Reflectance spectroscopy is important for placing lunar samples into a regional and global geologic context. To this end, the ultraviolet–visible (UV–vis) color ratio, used to estimate the TiO 2 composition of mature mare basalts, has been one of the most widely used spectral parameters in lunar exploration. We examine the correlation between UV–vis color and TiO 2 content using a combination of Clementine, Lunar Prospector, and sample data to document the extent to which color is dependent upon TiO 2. Examination of the remotely sensed data reveals that the correlation between UV–vis ratio and TiO 2 composition is best represented by a sigmoidal trend rather than the canonical linear or curvilinear correlation. With this information, we are then able to evaluate between two models that propose different explanations for the relationship between UV–vis color and TiO 2. The first model attributes the correlation between TiO 2 and UV–vis color to spectrally neutral opaques (i.e., ilmenite), while the other emphasizes the effect of Fe–Ti charge-transfer in lunar glasses and dual scattering mechanisms between high- and low-Ti basalts. We do not find evidence in the spectral data to support the occurrence of Fe–Ti charge-transfer in lunar glass as the principal cause for color in high-Ti basalts. The data also do not substantiate the existence of different scattering mechanisms (e.g., volume v. surface scattering) between high- and low-TiO 2 basalts. Instead, our analyses substantiate that the spectral effects of ilmenite exhibit a major influence over the UV–vis ratio. By including sample data we find that in addition to ilemenite/TiO 2 content, factors such as FeO content, ilmenite grain size, modal abundance of plagioclase, and the olivine-to-pyroxene ratio in a mare soils can influence the UV–vis continuum. These findings point to promising avenues of research that future UV–vis spectral techniques can exploit in order to yield more accurate TiO 2 estimates and potentially additional petrologic information.