North America at the height of glaciation, 20-25,000 years ago. So much water went into making the ice sheet that sea level dropped more than 250 feet. The solid line shows the resulting increased size of North America.

Eventually, the ice sheet melted away and the material it had scraped off was redeposited in confused heaps of sediment. In Agassiz's view, then, drift was basically glacial sawdust. Scratches in the bedrock were ascribed to bits of grit and rock caught in the moving ice as it gouged the land, a mechanism similar to what Lyell proposed in the iceberg theory. Erratics were large blocks broken off the bedrock that got caught up in the flow of the glacier and, when the ice melted, were deposited miles from their origin. (This idea was originally suggested to a friend of Agassiz by a Swiss peasant farmer, who had cooked up this crazy theory about why the boulders in his field looked exactly like the rock of a mountain some miles away.) Agassiz's ideas at first were met with disbelief and even hilarity. But glaciation explained more of the observations than any other notion that had yet (or has since) been proposed, and gradually the theories about drifting icebergs and the Noachian Deluge were abandoned.

Several new observations helped confirm the theory of continental glaciation in the minds of geologists. One was the discovery of permanent icecaps on Greenland and Antarctica, which provided an observable analog for Agassiz's ice sheets. Another was the recognition that the surface deposits in southern regions of Europe and North America were quite different from those to the north. In northern regions like Connecticut, drift sits directly on hard bedrock. But in the south, there is no blanket of drift. Instead, the surface materials slowly grade downward from soil to weathered bedrock (called saprolite) and finally to fresh, hard bedrock unaffected by weathering. Thick saprolite has been found in the north, but only in a few rare instances. The landscapes are also different: drumlins and streamlined hills are never found in the south. The dividing line between the northern region with drift, drumlins, and little saprolite and the southern region with thick saprolite, no drift, and no drumlins is a clear boundary reaching across almost the whole continent; roughly, it stretches from Cape Cod to Long Island, inland along the Ohio River to its confluence with the Mississippi, and then northwestward along the Missouri River to the Rocky Mountains. No convincing explanation could be found for why a supposedly global flood would affect the north so differently from the south. The only

125