** **

TIMS measurements of 234U/238U, 230Th/232Th and 232Th/238U ratios in Fe-Mn crusts have been undertaken to study the behavior of U and Th in these Fe-Mn deposits and to evaluate the potential interest of Fe-Mn crusts to recover Th isotope ratios of sea-water.

U and Th isotope compositions and concentrations have been measured in subsamples from depth profiles of the outermost margin of three crusts : D18-1 (west equatorial Pacific), SO-79 145 KD (Peru bassin), Ant 109 D-C (south west Indian).For the D18-1 Fe-Mn crust, decrease of U and Th isotope ratios with depth gives consistent growth rates of about 7.2 mm/Ma, and defines, in a plot of Ln[(234U)excess/238U] against Ln[230Th/232Th], a linear correlation with a slope of **l**234U/**l**230Th, which correponds to the alignment expected if all the fractions precipitated with the same U and Th isotope ratios and remained closed after their formation (Chabaux *et al.*, 1995). On the other hand, for the two other crusts, decrease of U and Th isotope ratios with depth gives significantly different ages, and defines an aligment with a slope lower than the theoritical slope of **l**234U/**l**230Th in the Ln[230Th/232Th] - Ln[(234U)excess/238U] diagram. A simple growth model, by precipitation of metal oxydes with constant Th and U initial ratios and an evolution in closed-system, cannot be retained for these two Fe-Mn crusts. For the three crusts studied here, 232Th/238U ratio also decreases of × 20% from the surface to O.5-1mm. Interpretation of all these results in term of changes of oceanic Th flux with time (following the suggestion of Huh and Ku, 1990) is difficult. If 232Th/238U decrease represent a decrease of oceanic 232Th flux from now to ×120 ka, 238U-234U-230Th disequilibria in SO-79 145 KD and Ant 109 D-C Fe-Mn crusts would indicate a much more important decrease of oceanic 230Th flux during the same time period, which seems to be unprobable. An alternative explanation is that all these results show that Fe-Mn crusts do not

behave systematically in closed system for U-Th : simple U-diffusion models (Chabaux *et al.*, 1995; Ku *et al*., 1979) account for the exponential decrease of 234U/238U and 230Th/232Th ratios with depth and for a linear correlation of slope inferior to **l**234U/**l**230Th in the Ln[230Th/232Th] - Ln[(234U)excess/238U] diagram - Decrease of 232Th/238U

with depth can be due to a 238U mobility or to a 232Th redistribution which goes with the mineralogical reorganization of surface layers during fixation of Fe-Mn oxides (Kusakabe and Ku, 1984).

In consequence, determination of Th isotope ratios of seawater from Fe-Mn crusts cannot be made by extrapolation of Th depth profiles to zero surface neither by using 234U/238U isotope ratio to correct Th isotope ratios for radioactive decay. The only reliable method is to use U-Th correlation defined in the Ln[230Th/232Th] - Ln[(234U)excess/238U] diagram and to extrapol the Th isotope ratio by assuming a present day seawater value for U.

With this method a variation of Th activity ratios in seawater is pointed out between the different oceans, with the lowest values in the Atlantic ((230Th/232Th) × 20-35) and the highest values in the Pacific ((230Th/232Th) up to 600). At least for Pacific, variations of Th activity ratios can be related to a variation of intensity of 232Th flux into seawater as a broad relationship exists between seawater Th isotopic ratios and Th/U ratios of Fe-Mn crusts. This simple logical for Th isotope variations does not seem to be apparent for the Indian ocean.

Chabaux, F., Cohen, A.S., O'Nions, R.K. & Hein, J.R., *Geochim. Cosmochim. Acta* **59**, 633-638 (1995).

Huh, C.A. & Ku, T.L., *Paleoceanography* **5**, 187-195 (1990).

Ku, T.L., Omura, A. & Chen, P.S., In *Marine Geology and Oceanography of the Pacific Manganese Province* (Bischoff, J.L. & Piper, D. Z., Eds.) 791-814 (1979).

Kusakabe, M. & Ku, T.L., *Geochim. Cosmochim. Acta* **48**, 2187-2193 (1984).