ABSTRACT A nonlinear model consisting of segmented cubic and linear functions was used to describe cumulative ash as a function of wheat's physical characteristics (wheat size, 1,000 kernel weight, protein content, whole wheat ash, pearling value, and Zeleny sedimentation). The protein content of the flour (i.e., protein recovery of the mill) was studied as a function of the whole wheat protein content. The empirical models were developed using data from 54 samples of hard red winter wheat milled in the pilot mill of the Department of Grain Science and Industry at Kansas State University, Manhattan, Kansas. The empirical models developed were implemented into an electronic spreadsheet capable of estimating milling yields, ash content, and protein content of straight grade, patent, and remaining clear flours from the ash curve.