It remains to be seen whether any of it will matter when the first jolt of a real seismic wave knocks Southern Californians off their feet; nothing gets the public's attention like an earthquake. In that way, nature handed the ShakeOut planners a gift: At 11:42 a.m. on July 29, 2008, a previously undetected fault sprang loose perilously close to downtown Los Angeles, producing a 5.4 magnitude earthquake that knocked cans off grocery-store shelves and set wood-frame bungalows swaying like boats. It was the largest quake in the Los Angeles Basin since 1994, when a 6.7 on the Northridge fault claimed 60 lives, collapsed two freeways and caused $40 billion in damage. Suddenly, Lucy Jones, the USGS chief scientist who became famous when she helped calm Southern Californians during a spate of earthquakes in the '90s, was back on the radio. "Some people have forgotten they live in earthquake country," observed Jones, the architect of the Multi-Hazards Demonstration Project. A drill may be the only way to remind them.

For the purposes of the ShakeOut, Hudnut and his colleagues mapped the hypothetical earthquake on the segment of the San Andreas that stretches from the Salton Sea to Lake Hughes north of Los Angeles. No part of this segment has erupted since 1857; some stretch of it slipped in 1812, but nobody knows how much or exactly where. The San Andreas' far southeastern extreme, the segment Fairbanks couldn't follow, has been quiet since about 1680.

Geologists still don't know what to make of that stillness. Faults come in different flavors and run on different schedules; some "thrust" faults only push out of the ground every 1,000 years; some strike-slip faults, where the ground moves laterally on either side of the fault, shift a little bit every 20 years. By digging deep into the ground and looking for signs of past quakes, geologists have recently uncovered a possible pattern in the southeastern strand of the San Andreas. From 800 to 1680, it broke on an average of every 150 years. So what does it mean that it hasn't moved for the last three centuries?

One of SoSAFE's goals is to find out. "A geologist once described it as being 10 months pregnant," Jones says. "But I think it's more like being 20 months pregnant. At which point you know you're not dealing with a normal pregnancy."

Sixty years after Lawson completed his 448-page report on the San Andreas Fault, international networks of seismometers, installed to register underground nuclear tests, unexpectedly revealed the contours of the long seams of tectonic plates that form the earth's crust. The plates form continents and thrust underneath seas; their natural instability roils up in volcanic ridges and triggers tsunamis. Where they slide up next to each other, the ground springs with thousands of cracks. 

The San Andreas Fault marks the slippery boundary between the Pacific and North American plates. Where the fault is relatively straight, it simply creeps: the plates inch in opposite directions a little at a time, and no one can feel what's happening. Where the edges of the plates form more jagged boundaries, the fault gets stuck, bends and builds up strain. That rock eventually snaps back into place, sending waves of seismic energy reverberating through the earth.

Some segments of the San Andreas build up strain for a century or two before exploding in magnitude sevens. Other segments, such as the geologically famous stretch that runs through the town of Parkfield in Central California, bend and spring back in fives or sixes every 20 odd years. Some have so far eluded geologist's understanding.

"California only has written records dating back to 1769," says Glenn Biasi, a research seismologist at the University of Nevada who works on a team of scientists assembling a chronology of quakes on the southern San Andreas. "So we have to go through layers (of earth) to get estimates on the dates of older earthquakes. Then we try to figure out how regular they are." In other words, to compile a historical record of earthquakes, geologists have to dig.