The present research was designed to evaluate an ultrasonic technique for measuring the drawability of low carbon steel sheets nondestructively. This technique, which uses electro magnetic acoustic transducers (EMAT's) is embodied in a fully automatic ultrasonic device called Ultra-Form, which was developed by Iowa State University in 1990. EMAT transducers generate the ultrasonic pulses in the sheet (without any need for a coupling medium) whose velocities are then related to the microstructural factors (crystallographic texture) that control drawability in rolled sheets. Drawability (usually interpreted by measuring r-values of low carbon steels) is an important property indicative of the formability of steel sheets used in the auto and appliance industries. It is measured by the destructive tension test which is sensitive to the experimental errors and is time consuming. From a quality control perspective, there is a considerable industrial interest in further development of nondestructive on-line measurements of r-values during production (at the steel mill, or during the stamping operations). Towards this end, a collaborative research between LTV Steel Company, National Steel Corporation, Ford Motor Company and Ames Laboratory was conducted. This included theoretical and experimental research on a total of 260 steel samples cut from 26 lifts (~10-12 samples of 20 in. by 20 in. (50 by 50 cm) per lift) of 5 types of commercially produced steels of Interstitially Free (IF), Aluminum-Killed Drawing Quality (AKDQ), Bake Hardenable (BH), High Strength Low Alloy (HSLA) and Phosphorus High Strength (HS) grades. These investigations also included tensile measurements on the steels (one per lift) and neutron diffraction (one per lift for 10 selected lifts). This research evaluated the ultrasonic technique with respect to its accuracy and reproducibility, sensitivity to the microstructure of the steel sheets (texture), and the factors that may affect the relative value of its predictions of plastic properties. The experimental correlations and the repeatability data (obtained from sequential ultrasonic measurements carried out on the nominally identical samples, 10-12 per lift) point to an excellent sensitivity of ultrasonic waves to the texture of the steels, and a sound electromagnetic coupling between the EMAT's and the sheets, which is important for an on-line system. The data also indicate that the most important factor affecting the ultrasonic r-value predictions is the accuracy of the single crystal elastic constants (which should be representative of the chemical composition of the steels under study). The choice of modeling (in prediction of plastic properties) also plays an important role. Here the relaxed constraint model provided a significant improvement over the empirical correlations of Mould-Johnson. The diffraction data indicate that the 6th order coefficients (mostly W[subscript]600 and W[subscript]620) improve predictions of r-values (mostly for the IF steels) and that they should be included in the analysis for future use.