The solutions of Eady's 1949 model of baroclinic stability are extended numerically to include the non-geostrophic perturbations which wore not covered by the analysis in Part I. It is found that the largest growth rates are never associated with these new perturbations, so the tentative conclusions of Part I are verified. The more exact numerical solutions lead only to slight quantitative modifications of the results of Part I. If we let Ri be the Richardson number, then the largest growth rates are associated with “geostrophic” baroclinic instability if Ri>0.950; with symmetric instability if ¼<Ri<0.950; and with Kelvin-Helmholtz instability if 0<Ri<¼. Geostrophic baroclinic instability and symmetric instability can exist simultaneously if 0.84<Ri<1, and symmetric instability and Kelvin-Helmholtz instability can exist simultaneously if 0<Ri<¼