Attention is drawn to the largest eruptions with dense rock equivalent (D.R.E.) volumes approx- imating 50 km3. These have been strongly clustered during the last 40 000 years, with 11 recognised in the period 35 000 to 22 000 years ago, and 10 in about the last 9000 years, but none in the period 22 000 to 9000 years ago. With the latest such eruption, at Tambora, only 175 years ago, it is likely that we remain in a period of increased activity. During the last 100 000 years, roughly the period during which modern Homo sapiens has existed as a species, there has only been one eruption an order of magnitude larger (>_ 500 km3 D.R.E.). This took place at Lake Toba, Sumatra, about 75 000 years ago, and is estimated to have deposited ash to a thickness of a centimetre or so in parts of Ethiopia, Somaliland, Saudi Arabia and Iran, and to a thickness of a few millimetres in the eastern Mediterranean. The average interval between global eruptions of this magnitude is probably about 40 000 to 50 000 years._x000D_
Insufficient work on the Quaternary history of volcanoes has been carried out to suggest where the next such eruption might occur. At Yellowstone, which has been well studied, the average interval between very large explosive eruptions is about 700 000_x000D_
120 000 years._x000D_
As a prerequisite for magnitude-frequency analysis of eruptions, and hazard and risk impact assessments, it is important to classify past eruptions in a consistent manner. In an attempt to find a universally applicable method to estimate order of magnitude D.R.E. volumes from the distribution of airfall tephra, an empirical correction has been derived for the elongation of tephra dispersal fans due to wind. Dispersion distances for various thicknesses of tephra, recalculated for windless conditions, have been obtained from a study of about 40 tephra isopach maps for which D.R.E. volumes have been calculated, and a range of distances derived, equivalent to D.R.E. volumes from 106 m3