Tensile deformation characteristics of cast aluminum alloys have been investigated extensively. Cast Mg alloys have remained mostly neglected by researchers, despite their potential for weight savings. This present study is motivated by this gap in the literature and consists of two stages; in Stage 1, analysis of tensile data gathered from literature were reanalyzed, and in Stage 2, data generated from tensile testing of 60 specimens of AZ91 Mg alloy castings in both T4 and T6 conditions were analyzed to characterize work hardening behavior. In Stage 1, more than 1600 data were collected from the literature for various Mg alloy families. After plotting these data in yield strength-elongation charts, highest points were identified and interpreted as the maximum ductility, i.e., ductility potential (eFmax). The trend in maximum points indicated a linear relationship with yield strength (σY), expressed as; eF(max) = 41.8 - 0.106σY (1) This ductility potential equation can be used as a metric to compare elongation obtained from tensile specimens to measure the structural quality of Mg alloy castings. Moreover, results indicated that ductility potential was not affected by heat treatment, grain size (within 30-120 μm), casting geometry, size, the type of casting process nor chemical composition. In Phase 2, AZ91 cast Mg alloy specimens in T4 and T6 conditions were tested in tension to obtain stress-strain data for each specimen. Fits of four constitutive equations, namely, the Hollomon, Voce, Ludwik and Swift, to true stress-true plastic strain data in the elastoplastic region were characterized for the specimens with highest elongation values for T4 and T6 specimens. The coefficient of determination, R2, values for all equations were in excess of 0.99, suggesting that all four equations provide excellent fits to tensile data in both conditions. The change in work hardening rate with true stress was investigated for all specimens by using Kocks-Mecking (KM) plots. It was determined that work hardening behavior of Mg alloy castings in T4 and T6 is distinctly different. In T4 specimens, there is a plateau in work hardening rate at approximately E/25 which was observed in all specimens. The presence of this plateau is consistent with results given in the literature for pure Mg. However, this plateau was not observed in any of the T6 specimens. The reasons for the absence of the plateau in T6 specimens are unknown at this time. In both T4 and T6 specimens, the KM work hardening model in which work hardening rate changes linearly with true stress was found to be applicable. This is the first time that KM model was found to be valid for Mg alloys. Moreover in all specimens, there was a sudden drop in work hardening rate just prior to final fracture. This drop was first hypothesized to be due to structural defects in specimens, which was subsequently validated via fractography. Structural defects were found in all specimens whose fracture surfaces were investigated, indicating low to medium levels of quality. The quality index method, originally developed for cast aluminum alloys as the ratio of elongation to ductility potential, was found not to be applicable to Mg alloys, at least in its original form. This is due to the fact that work hardening behavior of cast aluminum alloys follows the KM model and there is no plateau where work hardening rate is constant. Hence the work hardening behavior of cast aluminum alloys and AZ91 specimens in T6 condition was similar. However the plateau of constant work hardening rate had a strong effect on elongation in T4 specimens. Therefore quality index analysis, which is supposed to be independent of alloy condition, did show that T4 and T6 specimens had different quality index levels. This finding contradicted the result from Stage 1 that aging has no effect on ductility potential. However because of the presence of structural defects in all specimens, quality index levels were low (0.30-0.45). Therefore it is unclear at this point whether the work hardening behavior of T4 and T6 specimens would still be different if elongation values were in the proximity of the ductility potential line. More research is needed to characterize work hardening behavior of cast Mg alloys in the absence of major structural defects and also address other questions raised in this study.