Tarney & Windley (1977) questioned the andesite model for crustal growth (Taylor & White 1966; Taylor 1967; Jakes & Taylor 1974; Taylor 1977). Much of their argument is based on rare earth element (REE) data from Archaean granulites, which they claim are representative of the lower continental crust.
The most damaging comment against the andesite model is their calculation (Tarney & Windley 1977, fig. 1) of the REE pattern in the lower continental crust. Using a value for andesite (representing the bulk crust) and average sedimentary rocks (representing the upper crust)they calculated a lower crustal composition which has a strong positive Eu-anomaly and extremely low LREE (in fact there is no La or Ce!). They stated that their andesite value was taken from Nance & Taylor (1976); this is incorrect, since no andesite value is reported in that paper, nor is an andesite value reported in the only other paper by those authors (Nance & Taylor 1977).
Tarney & Windley do not tabulate their REE values, but estimates can be made from their chondrite-normalized plots. Their values are compared (Table 1) with the estimates of Taylor (1977, 1979) for the bulk continental crust, based on the andesite model.The upper crustal estimate (Nance & Taylor 1976) and the correct calculation for the lower crust (assuming the upper crust is 1/3 of the whole crust) are also given in Table 1.
These values are plotted in Fig. 1. In marked contrast to Tarney & Windley's calculations, the REE pattern for the lower