In recent volcanologic literature, the terms forecast and prediction have generally been considered synonyms. Wadge and Guest , however, in assessing the possibility that Mount Etna would erupt before May 1982, stated that “these are not predictions of specific events but general forecasts … based on the behavior of the volcano during the past seven years.” Lockwood et al.  used the term forecast in anticipating an eruption of Mauna Loa before the summer of 1978 on the basis of historical records. In contrast, Wood and Whitford-Stark  used the terms forecast and prediction synonymously when they anticipated an eruption of Krafla before the end of May 1982 by projecting records from 1975 to the end of 1981; in terms defined here, this statement was a forecast. The fact that all three of these forecasts proved incorrect indicates the relative uncertainty of simply projecting past records and it suggests the desirability of distinguishing, whenever possible, such general statements from more specific predictions based on repeated measurements of changing phenomena on a short time scale.