Grassland root responses to mowing and fertility are less well known than shoot responses, even though as much as 90% of productivity in semiarid grasslands occurs belowground. Thus, understanding root responses may aid the management of invasive grassland species such as Agropyron cristatum (L.) Gaerth (crested wheatgrass). We asked whether root responses reflect shoot responses to mowing and fertility in native grassland with and without a major component of crested wheatgrass. We subjected grasslands in northern Montana to 5 yr of mowing at two nitrogen (N) levels and followed root responses with minirhizotrons. Surprisingly, the roots of both native and invaded grasslands were unaffected by mowing and N addition, despite significant changes in shoot mass across both vegetation types. Root length was significantly greater beneath areas heavily occupied by crested wheatgrass (363 m · m⁻² image ± 200, mean ± standard deviation [SD]) than areas comprising largely native grassland (168 m · m⁻² image ± 128 SD). Also, no interactions occurred between year and any other factor, indicating that there were no changes in belowground responses over the 5 yr examined. In contrast, shoot mass was significantly reduced by mowing (not mowed, 612 g · m⁻² ± 235 SD; mowed, 239 g · m⁻² ± 81 SD) and was significantly increased by N addition (no added N, 380 g · m⁻² ± 215 SD; added N, 488 g · m⁻² ± 287 SD). In conclusion, 5 yr of mowing decreased shoot mass, but not root mass. On the other hand, 5 yr of N addition increased shoot mass, but not root mass. Given that most production and competition in grasslands occurs belowground, this suggests that mowing may not be a successful tool for reducing crested wheatgrass root length, regardless of soil fertility.