Data from volcano-monitoring studies constitute the only scientifically valid basis for short-term forecasts of a future eruption, or of possible changes during an ongoing eruption. Thus, in any effective hazards-mitigation program, a basic strategy in reducing volcano risk is the initiation or augmentation of volcano monitoring at historically active volcanoes and also at geologically young, but presently dormant, volcanoes with potential for reactivation. Beginning with the 1980s, substantial progress in volcano-monitoring techniques and networks - ground-based as well space-based - has been achieved. Although some geochemical monitoring techniques (e.g., remote measurement of volcanic gas emissions) are being increasingly applied and show considerable promise, seismic and geodetic methods to date remain the techniques of choice and are the most widely used. Availability of comprehensive volcano-monitoring data was a decisive factor in the successful scientific and governmental responses to the reawakening of Mount St. Helens (Washington, USA) in 1980 and, more recently, to the powerful explosive eruptions at Mount Pinatubo (Luzon, Philippines) in 1991. However, even with the ever-improving state-ofthe-art in volcano monitoring and predictive capability, the Mount St. Helens and Pinatubo case histories unfortunately still represent the exceptions, rather than the rule, in successfully forecasting the most likely outcome of volcano unrest.