Earth and Planetary Science Letters', 114 (1992) 149 157 149 Elsevier Science Publishers B.V., Amsterdam [CH] Th-Sr isotopic relationships in MORB K.H. Rubin and J.D. Macdougall Scripps Institution of OceanograPhY, 9500 Gilman Drive, La Jolla, CA 92093-0220, USA Received April 24, ]992; revision accepted October 8, 1992 ABSTRACT Although MORB are well known to exhibit a limited range in isotopic ratios such as 87Sr/86Sr, a compilation of existing (23°Th/Z32Th) data, which reflect present Th/U in the source, shows a large dispersion. This is in contrast to the, admittedly limited, data available for oceanic islands, for which there is a broad correlation between 8TSr/S6Sr and (23°Th/Z32Th). The distribution of data for MORB is approximately bimodal, with relatively depleted samples plotting from within to well above the band defined by the oceanic island correlation (S7Sr/S6Sr = 0.70230-0.70270 and (23°Th/X32Th) = 1.1-2.3) and relatively enriched MORB plotting from within to below this band (SVSr/S6Sr = 0.70270-70320 and (Z3°Th/ 232Th) = 0.8-1.4). The data are consistent with the generally accepted idea that E-MORB are generated by the interaction of depleted mantle with more enriched material having higher Th/U and ~VSr/S6Sr. However, the large range in source Th/U implied by the thorium isotopic compositions of both enriched and depleted MORB requires complex melting and/or other recent source modification processes. Much of this variability may be caused by multiple, small-degree melt extraction at the spreading ridge, a process which would affect all incompatible trace element concentrations and ratios. Introduction The thorium isotope ratio 23°Th/232Th is unique among radiogenic isotope tracers because it reflects a present-day characteristic of the source: its U/Th ratio. This is a result of the fact that 23°Th (half-life = 75,200 yr) is a decay prod- uct of 238U and is in secular equilibrium in man- tle that has not undergone chemical disturbances for approximately 350 ka or more. Since the Th isotope ratio responds quickly to changes in man- tle Th/U, it yields quite different information to that supplied by Pb isotopes, which reflect the long-term integrated history of Th/U in the source. Similarly, other systems with long-lived parent isotopes (e.g., Rb-Sr and Sm-Nd) provide information only about long-term source charac- teristics; their present parent-daughter ratios may be very different from those inferred from the isotope ratios. Thorium isotope studies are, therefore, uniquely useful for investigating petro- Correspondence to: K.H. Rubin, School of Ocean and Earth Sciences and Technology, University of Hawaii, 2525 Correa Road, Honolulu, HI 96822, USA. genic processes in general, and very recent source depletion or enrichment events in particular [1,2]. The activity ratio (23°Th/232Th) in freshly erupted lava is determined by Th/U in the source if: (1) the combined timescales of melting and mantle residence time, including source-to- surface transport, are short compared with the timescale of 23°Th decay; and (2) 23°Th and 23SU in the source materials are in secular equilibrium prior to melting. E~idence that the former is true for most mid-ocean ridge basalts (MORB) comes from the widespread occurrence of 226Ra ex- cesses (half-life = 1.62 ka) in freshly erupted lavas [3-5]. Also, because the residence time in the mantle of elements such as Th and U is long, compared with the half-life of 23°Th, it is reason- able to infer that the MORB source is in radioac- tive equilibrium at the onset of melting. It should be pointed out that some models predict that (23°Th/232Th) in freshly erupted lavas underesti- mates Th/U when the melting rate is small rela- tive to the 23°Th decay constant [6,7]. However, even if this is the case, the difference between the actual source Th/U and that calculated from (23°Th/232Th) is unlikely to be more than 10-15% on average [7] and would not affect the conclu- sions of this paper. 0012-821X/92/$05.00 © 1992 - Elsevier Science Publishers B.V. All rights reserved