historybriefs
all subjects →
13:00in productionCh. 1 · 5:12 in the morning, along three hundred miles of fault/ 13:00 · ceiling 15 min
Crime & disaster · 20th century

1906 San Francisco earthquake

1906

San Francisco's official 1906 death toll sat at a few hundred for nearly a century before the city admitted, in 2005, that it had simply never counted most of the dead in Chinatown.

The earthquake that struck San Francisco at 5:12 on the morning of 18 April 1906 registered somewhere between magnitude 7.7 and 8.3 and ruptured close to three hundred miles of the San Andreas Fault, but most of the city's destruction came not from the shaking itself, which is estimated to account for as little as five to twenty percent of the damage, but from the fires that followed and destroyed some twenty-five thousand buildings. Early death tolls of a few hundred left out most of the dead in Chinatown, and San Francisco's Board of Supervisors did not formally acknowledge a toll above three thousand until 2005. The disaster also drove the geologist Andrew Lawson's foundational research connecting the earthquake directly to the fault he had already identified in 1895.

Chapters & takeaways6
  1. 0:08
    5:12 in the morning, along three hundred miles of fault

    The earthquake struck at 5:12 a.m. on 18 April 1906, rupturing roughly 296 miles of the San Andreas Fault and shaking a region from Oregon to Los Angeles.

  2. 2:10
    Fire, not the shaking, did most of the damage

    Officials estimated that fires following the quake, rather than the shaking itself, caused between eighty and ninety-five percent of the total destruction across 490 city blocks.

  3. 4:20
    A death toll that left out Chinatown

    Early counts of 375 to 500 deaths excluded hundreds of fatalities in Chinatown, and the true toll remained officially unresolved for decades.

  4. 6:30
    Soldiers, relief camps, and two-dollar rent

    General Frederick Funston mobilised federal troops and the Army built 5,610 relief houses across eleven camps, housing twenty thousand displaced residents at two dollars a month.

  5. 8:40
    A fault named a decade before it broke

    Andrew Lawson had already identified the San Andreas Fault in 1895, and his 1908 report mapping the 1906 rupture proved the fault was what had caused the earthquake.

  6. 10:50
    A formal correction, ninety-nine years later

    San Francisco's Board of Supervisors officially recognised a death toll above three thousand in 2005, acknowledging that marginalised communities had been undercounted from the start.

Worth your time?

Yes. Study the whole thing.

4.5/ 5
What works
  • the 2005 official correction to the death toll is treated as a real historical event in its own right, not a footnote
  • the fire-versus-shaking damage split gives a genuinely counterintuitive fact its full weight
What does not
  • the range for the earthquake's magnitude, 7.7 to 8.3, is left as a range rather than narrowed for the reader
Study it if
  • readers interested in how disaster death tolls get revised decades later
  • anyone who wants the seismology history connected directly to the disaster that produced it
Skip it if
  • readers wanting primarily a rebuilding and urban-planning account of the recovery years
The written brief4 min read

5:12 in the morning, along three hundred miles of fault

The earthquake struck at 5:12 in the morning on 18 April 1906, preceded by a foreshock some twenty to twenty-five seconds ahead of the main shock. Its magnitude has been estimated at 7.9 on the moment magnitude scale, though the range of estimates runs from 7.7 to 8.3 depending on the method used, and the rupture propagated along roughly 296 miles of the San Andreas Fault. The shaking itself was felt across a remarkably wide area, from Oregon in the north to Los Angeles in the south and as far inland as central Nevada, giving contemporaries across a large stretch of the American West direct experience of an earthquake whose epicentre lay specifically beneath San Francisco and the fault running through it.

Fire, not the shaking, did most of the damage

Contrary to the popular image of an earthquake-flattened city, most of San Francisco’s physical destruction came from fire rather than ground shaking. Authorities estimated that fires accounted for somewhere between eighty and ninety-five percent of the total destruction, with more than thirty separate fires eventually destroying approximately 25,000 buildings across 490 city blocks. The financial cost of this destruction was estimated at 350 million dollars in 1906 terms, a figure equivalent to roughly 9.1 billion dollars in 2024 money. This distinction between earthquake damage and fire damage mattered practically as well as historically, since it meant that firefighting capacity and water supply, both severely compromised by the earthquake’s damage to pipes and infrastructure, ended up determining the disaster’s ultimate scale more than the initial shaking did.

A death toll that left out Chinatown

The earthquake’s human toll proved far harder to pin down accurately than its physical damage. Early estimates placed deaths at somewhere between 375 and just over 500, but these figures excluded hundreds of fatalities in Chinatown that went unrecorded at the time, a gap reflecting how little institutional attention marginalised communities received amid the broader disaster response. Later estimates for total deaths ranged as high as 3,000 or more, but the toll remained officially uncertain for decades. It was not until 2005 that San Francisco’s Board of Supervisors formally recognised a death toll exceeding 3,000, an official acknowledgement, nearly a century after the event, that earlier casualty figures had systematically undercounted the disaster’s actual human cost among the city’s less privileged residents.

Soldiers, relief camps, and two-dollar rent

The immediate response drew on federal military resources stationed nearby. General Frederick Funston mobilised troops from Angel Island, and more than 4,000 soldiers eventually assisted with firefighting and relief operations across the devastated city. The Army constructed 5,610 relief houses spread across eleven camps, providing shelter for some 20,000 displaced residents at a nominal rent of two dollars a month, a scale of organised emergency housing that reflected both the severity of the displacement and the resources the federal government was willing to commit to the recovery. Reconstruction of the city proceeded steadily over the following years and was largely complete by 1915, a milestone marked publicly by the Panama-Pacific International Exposition, staged explicitly to celebrate San Francisco’s recovery from the disaster.

A fault named a decade before it broke

The earthquake also produced a lasting scientific legacy. Andrew Lawson, a professor at the University of California, Berkeley, had already identified the San Andreas Fault in 1895, well before the 1906 earthquake occurred. Following the disaster, Lawson surveyed and mapped surface ruptures across the affected region, and when he plotted these locations, they formed a line closely matching the fault he had already mapped over a decade earlier, allowing him to conclude with confidence that this specific fault had caused the earthquake. His subsequent 1908 report went further, demonstrating a clear correlation between the intensity of shaking experienced in different areas and the underlying local geology, a finding that became foundational to how seismic risk is assessed and zoned in earthquake-prone regions today.

A formal correction, ninety-nine years later

Later research built on Lawson’s initial identification of the fault’s significance. By 1953, geologist Thomas Dibblee’s work had demonstrated that hundreds of miles of lateral movement had occurred along the San Andreas Fault over geological time, evidence for large-scale horizontal plate motion that was considered a radical proposition when first advanced but has since been fully supported by the development of modern plate tectonic theory. Taken together, the 1906 disaster occupies an unusual place in the history of both American cities and earth science: a catastrophe whose true human cost was only formally acknowledged a century later, and whose physical evidence directly enabled scientific advances that reshaped how earthquakes themselves are understood.

Same strand · Crime & disaster4 of 53
Up next in History

1989 Loma Prieta earthquake

1989 · 13:00

A World Series broadcast accidentally saved lives during the 1989 Loma Prieta earthquake, since fans leaving work early for the game emptied a freeway that on an ordinary rush hour carried 195,000 vehicles a day past a viaduct that collapsed minutes later.

13:00