When Alfred Wegener proposed in 1912 that the continents had once formed a single landmass before drifting apart, he supported the claim with evidence that geologists found individually familiar but collectively startling: matching fossil species stranded on continents now separated by oceans, coastlines that seemed to interlock like puzzle pieces, and rock strata that lined up across the Atlantic as though sliced by one blade. Wegener, trained as a meteorologist rather than a geologist, argued that drift explained these patterns more economically than the reigning theory of sunken land bridges, which required entire tracts of continental crust to rise and vanish beneath the sea without leaving any trace of their submersion.

Yet the geological establishment greeted the hypothesis with something closer to derision than debate. Critics did not merely doubt Wegener's evidence; many dismissed outright the possibility that solid continental rock could plow through solid oceanic rock, since Wegener could not specify a mechanism capable of moving landmasses across the globe. A symposium convened by the American Philosophical Society in 1926 was organized largely to repudiate the idea, and several prominent geologists characterized it as the sort of theory that appeals to outsiders precisely because it skips over the technical difficulties insiders would have caught. Wegener spent his remaining years gathering further corroborating evidence, from glacial striations to matching mountain ranges, without ever persuading the discipline's leading figures, who remained committed to a stationary earth until well after his death in 1930.

Only in the 1960s, once the discovery of mid-ocean ridges and magnetic striping on the seafloor supplied the missing mechanism, did plate tectonics rehabilitate what had been Wegener's outsider's intuition. The belated vindication suggests that the geologists who rejected continental drift were not simply mistaken about the evidence Wegener assembled; they were, in a narrower sense, right to insist that a hypothesis without a mechanism remains scientifically incomplete, even when the underlying observation eventually proves correct.