A survey is given of the dimensions and composition of the present continental crust. The abundances of immobile elements in sedimentary rocks are used to establish upper crustal composition. The present upper crustal composition is attributed largely to intracrustal differentiation resulting in the production of granites senso lato. Underplating of the crust by ponded basaltic magmas is probably a major source of heat for intracrustal differentiation. The contrast between the present upper crustal composition and that of the Archean upper crust is emphasized. The nature of the lower crust is examined in the light of evidence from granulites and xenoliths of lower crustal origin. It appears that the protoliths of most granulite facies exposures are more representative of upper or middle crust and that the lower crust has a much more basic composition than the exposed upper crust. There is growing consensus that the crust grows episodically, and it is concluded that at least 60% of the crust was emplaced by the late Archean (ca. 2.7 eons, or 2.7 Ga). There appears to be a relationship between episodes of continental growth and differentiation and supercontinental cycles, probably dating back at least to the late Archean. However, such cycles do not explain the contrast in crustal compositions between Archean and post-Archean. Mechanisms for deriving the crust from the mantle are considered, including the role of present-day plate tectonics and subduction zones. It is concluded that a somewhat different tectonic regime operated in the Archean and was responsible for the growth of much of the continental crust. Archean tonalites and trond-hjemites may have resulted from slab melting and/or from melting of the Archean mantle wedge but at low pressures and high temperatures analogous to modern boninites. In contrast, most andesites and subduction-related rocks, now the main contributors to crustal growth, are derived ultimately from the mantle wedge above subduction zones. The cause of the contrast between the processes responsible for Archean and post-Archean crustal growth is attributed to faster subduction of younger, hotter oceanic crust in the Archean (ultimately due to higher heat flow) compared with subduction of older, cooler oceanic crust in more recent times. A brief survey of the causes of continental breakup reveals that neither plume nor lithospheric stretching is a totally satisfactory explanation. Speculations are presented about crustal development before 4000 m.y. ago. The terrestrial continental crust appears to be unique compared with crusts on other planets and satellites in the solar system, ultimately a consequence of the abundant free water on the Earth.