Red light and dark adaptation

Full dark adaptation takes twenty to forty minutes and one glance at a white phone screen resets most of it. Use deep red light, as dim as you can work with.

Two mechanisms are at work, on very different timescales. Your pupil opens within seconds, which is the part people notice. Underneath it, the light-sensitive pigment in your rod cells regenerates far more slowly, and that is the part that matters for faint detail.

The vision literature puts numbers on it. Cones adapt in roughly five to seven minutes but are not much use in very dim light. At around ten minutes there is a "rod–cone break" where the rods take over, and peak sensitivity arrives somewhere between twenty and thirty minutes. Rhodopsin rebuilds to about half in five minutes but needs close to half an hour to saturate. Bright light bleaches it essentially instantly, which is why one careless glance costs you so much.

Red light works because rhodopsin is relatively insensitive to long wavelengths, above roughly 650 nm. That lets your cones do the reading while the rods stay dark-adapted, which is why observatories, aircrews and submariners all standardised on it.

Brightness matters as much as colour. A bright red torch is still bright. The goal is the dimmest light that lets you do the task, which is usually far dimmer than people expect.

Phones are the usual culprit at a dark site — a white screen at full brightness undoes half an hour of waiting in a second. This site has a red-light mode in the header for exactly that reason: it renders the whole page in a single red channel on black, emits no blue or white at all, and switches instantly rather than fading, because even a brief bright frame during a transition would defeat the point.