Melatonin, the hormone of the pineal gland and product of numerous other cells and tissues, has attracted considerable interest, especially with regard to its functions as regulator of circadian clocks, antioxidant, immune modulator, neuroprotectant, and agent influencing metabolic control mechanisms, with consequences to metabolic syndrome, insulin resistance and obesity. However, the effects of melatonin are not as uniform as it may appear on the basis of countless reports that emphasize the protective potential of this molecule. In fact, quite a number of actions can be entirely opposite, depending on experimental systems and conditions. Sometimes, these contrasting effects have been perceived as being highly controversial. However, this is not necessarily the case if the causes of the differences, conditions and deviating properties of experimental systems are appropriately considered. This review aims to outline these causes and to resolve discrepancies that seem to exist at first glance. As an immune modulator, melatonin can behave either as an anti- or pro-inflammatory agent, depending on the type of challenge. Since inflammation causes enhanced formation of reactive oxygen and nitrogen species, melatonin may also behave in a pro-oxidant fashion, contrary to otherwise observed anti-oxidant actions. In nocturnally and diurnally active mammals, melatonin is either positively or negatively associated with exercise, wakefulness and food intake. Therefore, many findings in laboratory rodents cannot be translated to humans, especially concerning sleep and glucose tolerance. Tumor and non-tumor cells can respond in opposite ways to melatonin, in a pro-apoptotic or anti-apoptotic way, in down-regulation (in cancer cells) or up-regulation (in aging tissues) of oscillator components such as CLOCK or sirtuin 1 (SIRT1). Dysregulated oscillators in which clock genes with anti-tumor properties such as Per2 are epigenetically silenced behave differently than clocks that are fully operating or only moderately flattened in their amplitudes because of aging.