Streck et al. (2007) state that primitive andesite lavas (Mg# > 0.6; 53–63 wt% SiO2) from Mt. Shasta are not primary melts, but instead are mixtures of evolved, high-Sr/Y, low-Mg# dacite with primary, low-Sr/Y, high-Mg# basalt. They propose that this mixing forms primitive andesites worldwide, and they dismiss the idea that primitive andesite is an end member in arc magmatism.
Magma mixing is evident in primitive Aleutian andesites (e.g., Kay, 1978; Kelemen et al., 2003a; Yogodzinski and Kelemen, 1998) and xenoliths (e.g., Conrad et al., 1983; Yogodzinski and Kelemen, 2007). Streck et al. imply that our data support their hypotheses, but Aleutian mixing requires a primitive andesite end member with the highest Mg#, Cr, Ni, Th, Ba, Sr and light rare Earth elements (REE), and lowest Y and heavy REE. We show this with Sr/Y versus Mg# (Fig. 1A). High-Sr/Y lavas are mixtures of a low-Mg#, low-Sr/Y component and a high Sr/Y, high Mg# component. The high Sr/Y component is an andesite (Fig. 1B) with Mg# ~0.7, in Fe/Mg equilibrium with mantle olivine. There are no low Mg# lavas with higher Sr/Y, so the Streck et al. process did not form end-member Aleutian primitive andesites.
Mt. Shasta data (Fig. 2) are similar to that from the Aleutians, with a mixing trend extending toward a high-Sr/Y, high-Mg# andesite end member, nearly perpendicular to the mixing proposed by Streck et al. If their process formed any lavas at Mt. Shasta, they are rare. In any case, by analogy with the Aleutians, their dacite end member (Mg# 0.61) formed from a primitive andesite.
Streck et al. also misrepresent Pb isotope data for arc lavas. Citing our work (Kelemen et al., 2003b), Streck et al. (2007, p. 353) write, “elevated Pb isotopic signatures of [primitive andesites] worldwide could be interpreted in terms of a crustal contribution.” However, Kelemen et al. (2003b, our Figure 9; 2003a, our Figures 7–11) show that Aleutian primitive andesites have the most depleted, least “crustal” Sr, Nd, and Pb isotopes of any arc magmas worldwide.