Many studies of climate variability in the Tropical Ocean have used high-resolution chemical tracer records contained in coral skeletons. The complex architecture of coral skeletons may lead to the possibility of biases in coral records and it is therefore important to access the fidelity of coral geochemical records as environmental proxies. Coral skeletal records from the same coral colony, and even the same corallite, may show large variation due to differing extension rates, formational timing of the skeletal elements, colony topography, and sampling resolution. To assess the robustness of the skeletal record, we present d13C and d18O data based on different sampling resolutions, skeletal elements, and coral colonies of Montastraea faveolata species complex, the primary coral used for climate reconstruction in the Atlantic. We show that various skeletal elements produce different isotopic records. The best sampling rate to resolve the full annual range of sea surface temperature (SST) is 40 samples per year. This sampling frequency also consistently recovered SST variability measured at weekly intervals. A sampling rate of 12 times per year recovered 84% of the annual range recording average monthly SST changes through the year. Six samples per year significantly decreased the ability to resolve the annual SST range. The d18O recorded from two adjacent colonies were very similar, suggesting that this isotope can be trusted to record environmental changes. The d13C, on the other hand, remained highly variable, perhaps as a result of the activity of symbiotic algae (zooxanthellae).