Geochemical constraints on crustal anatexis: A case study from the Pan-African Damara granitoids of Namibia

Major and trace element models of recently published vapour-absent mica dehydration melting experiments are used to identify granitoids generated by muscovite and biotite dehydration melting, and to distinguish between plagioclase-limited and biotite-limited, biotite dehydration melting. In the case of granitoids from the Pan-African Damara mobile belt (Namibia), many of the leucogranites and Salem-type granitoids may be modelled by biotite dehydration melting. The low Rb/Sr granitoids (e.g. Donkerhuk Onanis, Salem Onanis, Donkerhuk Nomatsaus, Salem Goas) probably reflect feldspar-limited, biotite dehydration melting (a pelitic source) whereas the high Rb/Sr suites (e.g. Bloedkoppie leucogranite, Stinkbank leucogranite, Salem Swakopmund, Leucocratic Stinkbank granite) reflect biotite-limited, biotite dehydration melting (a greywacke source). Alaskites from the Damara belt have major element compositions which are consistent with muscovite dehydration melting, and their positive Eu anomalies are linked to high K2O reflecting K-feldspar entrainment. Combined Zr and LREE (light rare earth element) solubility models indicate that insufficient time (probably less than 10(4) years) had elapsed between melt generation and melt extraction to ensure that the alaskite melts attained their equilibrium concentrations of Zr and the LREFs. In contrast, the leucogranites and Salem-type granites have attained their equilibrium inventories of these trace elements. Combined Fe2O3 and MgO contents in some samples from two granitoids (the Salem Goas and Donkerhuk Onanis intrusions) are higher than those readily attainable by biotite dehydration melting indicating either: (1) that they contain a contribution from melts generated by incipient garnet breakdown or; (2) that they contain small amounts of an entrained ferromagnesian phase.