Starch is a nutritionally important carbohydrate in feeds that is increasingly measured and used for formulation of animal diets. Discontinued production of the enzyme Rhozyme-S required for AOAC Method 920.40 invalidated this method for starch in animal feeds. The objective of this study was to compare methods for the determination of starch as potential candidates as a replacement method and for an AOAC collaborative study. Many starch methods are available, but they vary in accuracy, replicability, and ease of use. After assays were evaluated that differed in gelatinization method, number of reagents, and sample handling, and after assays with known methodological defects were excluded, 3 enzymatic-colorimetric assays were selected for comparison. The assays all used 2-stage, heat-stable, a-amylase and amyloglucosidase hydrolyses, but they differed in the gelatinization solution (heating in water, 3-(N-morpholino) propanesulfonic acid buffer, or acetate buffer). The measured values included both starch and maltooligosaccharides. The acetate buffer-only method was performed in sealable vessels with dilution by weight; it gave greater starch values (2-6 percentage units of sample dry matter) in the analysis of feed/food substrates than did the other methods. This method is a viable candidate for a collaborative study.