Excerpts from the report: The falling number test was developed by Hagberg as a simple, quick index for the alpha-amylase activity of grain. Falling number is the time, in seconds, required to stir and allow the viscometer-stirrer to fall a fixed distance through a hot aqueous flour suspension being liquefied by the enzyme. Perten has shown the relationship between falling number, amylograph Brabender units, diastatic activity, and other measures of alpha-amylase activity. The purpose of this study was to locate all significant sources of error in the falling number test and to establish controls for them. The number of samples used to investigate a source of error was minimized because of the amount of work required to examine all possible factors which might influence the falling number results. An effort was made to choose samples covering a broad range of falling number values for each experiment but at the same time to select them at random. In some experiments the difference between the lowest and highest mean value approached the least difference for significance. If such experiments were repeated with a much larger number of samples, a significant difference might be shown for this particular source of error. Since most of the experiments were carried out using three types of commercial flours--bread, all-purpose, and cake--the conclusions drawn are especially applicable to them. Work done with wheat was for the purpose of observing the effect of granularity resulting from the method of milling the wheat.