Fire-history data for ponderosa pine forests in the western U.S. have uncertainties and biases. Targeting multiple-scarred trees and using recorder trees when sampling for fire history may lead to incomplete records. For most of the western U.S., research is insufficient to conclude that high-severity fires did or did not occur in these forests prior to EuroAmerican settlement, because the needed data are not commonly collected. The composite fire interval is shown here to be misleading, but this can be remedied in part with interval estimates by fire size class. These problems mean that an assumption-that high surface-fire frequencies will restore and maintain the structure of these forests-lacks a foundation in reliable fire-history research.