Fuel treatments are now mandated by federal policy to reduce hazardous fuels on federal forest lands. More information is needed on the effectiveness of fuel treatments, especially in mixed-severity fire regimes. In this study I had the rare opportunity to quantify the relationship between fuel structure and fire severity using pre-fire surface and canopy fuels data and fire severity data after an intense wildfire. The study area is in a mixed-severity fire regime, the mixed-evergreen forest of southwestern Oregon. Thinning reduced canopy fuels, decreasing the potential for crown fire spread, but the presence of activity fuels increased potential surface fire intensity, so increases in canopy base height did not decrease the potential for crown fire initiation. Thinning followed by under-burning reduced canopy fuels and surface fuels, thereby decreasing both crown fire spread potential and the potential for crown fire initiation. However, crown fire is not a prerequisite for high fire severity; damage and mortality of overstory trees in the wildfire was extensive despite the absence of crown fire, and the low predicted crown fire potential before and after the fuel treatment. Damage to and mortality of overstory trees were most severe in thinned treatments (80 - 100% mortality), least severe in the thinned and under-burned treatment (5% mortality), and moderate in untreated stands (53-54% mortality) following a wildfire in 2002. Fine fuel loading was the only fuel structure variable significantly correlated with crown scorch of overstory trees. Percentage crown scorch was the best predictor of mortality 2 years post-fire. Efforts to reduce canopy fuels through thinning treatments may be rendered ineffective if not accompanied by adequate reduction in surface fuels.