The effect of sizing mixtures on the slubby portions of a staple yarn was determined indirectly by the application of the size to a cotton roving and subsequent testing of the roving for strength, ultimate elongation, and stiffness. In contrast to the methods which employ yarn as the base material for the application of the size, large differences between sizing mixtures were found by this procedure. The evaluation method which was investigated proved useful for the screening of sizes for knitting yarns in which the failure to give good knitting quality was caused by the presence of slubs.