Why dusk is the sensitive moment
The SCN is the brain's master daily pacemaker, and melatonin is a hormone the pineal gland releases on a nightly schedule. The clock's sensitivity to melatonin is not constant across the day: it is phase-dependent, meaning the same dose can push the clock earlier at one circadian time and do little or nothing at another [3]. In rat SCN tissue kept alive in a dish, melatonin applied at the right circadian time advanced the phase of the clock's own electrical rhythm — direct evidence that melatonin can reset the pacemaker itself, without any other body signals involved [2].
The reason dusk stands out is that this is when endogenous melatonin is rising. One account holds that during a standard 12-hour light/12-hour dark laboratory cycle, the clock's melatonin sensitivity would be minimal during much of the day, so the hormone's resetting action is concentrated where it overlaps with the natural evening rise [3]. That framing treats melatonin as a feedback signal: the nightly surge tells the clock where it is in the cycle and nudges it accordingly.
Resetting is not the only thing melatonin does to the clock
Melatonin has at least two separable actions on the SCN: an acute inhibition of neuronal firing, and a slower phase-shifting effect on the clock's timing [4]. These are distinct processes mediated by melatonin receptors in the SCN, and keeping them apart matters for interpreting timing experiments — a drop in firing is not the same as a change in phase [4].
The phase-shifting action is what the question is about, and it has been traced to specific receptor chemistry. In cultured cells carrying circadian reporter genes, melatonin lengthened the period of the BMAL1 and PER2 rhythms and delayed circadian phase, while an MT1 inverse agonist — a molecule that pushes the same receptor the opposite way — advanced phase [5]. That the two compounds moved the clock in opposite directions supports the idea that they act through the same receptor system, though this is a mechanistic interpretation rather than a proven purpose.
Do dish and whole-animal results line up?
On the core point they do. The in vitro SCN work established that melatonin directly resets the clock at a restricted, phase-dependent window [2][3], and the cell-based reporter work reports phase shifts from melatonin that corroborate in vivo data, with melatonin delaying phase in that system [5]. So both laboratory preparations and living-animal observations point to melatonin as a genuine phase-setting signal rather than an artifact of one setup.
The agreement has limits worth stating plainly. The dish experiments isolate the SCN, while living animals add the rest of the body, and the comparison in the cell work is with in vivo findings rather than a matched side-by-side experiment [5]. Timing also differs by preparation: the SCN recordings describe an advance at a sensitive circadian time [2], whereas the reporter-cell study describes a delay [5]. Those are not necessarily contradictory — phase response commonly depends on when the signal arrives and on which tissue is being measured — but they mean "lab and live agree" should be read as agreement on the existence and phase-dependence of resetting, not on an identical shift at an identical hour.
The clock is not the only place melatonin acts
One reason the SCN story is incomplete is that melatonin receptors are expressed outside the brain. In the reporter-cell experiments, cellular circadian rhythms responded to MT1-targeting molecules even though those cells were not dependent on the SCN, and the authors conclude that melatonin's direct effects likely extend beyond the SCN [5]. This is a caution about attributing every whole-animal phase shift to the suprachiasmatic clock.
Developmental timing adds another layer. Melatonin given during the fetal stage affected circadian clock gene expression in the SCN but not in the liver, indicating that the same hormone can act on the brain pacemaker and spare at least one peripheral clock at that life stage [1]. Together these findings suggest the SCN is a key but not exclusive target, and that which clock responds depends on when melatonin is present and which tissue is examined [1][5].
Sources used in this answer
Melatonin administered during the fetal stage affects circadian clock in the suprachiasmatic nucleus but not in the liver
Melatonin administered during fetal development affected circadian clock gene expression in the rat suprachiasmatic nucleus but not in the liver.
Melatonin directly resets the rat suprachiasmatic circadian clock in vitro
Melatonin directly reset the rat suprachiasmatic clock in vitro, producing phase advances in SCN electrical activity at specific circadian times.
Circadian actions of melatonin at the suprachiasmatic nucleus
The SCN's sensitivity to melatonin is phase-dependent, and under a 12-hour light/12-hour dark laboratory cycle that sensitivity would be minimal during much of the day.
Molecular dissection of two distinct actions of melatonin on the suprachiasmatic circadian clock
Melatonin has two distinct actions on the SCN clock — acute inhibition of firing and phase-shifting — both mediated by SCN melatonin receptors.
Circadian Effects of Melatonin Receptor-Targeting Molecules In Vitro
In U2OS circadian reporter cells, melatonin lengthened BMAL1 and PER2 periods and delayed phase while an MT1 inverse agonist advanced phase, and the effects did not depend on the SCN.