Late Precambrian crustal evolution in the North Eastern Desert of Egypt occurred in a strongly extensional tectonic environment and was accompanied by abundant bimodal igneous activity. The extrusive and intrusive expressions of this magmatism, known as the Dokhan Volcanics and Pink Granites, respectively, were studied in detail from two areas. The Dokhan Volcanics and associated feeder dikes consist of a “mafic” suite dominated by andesites (∼60% SiO2) and smaller volumes of basalt and a “felsic” suite composed of rhyolite tuffs, ignimbrites and hypabyssal intrusions (∼72–78% SiO2). The rocks of the mafic suite display calc-alkaline trends on an AFM diagram but are enriched in incompatibles such as TiO2, P2O5, K2O, Rb, Sr, Ba, Zr, Y, Nb, and LREE. Rare earth element patterns are steep, with (Ce/Yb)n = 7.7 to 16.8. They contain moderate Ni (60 ppm) and Cr (95 ppm), indicating limited low-P fractionation. The melts of the mafic suite are interpreted to have formed either by ≤25% batch melting of eclogite or by ∼10% batch melting of LREE-enriched garnet lherzolite. The rocks of the felsic suite include Dokhan rhyolites and the epizonal Pink Granites. These contain 72–78% SiO2, are metaluminous and peraluminous, and have the high K2O/Na2O and FeO*/(FeO*+MgO) characteristic of post-tectonic, “A-type” granites. They are moderately enriched in incompatible elements, but their REE patterns overlap with those of the mafic suite, from which they can be distinguished by deep europium anomalies (Eu/Eu*=0.08–0.64) and flat HREE patterns=((Yb/Er)n=0.90–1.16). They share with the rocks of the mafic suite isotopic characteristics of depleted mantle, precluding anatexis of much older continental crust. The europium anomalies covary with Sr contents and indicate that plagioclase control was important, while the flat HREE patterns preclude residual garnet in the source. Hence the felsic melts could not have formed by anatexis of garnet-bearing mafic lower crust. Such melts could have formed by anatexis of amphibolite-facies crust, an interpretation which is not favored because the melts are not saturated with P2O5. Alternatively, the felsic melts may have formed via low-P fractional crystallization of the mafic melts, with about 2/3 removal of mostly plagioclase and amphibole along with minor apatite and zircon. This may have been accompanied in the latest stages of magmatic evolution by liquid-state fractionation such as thermo-gravitational diffusion or halide complexing.