Radical chemistry at a forested site in northern Michigan has been examined using measurements made by a radical amplifier system in the summer of 1997. The radicals show a diurnal variation with the maximum concentration being measured several hours after solar noon. Low but observable concentrations were present at night, but these levels are inconsistent with the previously postulated active nighttime chemistry driven by high concentrations of highly reactive hydrocarbons. A box model, constrained by measured concentrations of CO, O3, NO, NO2, isoprene, and peroxyacetylnitrate, was found to simulate both the maximum concentrations and their diurnal variation. Model sensitivity studies indicate that 54% of the radical production comes from sources other than ozone photolysis, and the presence of isoprene has an impact on the radical concentration through impacts on both radical production and loss.