A new methodol. for the estn. of drop size distributions in dil. polydisperse polymer blends is proposed. The procedure exploits the capabilities of large amplitude oscillatory shear (LAOS) flows to discriminate the behavior of drops of different size. The methodol. is tested on synthetic data obtained by mimicking rheol. responses of polydisperse immiscible polymer blends. The procedure allows a quant. description of the original distribution, and is also robust to exptl. noise