Satellite-line emission from OH molecules has been detected for the first time in dust clouds. Although the main lines have strengths characteristic of LTE, the satellite lines do not. The 1720- IHz transition is anomalously strong compared with the main-line emission, while the 1612-MHz transition is anomalously weak. This situation applies to all three positions so far observed in two different clouds. The anomalies are qualitatively of the type that can be explained by near- or far-infrared pumping of OH. The ratio of main lines, not perturbed on this theory, is used to derive column densities and excitation temperatures. These conditions are used to apply the near- and far-infrared theory, for which quantitative agreement with the observed satellite4ine anomalies seems possible in the future.