Unlike radiometric dating where the rate constant is independent of environment, amino acid racemization (AAR) is a chemical reaction where the rate constant is dependent primarily on temperature, as well as a far more modest vital effect, which can be constrained by monospecific sampling. Most AAR dating studies utilize carbonate fossils, where pH is constant, and some fossils approximate a closed system for biomolecules. Changes in indigenous proteins preserved in carbonate skeletons can be used to estimate the time elapsed since death of the organism. Racemization, the most widely used reaction in amino acid geochronology, refers to the inversion of L-amino acids to their D- configuration. For the protein amino acid L-isoleucine, racemization occurs about only one of the two chiral carbon atoms; this reaction (epimerization) proceeds from an initial D/L ratio near zero to an equilibrium ratio of about 1.30. The resolving power of D/L ratios as relative-age indices is a function of the reaction rate, limited by natural variability, defined as the spread in ratios in five well preserved shells of the same age from a single stratum. Analytical uncertainty within a single laboratory is relatively insignificant, but interlaboratory comparisons suggest caution in comparing ratios determined at different laboratories or by different techniques. Sampling appropriate portions of a shell is essential to minimize variability in D/L ratios. Site temperature exerts the greatest control on reaction rate, hence resolving power. For tropical sites (>25°C) age differences < I ka can be resolved within the Holocene and ca. 5 to 10 ka for older samples as the reaction rate decreases; equilibrium is attained after 150 to 300ka. The reaction rate is substantially lower at mid latitude sites (ca.+ 10°C), where equilibrium requires about 2 Ma and resolution within the last glacial cycle is no better than 10 to 20 ka. Arctic sites (<-10°C) show almost no measurable racemization within the Holocene, and in extreme cases even within the last 100 ka; equilibrium is not attained in 10 Ma. Careful collecting is essential. Ideally, samples have been buried by at least 1 m of overburden; 2 m is even better. Dense carbonate is optimal. Each amino acid racemizes at a different rate, allowing targeting amino acids with reaction rates optimal for particular time ranges.