The case for primary, mantle-derived carbonatite magma

There is much debate about whether carbonatite magmas are derived in 'secon dary' fashion through the advent of liquid immiscibility operating in the c rust on evolved nephelinitic magma, or whether they are derived in the mant le by direct partial melting of a carbonated periodtite. This paper briefly summarizes the epsilon(Sr)-epsilon(Nd) data for carbonatites in general an d evaluates the isotopic relationships between carbonatites and alkaline si licate rocks in several well-studied complexes from Africa. Available data for carbonatites younger that 200 Ma have a range in epsilon(Sr)-epsilon(Nd ) that is less than that found in oceanic basalts despite the fact that car bonatites traverse lithospheres that are much more complex than those in th e oceans. By contrast, for the Napak, Kerimasi, Shombole, Dorowa, Shawa and Spitskop complexes the alkaline silicate rocks show greater variability an d have more enriched epsilon(Sr)-epsilon(Nd) (higher epsilon(Sr), lower eps ilon(Nd)) values than their associated carbonatities. In general, the carbo natites have isotopic compositions that are closer to the more primitive si licate rocks, such as melilitites and olivine nephelinites, than to more ev olved nephelinites and phonolites. In the case of the Napak Complex the enr iched component was introduced from the lower crust whereas for the Dorowa and Shawa complexes of SE Zimbabwe, the component was derived from the sub- continental lithospheric mantle. These relationships indicate that the carb onatites must have existed as discrete magmas in the mantle and argue again st a derivation by liquid immiscibility in the crust. although a contrast i n isotopic composition does not rule out an immiscibility relationship at m antle depths and early in the evolutionary history of a melilitic or nephel initic magma, there is little experimental support for it. Existing experim ental data indicate the immiscibility between carbonate and silicate liquid s is favoured at low, crustal, pressures by that immiscibility is unlikely to occur in realistic mantle melts or their derivatives at mantle pressures . Many experimental data exists to show that magnesian carbonatite liquids form as the near-solidus melts of carbonated mantle peridotite at depths in excess of 75 km. We conclude that the calcitic dolomitic carbonatite magma s discussed in this paper are best considered as being derived from primary carbonatite magmas generated in the mantle by partial melting of carbonate d peridotite.