In early September 1975, two clearcuts (14 and 17 acres; 5.7 and 6.9 ha), two sets of 4 small clearcuts (1.5 acres; 0.6 ha each), and one shelterwood cutting (22 acres; 8.9 ha) were broadcast burned principally for seedbed preparation and fuel reduction on the Coram Experimental Forest. The objective was to develop a model for predicting duff reduction (amount and percent) and mineral soil exposure (percent) resulting from the fires, using only preburn measurements for inputs. In addition, the influence of the fires on soil heating, soil water content, and nonconiferous root mortality was quantified. Because of the relatively low maximum air temperatures and high precipitation on the study area during August, the fuels were relatively moist at the time of ignition. As a result, the fires generally spread poorly, especially in the shelterwood, except where the fuels were concentrated. The prescribed fires reduced the duff an average of 27 percent in the clearcuts and 11 percent in the shelterwood. Mineral soil exposure averaged 19 percent in the clearcuts and l0 percent in the shelterwood. The amount of water within the upper 4 inches (10. 2 cm) of soil was not changed greatly, there was no significant heating of the surface soil above 113°F (45°C), and root mortality was low. Multivariate regression analysis was used to derive prediction equations for duff reduction and mineral soil exposure. However, because of extremely poor correlation between the dependent variables, duff reduction and mineral soil exposure, and all combinations of the independent variables, no statistically valid equations resulted. The expected correlation between duff consumption and lower duff water content was not evidenced. This was attributed to the high water content of the duff itself, which averaged nearly 50 percent in the three blocks studied. These results substantiate Shearer's (1975) conclusions. Broadcast burning for seedbed preparation should not be attempted in this forest type when lower duff water contents exceed about 100 percent.