A Bouguer gravity map of an area of northeastern Georgia encompassing parts of the Inner Piedmont, Charlotte belt, and Carolina Slate belt has been constructed from gravity stations spaced ∼2 km apart. Anomalies in the Charlotte–Carolina Slate belts are due to metamorphosed mafic rocks in the upper crust (depth extent <5 km, positive anomalies) and unmetamorphosed Alleghenian granites that extend to depths of as much as 16 km and produce negative anomalies. One of these granites is the Danburg granite; the other granite is not exposed at the surface. The Elberton granite in the Inner Piedmont has too little gravity expression in this survey for a structural interpretation, because the density contrast between the Elberton and the country rock is too small. The Middleton-Lowndesville fault zone is the geologically mapped boundary between the Inner Piedmont and the Charlotte–Carolina Slate belts in this area; there is a sharp gradient of Bouguer gravity across the fault zone caused by the juxtaposition of the shallow mafic rocks of the Charlotte–Carolina Slate belts with Inner Piedmont rocks. There is also a dramatic difference in the nature of the gravity field across the fault zone due to the presence of near-surface sources on the southeast (Charlotte–Carolina Slate belts) side and the absence of such sources on the northwest (Piedmont) side. We interpret the Middleton-Lowndesville fault zone as a surface (that is, exposed) suture between two suspect terranes accreted onto the North American margin during the Paleozoic. The northwest part of the map is occupied by a broad gradient that is part of the Piedmont gravity gradient, a feature that runs along the entire southern Appalachians, crosscutting mapped boundaries between belts. We interpret this gradient as representing the continental margin of early Paleozoic North America and as being caused by the juxtaposition of sialic Granville mid- to lower-crustal material with more mafic material to the southeast. This buried suture is covered by overthrust terranes. We propose that thrusting must extend at least to a point southeastward of the suture.