Archaeologists studying shell assemblages from prehistoric sites along the Pacific coast of North America have been interested in the influences of collecting intensity and environmental variability on California mussel (Mytilus californianus) size. To determine the variation in mussel size within a shell assemblage, researchers have developed a variety of proxies of mussel valve length based on measurements of morphological features occurring at or near the valves umbo. We propose four additional measurements that can serve as proxies and evaluate their correlation with valve length using regression analysis. Of the four, anterior adductor scar length has the strongest correlation, and we present two examples of its application. We also evaluate a popular visual technique based on a set of outline drawings of valves of varying lengths, and we found that it systematically underestimated valve length but could be useful under certain circumstances. We conclude that the selection of a particular proxy of mussel valve length depends on the nature of the mussel shell assemblage being studied and the research context.