For both night and civil-twilight operations, § 107.29 puts the same equipment requirement on the aircraft: it must have lighted anti-collision lighting visible for at least 3 statute miles, with a flash rate sufficient to avoid a collision. Two specifics carry the weight. The light must be seeable from three statute miles — the same three-mile figure that sets the daytime flight-visibility minimum — and it must flash, at a rate fast enough to catch the eye and prevent a collision; a steady glow or a slow blink does not satisfy the rule. The logic is see-and-be-seen: a small drone is nearly invisible against a dark sky, and this light is how the pilot of a crewed aircraft picks it out in time to stay clear.

The rule gives the remote pilot in command exactly one piece of discretion, and it is narrow. You may reduce the intensity of the anti-collision lighting, but you may not extinguish it, if you determine that because of operating conditions it would be in the interest of safety to do so. The allowance exists for real reasons — a bright strobe near your control station can wash out your own night vision or bloom in the camera you are flying by — so the rule lets you dim it. What it never lets you do is switch it off, because the instant the light is dark the aircraft is invisible to exactly the traffic the light exists to warn.

Because this is an equipment requirement, it is settled before takeoff, not in the air: the aircraft either carries compliant anti-collision lighting or it does not fly after the light fades. And it applies across both after-light regimes — civil twilight and night alike — even though the pilot-training condition from the last section attaches only to night. The light is what lets others see you; the visibility and cloud minimums in the next section are what protect your own ability to see and avoid. In low light you need both halves working.