Abstract
The existing experimental evidence for resonances in heavy-ion reactions is interpreted in a unified way, based on the molecular resonance-window concept. Predictions are given for nuclear systems most likely to exhibit resonances.