1. Cover as much of the range as practical:
You never know what may happen in another time frame.

Example:
Testing vision after dark adaptation. If you only kept people in the dark for 5 minutes and not 15, you would miss the switch from cones to rods.