are currently synthesizing or hydrolysing triglyceride. Confounding many studies is the problem of leptin's diurnal rhythm. Because many studies only measured leptin during its morning nadir or examined the effects of insulin or specific nutrients provided after an overnight fast, important information on regulation may have been lost.