The use of 3H and 3He to estimate the apparent age of groundwater presumes that differences in their transport behavior have little effect on 3H-3H ages. Here we show that 3H and 3He are highly susceptible to diffusive fractionation, which can result in substantial errors in 3H-3He apparent ages of post-1950, prebomb peak groundwater. Previous analyses in homogeneous and weakly heterogeneous systems suggested that significant diffusive fractionation effects on 3H-3He ages only manifest at low velocities. We show that the potential for diffusive fractionation depends strongly on the subsurface heterogeneity, flow field, and 3H source. Results indicate that within aquitards dominated by diffusive mass transfer, diffusive fractionation can lead to artificially old apparent age dates. Along preferential flow paths, diffusive fractionation can lead to artificially young apparent age dates. Along preferential flow paths strongly affected by diffusive mass transfer, this phenomenon can result in the same 3H-3He apparent age for postbomb peak and older post-1950, prebomb peak groundwater. Therefore, where nonunique age dates are plausible, it may be difficult to distinguish postbomb peak groundwater from post-1950, prebomb peak groundwater affected by diffusive fractionation in the absence of additional hydrogeologic information sufficient to delineate the effects of diffusive fractionation.