The number of really big earthquakes in the 19th century is suspiciously lower than the number measured in the past century, raising the possibility that seismologists may be underestimating big earthquake risk based on historical data.