Introduction: The persistence of lithospheric thickness variations from its earliest era provides constraints on the thermal evolution of Mars [1,2]. Under these constraints, plausible thermal evolution models require cooling of the crust that could be accomplished by either hydrothermal cooling in a fractured portion of the upper crust or the existence of a stable stratified mantle where heat-producing elements are sequestered in a deeper portion of the mantle [2]. CRISM (Compact Reconnaissance Imaging Spectrometer for Mars) [3] and OMEGA (Observatoire pour la Mineralogy, l’Eau, les Glaces et l’Activité) [4,5] observations of phyllosilicate minerals excavated from depths of 4-5km have been proposed as evidence in support of the deep crustal hydrothermal cooling model [6]. The most prevalent of these phyllosilicates appear to be Fe/Mg smectites, which typically form at moderate temperature (<200°C) and pressure conditions expected in a hydrothermal geotherm. Here, we use CRISM data to further evaluate models of crustal cooling. In stratified mantle cooling, the geothermal gradient shows an approximately linear increase in temperature with depth, creating a correlation between mineralogy and depth. This model predicts higher temperature mineral phases with depth. In a hydrothermal circulation environment, however, the crustal temperature stays relatively constant with depth to the base of the hydrothermal layer. This system would not produce noticeable changes in mineralogy as a function of depth, though it could generate regional mineralogic variations reflecting regions of cool water downwelling and warmer water upwelling [6]. Impact craters provide a window to the interior where larger craters expose rocks from deeper sections of the crust. Here we systematically analyze the mineralogy in central peaks, rims, and ejecta of craters across Noachian/Phyllosian-age terrains to assess if there is any systematic variation in mineralogy exposed by impact. Data Sets: CRISM is a visible to near infrared imaging spectrometer onboard the Mars Reconnaissance Orbiter. CRISM can acquire high resolution targeted images that have a spatial resolution of 18-35m/pixel and cover 544 wavelengths from 0.362-3.92 μm [7]. Photometric and atmospheric corrections are applied to images to account for variations in observation geometry and atmospheric gas absorptions respectively [3]. We analyzed a subset of spectral data from 1.0 – 2.6 μm in corrected FRT (Full Resolution Targeted) and HRL (Half Resolution Long Targeted) CRISM images. Spectral parameters were used to highlight regions of a CRISM observation that expressed characteristic mineralogical features [8]. We collected spectra from areas with strong spectral parameter values and ratioed these to spectrally neutral areas in the same image columns. These ratio spectra were then compared to standard laboratory spectra. We examined CRISM observations of 137 craters in Terra Tyrrhena and Noachis Terra, two areas of Mars that are comprised of mostly Noachian/Phyllosian-aged terrains (Fig. 1).