The production of superheated melt during hypervelocity impact events has been proposed to be a common occurrence on terrestrial planetary bodies. Recent direct evidence of superheated impact melt temperatures exceeding > 2370 degrees C from the Kamestastin (Mistastin Lake) impact structure, Canada, was based on a single impact glass sample. Such high superheated melt temperatures have strong implications for the evolution of crustal material, the thermal history of impact cratering events, and the rheology of impact melt. However, although widely predicted in previous studies, with the exception of the Mistastin Lake impact glass, there is little direct evidence for superheated temperatures in multiple settings across an impact structure. Therefore, an outstanding question is how heterogeneous are superheated conditions across a single impact structure. In this work, we analyze the crystallographic orientations and microstructures of zircon grains and the precursor parent phases of baddeleyite crystals, from four different samples representing the entire melt-bearing stratigraphy at Mistastin: an impact glass, a vesicular clast-poor impact melt rock, a clast-rich impact melt rock, and a glass-bearing impact breccia. Using electron microprobe analysis followed by electron backscatter diffraction, we discovered that four zircon grains with vermicular coronae of baddeleyite crystals from the impact glass contain evidence for a cubic zirconia precursor, indicative of temperature conditions > 2370 degrees C. We also report evidence of superheating up to 1673 degrees C in the glass-bearing impact breccia. In addition, we also report the first occurrence at Mistastin of the high-pressure zircon polymorph reidite and former reidite in granular neoblastic (FRIGN) zircon in grains from the glass-bearing impact breccia, implying minimum peak shocks from 30-40 GPa. The identification of superheating from two localities at Mistastin demonstrates (1) that superheating is not restricted solely to rapidly cooled impact melt rock samples and is therefore more distributed across impact structures, and (2) we can investigate the P-T evolution pathways of impact melt from different impact settings, providing a clearer picture of the thermal conditions and history of the impact structure. (C) 2022 Elsevier B.V. All rights reserved.