Honey mesquite (Prosopis glandulosa Torr.) may maintain apical dominance after a treatment that causes partial top-kill (PTK) and leaves canopies with “stem flagging.” In contrast, top-killing treatments stimulate multistemmed regrowth (i.e., basal sprouting; BSP). Because this difference may impact competition with grasses, a better understanding of physiology associated with PTK and BSP canopies compared with untreated canopies is needed. We quantified predawn leaf water potential (Ψ_PD), leaf-level gas exchange rates (photosynthesis [A] and stomatal conductance [g_s]), and whole-tree stomatal conductance (G_s) of untreated, PTK, and BSP mesquites 2–11 yr after aerial herbicide application in a north Texas savanna. Total leaf area was nearly 4 × greater in untreated and BSP compared with PTK trees. In a few situations where soil moisture stress was greatest, untreated mesquites exhibited more negative Ψ_PD and lower leaf-level gas exchange than did PTK mesquites. BSP mesquites occasionally had greater Ψ_PD and leaf-level gas exchange than untreated mesquites. Since imbalances in root-to-shoot ratios caused by PTK were largely not manifested at the leaf-physiology level, PTK mesquites likely adjusted rapidly to herbicide disturbance. When g_s was scaled to the whole canopy, G_s estimates were 3–4 × greater in untreated and BSP than PTK trees. Thus, canopy leaf area was the primary driver of differences in whole-canopy G_s. It is LIKELY that transpiration would be lower in a stand of untreated mesquites compared with PTK mesquites of similar stand density and basal areas before treatment, as results from comparing untreated and PTK mesquites indicate. Partially top-killed trees that maintain apical dominance and do not resprout should compete less with grasses for water and light and facilitate species diversity and heterogeneity.