The proposition that mid-plate ('hot spot') magmatism is strongly discouraged by a combination of thick lithosphere and rapid plate motion is examined empirically on a global basis and found to be inadequate. Although hot spots perhaps tend to 'avoid' cratonic areas, the nonrandom worldwide distribution of hot spots appears primarily to reflect the state of the sublithospheric mantle. Plate thickness or speed is of lesser importance. The high level of African mid-plate volcanism since 25 m.y. B.P. appears to be part of a global pattern controlled by events in the mantle rather than a result of the Africa plate coming to rest over sublithospheric thermal sources about 40 m.y. B.P.