Following the report of solar-like oscillations in the G0 V star η Boo (Kjeldsen et al. 1995), a first attempt to model the observed frequencies was made by Christensen-Dalsgaard et al. (1995). This attempt succeeded in reproducing the observed frequency separations Δv and δv02, although there remained a difference of ∼ 10 μHz between observed and computed frequencies. In those models, the near-surface region of the star was treated rather crudely: convection was described by means of a local mixing-length theory neglecting turbulent pressure, and the oscillations were assumed to be adiabatic. These approximations are likely to affect both Δv and absolute frequencies.