Description: Planktonic foraminiferal tests of the spinose species Orbulina universa, of the non-spinose Globorotalia tumida-menardii complex, and of a mixed species assemblage (grain size fraction 200-400 μm) were isolated from Sierra Leone Rise core 13 519-2 and analyzed for free, total, and bound (by difference) amino acids to study the isoleucine epimerization mechanism in fossil foraminiferal tests and to define empirical calibration curves for dating deep-sea sediments over the past 900,000 years. Total isoleucine epimerization curves typically separate into three "linear" segments of decreasing apparent rates with increasing time and exhibit a pronounced "species effect". The degree of epimerization attained at time is considerably lower in 0. universa than in G. tumida-menardii while the mixed species results scatter between the limits delineated by the two monospecific curves. Total allo/iso ratios are closely related to the proportion of free to total isoleucine accumulating in the tests indicating that the rate of hydrolysis of matrix proteins and peptides controls the overall epimerization reaction. The results are consistent with experimental evidence (KRIAUSAKUL & M1TTERER 1978, 1980 a, b) whereupon isoleucine epimerizes at a rapid rate in terminal positions but at slow rates in interior positions as well as in the free state. Notwithstanding free isoleucine exhibits the highest degree of epimerization due to preferential hydrolysis of extensively epimerized terminal isoleucine. Species-specific hydrolysis and epimerization rates are maintained until about 50% of bound isoleucine have been hydrolyzed to the free state corresponding to a total allo/iso ratio of about 0.5. Remaining peptide units appear to be more resistent against hydrolysis and separate species then show the same apparent epimerization rate dominantly controlled by the slow conversion rate in the free state until equilibrium is achieved in Miocene samples under deep-ocean temperature conditions (KING & HARE 1972 a). The degree of epimerization attained at comparable time in separate species will, however, remain different due to different initial rates of hydrolysis. Selective leaching of free isoleucine from the tests results in a lowering of total allo/iso ratios and apparent rate constants when significant amounts have been generated by diagenetic hydrolysis. Comparison with results based on Pacific core V 28-238 (KING & NEVILLE 1977) indicates an inverse relationship between the intensity of surficial carbonate dissolution and post-depositional leaching. Tests deposited under less corrosive conditions (Atlantic core 13519-2) will more significantly deviate from a closed system during diagenesis than those deposited in more corrosive bottom waters (Pacific core V 28-238). Carbonate dissolution removes proteinaceous lamellae from more external shell structures, which otherwise - if preserved during deposition - will preferentially be subject to leaching of free amino acids during diagenesis. This effect may qualify the applicability of empirical epimerization calibration curves to different depositional environments although the general agreement of the epimerization curves obtained for both sites is promising for future research. Identical apparent epimerization rates in G. tumidamenardii tests younger than 120,000 years (when leaching is stiH insignificant) at both Atlantic and Pacific sites are not consistent with the presently different bottom water temperatures in these regions. It is concluded that glacial North Atlantic Deep Water was cooler than at the present at the Sierra Leone Rise.