Abstract
General principles for the selection of samples in the calibration of near-infrared (NIR) spectrophotometers are discussed. The discussion ends with suggestions of alternative strategies, which are illustrated by examples.