The ocean is a major component of global heat transport and represents a large exchangeable reservoir of CO_2. The importance of these effects on climate can be quantified with records of ocean temperature, chemistry and dynamics spanning past climate change. One approach to reconstruct past ocean conditions relies on the chemical composition of CaCO_3 skeletons from coral. Despite the utility of these geochemical proxies, several lines of evidence suggest that biomineralization, the process corals use to build their skeletons, also influences composition, complicating the interpretation of past records. Coral grown under constant environmental conditions, either collected from the deep-sea or cultured in the laboratory, are used to quantify and spatially map the effects of biomineralization on skeletal composition. In modern deep-sea coral, Mg/Ca increases with decreasing Sr/Ca in most the skeleton, consistent with closed-system (Rayleigh) precipitation. Results also show composition strongly follows skeletal architecture. Centers of calcification (COCs) are small regions of disorganized crystals thought to be the initial stage of skeletal extension. Unlike the rest of the skeleton, Mg/Ca ratios vary more than two fold within the COCs while Sr/Ca is near constant. Our data provide new constraints on a number of possible mechanisms for this effect. In a complementary set of experiments the nanoSIMS, a new instrument capable of accurate sub-micron compositional analysis, is applied to adult cultured surface coral (1) mapping the pattern of metal ion incorporation in new growth and showing that the calcifying fluid is likely in direct exchange with seawater; and (2) testing the sensitivity of Me/Ca ratios to aragonite saturation (Omega). Despite a large range of Omegas and calcification rates, the average Sr/Ca of nanoSIMS spot measurements in cultured coral are within 1.2% (2 sigma std. dev. of the 5 means). These data suggest that temperature is a more significant control on Sr/Ca than aragonite saturation between Omega = 2.5--5. Within the framework of a closed-system (Rayleigh) model for biomineralization the results constrain explanations for the sensitivity of coral calcification rates to ocean acidification, improving our understanding of how anthropogenic CO_2 will impact coral reefs.