The hypothesis that early formed d¿collement folds localize thrust ramps, explaining the regular spacing of major thrust ramps observed in many mountain belts, is examined. The model for folding instability consists of a rigid basement, a weak d¿collement zone, and a stiff thrust sheet. The model incorporates the effects of topographic stresses, and syntectonic erosion and redeposition and uses a linearized power law rheology for the thrust sheet. This approximation is valid for small angles of fold limb dip (<15¿) and is appropriate for analysis of the wavelength selection of folds at the onset of deformation. The spacing of major thrust ramps and the stratigraphic thicknesses of the d¿collement zone and thrust sheet from the Wyoming portion of the Sevier fold-and-thrust belt are used to test the model. The ratio of the fold dominant wavelength to thrust sheet thickness, Ld/h2, increases with the ratio, R, of the viscosity of the stiff layer to that of the d¿collement zone and decreases with the power law stress exponent, n2. The fold amplification factor increases with both n2 and R. As the topographic decay ratio, D, increases, Ld/h2 increases at fixed n2 and R. Acceptable models for the Absaroka and Darby thrust sheets have n2≥3. Only when R<1000 and significant erosion and redeposition of sediments accompany the folding of the thrust sheet (D≥103) does the required R fall below a reasonable limit. Thus a model in which d¿collement folding is accompanied by syntectonic erosion and redeposition precedes thrusting is consistent with the available observational constraints. ¿ American Geophysical Union 1996