Dissapearing solar filaments have long been suspected as an indicator of terrestrial magnetic disturbances. However, because filament disappearances are a common solar event and because they failed as a candidate source for M region (recurrent) magnetic disturbances, their potential utility as a forecasting aid for geomagnetic storms has largely been neglected. A search for possible solar sources of geomagnetic storms from June 1976 through June 1979 has revealed that a significant number of the storms, including the two largest, can only be associated with filament disappearances. This result is supported by the many recent papers studying SKYLAB and other observations of coronal transients which always find a strong correlation between those transients and eruptive prominences. By analyzing the physical characteristics of those disapearances which procede magnetic storms and those which do not, some tentative guidelines for forecasting geomagnetic disturbances have been developed based on evidence of a significant restructuring of the implied coronal magnetic field which could release solar wind plasma favorably positioned to impact the earth.