The evening wind-down

The night is mostly decided before it starts. This is the 60 to 90 minutes that does the deciding, reduced to the three levers with evidence behind them.

01Why the last hour decides the night

Falling asleep is not a switch, it is a handover. The circadian alerting signal has to fall, melatonin has to rise, and the body has to start losing heat through the skin of the hands and feet. That distal warming, the sign of heat leaving the core, is one of the best physiological predictors of how quickly sleep comes.¹ Everything in this protocol serves one of those transitions. Nothing in it is a ritual for its own sake.

02Light down

Evening light is the strongest brake on the handover. In a controlled crossover study, reading on a bright screen before bed suppressed melatonin by around half, delayed the circadian clock by about an hour and a half, lengthened the time to fall asleep and cut next morning alertness, all against the same book on paper.² Sensitivity varies enormously: in one experiment the evening light needed to halve melatonin differed more than fifty fold between people, and for the median person ordinary room light was enough.³ You do not know which end of that range you are on, so the safe assumption is the same for everyone: in the last hour, dim the room, warm the light, keep screens small and dim or off.

03Heat out

Warming the skin an hour or two before bed sounds like it should delay sleep, and does the opposite. A meta-analysis of 13 studies found that a warm shower or bath at 40 to 42.5 degrees, around ten minutes, taken one to two hours before bed, shortened the time to fall asleep by roughly nine minutes and improved rated sleep quality.⁴ The mechanism is the handover again: warming the periphery opens circulation in the skin, which accelerates the core heat loss the body was already trying to achieve.¹ Timing matters. The same shower taken directly before bed does less.

04Mind parked

The most common thing that keeps a demanding day awake is the day itself. In a polysomnography study, people who spent five minutes before bed writing tomorrow's to-do list fell asleep about nine minutes faster than people who wrote about tasks already completed, and the more specific the list, the faster the onset.⁵ Writing the day down is not journalling for its own sake. It is offloading open loops so the arousal system can stand down.

05What is not a wind-down

06The research horizon

The evening literature has been busy putting numbers on things this protocol treats as levers. Screens: a 2025 meta-analysis covering 548,338 adults quantified the cost at roughly 13 minutes of later bedtime per hour of evening screen use, with a tripled odds of trouble falling asleep among heavy users.⁷ Thirteen minutes an hour sounds survivable until you count the hours. Temperature: a study logging around 11,000 nights in real bedrooms found sleep efficiency stable between 20 and 25 degrees and dropping 5 to 10 percent as rooms warmed past that, with large individual variation, which is why the protocol says experiment rather than prescribing one number.⁸ And one honest null: a 2025 randomised trial of pre-sleep worry-postponement cut daily worry by about 15 minutes but did not measurably improve sleep itself.⁹ Less worry is its own reward, so the practice stays; the sleep claim waits for better data. That is how this Library works: the numbers move, the page moves with them.

07Where LAYER THREE sits

The evening layer is taken 60 to 90 minutes before sleep, which is not an accident: it is the same window this protocol occupies. Magnesium bisglycinate, ashwagandha, L-theanine, valerian and lemon balm are the formulation's side of the downshift, five ingredients working the same handover from the inside. The light, the heat and the list are yours and carry the most weight. One window, everything pushing the same direction.

How we handle evidence. Every lever in this protocol carries its primary source and grade. The formulation is described alongside the behaviour, not in place of it. Educational content, not medical advice.