Wheat contributes significantly to human nutrition and livelihoods around the world, but is highly susceptible to drought stress, which is expected to become more prevalent as the climate changes. Therefore, it is increasingly important to assess wheat genotypic variability for tolerance to water deficit and design screening techniques for use in breeding programs. However, genetic variation in whole plant water use efficiency (WUEwp, - biomass per water used) is not always correlated with variation in leaf level water use efficiency (WUEi - assimilation per stomatal conductance), increasing the difficulty of phenotypic prediction. To understand the disconnect, a mix of spring wheat cultivars and landraces from the Watkins collection were examined for variation in the mechanisms regulating biomass gain (BM) and water use (WU) as components of WUEwp. Specifically, the impact of leaf age and soil drying on stomatal conductance (gs) and assimilation (A) were assessed. Significant variation was observed for WUEwp (two-fold), with genotypes Krichauff and G1 (Watkins) consistently displaying high WUEwp and Gatsby low WUEwp. Increased WUEi was correlated with increased WUEwp, even though no significant genetic variation was observed for WUEi. Additionally, sustained A across leaf age was observed in Krichauff but not Gatsby, corresponding with their measures of WUEwp. Further, lower levels of photosynthetic limitation by rubisco (Vcmax) and decreased leaf biomass partitioning (LP) were correlated to higher WUEwp. While WUEi may not always predict variation in WUEwp, other variables (Vcmax, LP, and sustainability of A across leaf age) were strongly associated with the whole plant response. Thus, in vivo measures of photosynthetic limitation and other whole plant proxies, such as area-based estimates of leaf partitioning, serve as useful tools in predicting WUEwp.