Dynamic aspects of melatonin's actions merit increasing future attention. This concerns particularly entirely different effects in senescent, weakened oscillators and in dysregulated oscillators of cancer cells that may be epigenetically blocked. This is especially obvious in the case of sirtuin 1, which is upregulated by melatonin in aged tissues, but strongly downregulated in several cancer cells. These findings are not at all controversial, but are explained on the basis of divergent changes in weakened and dysregulated oscillators. Similar findings can be expected to occur in other accessory oscillator components that are modulated by melatonin, among them several transcription factors and metabolic sensors. Another cause of opposite effects concerns differences between nocturnally active laboratory rodents and the diurnally active human. This should be more thoroughly considered in the field of metabolic syndrome and related pathologies, especially with regard to type 2 diabetes and other aspects of insulin resistance. Melatonin was reported to impair glucose tolerance in humans, especially in carriers of the risk allele of the MT receptor gene, MTNR1B, that contains the SNP rs10830963. These findings contrast with numerous reports on improvements of glucose tolerance in preclinical studies. However, the relationship between melatonin and insulin may be more complex, as indicated by loss-of-function mutants of the MT receptor that are also prodiabetic, by the age-dependent time course of risk allele overexpression, by progressive reduction in circadian amplitudes and melatonin secretion, which are aggravated in diabetes. By supporting high-amplitude rhythms, melatonin may be beneficial in preventing or delaying diabetes.