Recent theoretical, experimental, and field studies conducted by the Bureau of Mines suggest that rock failures may be predicted and, in some instances, controlled. Theoretical studies that were later corroborated by laboratory and field studies predict that a seismic anomaly will precede rock bursts and/or roof falls in mines. The seismic anomaly is predicted to consist of an increase in the number of seismic events occurring in the immediate vicinity where the failure is nucleated. This increase will then be followed by a dramatic decrease of seismicity in a broad region that surrounds the location of the impending failure. The dimensions of this region allow a preliminary estimate of the time to failure (measured from the time the seismicity increase began) as well as an estimate of the energy that will be released at the instant of failure. Field observations indicating that this predicted behavior does precede rock bursts and a roof fall are presented in this report. Implications of results to the problem of failure control in mines that experience massive rock failures are discussed.