Understanding Neogene arc crustal thickening in the central Andes requires (1) some estimate of initial pre-Neogene (prior to 26 Ma) crustal thicknesses and (2) mechanisms that account for the remaining deficit in crustal thickening (10--30%). Mid-Miocene horizontal crustal shortening can explain most but not all crustal thickening in the modern central Andean arc. Systematic changes in geochemical and Sr, Nd, and selected Pb isotopic data of Late Cretaceous--Eocene (~78--37 Ma) and older arc magmatic episodes from north Chile provide new constraints on both. First, Andean crust may have been significantly thickened by long-term underplating of mantle-derived basalt from Jurassic to present. Second, estimated initial (late Eocene) crustal thicknesses of ~45 km are consistent with (1) amphibole- and garnet-bearing residual mineralogies for late Eocene syntectonic/posttectonic granitoids, (2) lower crustal P wave velocities of 7.3--7.7 km s-1 compatible with underplated mafic crust, and (3) results from recent experimental petrologic work showing garnet stability in mafic mineralogies ≥12 kbar (≥40 km crustal thickness). Analogous to older Andean magmatic episodes in north Chile, newly underplated basaltic crust may account for the remaining deficit in Neogene crustal thickening. Similar evolutionary patterns in geochemistry and initial Sr and Nd isotopic characteristics of Andean (200 Ma to present) magmatic rocks suggest that the Andean orogeny in this region evolved by a combination of processes of repeated arc migration, tectonic and magmatic crustal thickening, and igneous recycling which was controlled by periodically changing plate convergence rates and obliquity and corresponding changes in the rheologic behavior of the continental crust.