Paleomagnetic results are reported from the Nosib, Otavi, and Mulden groups of the Damara Super-group, a late Precambrian shelf sequence on the southern margin of the Congo craton in Namibia. Three magentizations were isolated in the Nosib group samples. In order of decreasing blocking temperature they are NQ1 (n = 6 sites, λ = 28 ¿N, ϕ = 323 ¿E, α95 = 15¿), NQ2 (n = 7 sites, λ = 51 ¿S, ϕ = 213 ¿E, α95 = 12¿), and NQ3 (n = 13 samples,λ = 09 ¿N,ϕ = 295 ¿E,α95 = 13¿). Overall precision of all three magnetizations upon tectonic correction suggests that they predate Pan-African (650--450 Ma) folding. Two magnetizations were isolated in the Otavi group samples, above the Nosib in stratigraphic sequence. The DC1 component of possible prefolding age (n = 4 sites, λ = 52 ¿S, ϕ = 186 ¿E, α95 = 35¿) has been over-printed by the DC2 magnetization (n = 10 sites, λ = 55¿S, ϕ = 044 ¿E, α95 = 15¿) of probable postfolding age. A single magnetization of probable pre-folding age was isolated in the overlying Mulden Group samples (n = 6 sites,λ = 12 ¿S,ϕ = 090 ¿E,α95 = 16¿). Together with previously published paleomagnetic data from Africa, the new data showed that no great relative movements have occurred between the Congo and Kalahari cratons during the interval of Pan-African tectonism in the Damara belt (McElhinny and McWilliams, 1977). Continental collision preceded by large relative displacements and closure of a wide ocean (e.g.,a Himalayan analog) is effectively ruled out for the Damara belt. We develop an alternative model consistent with the available paleomagnetic and geologic data, which invokes rifting, heating, and stretching of the lithosphere underneath the Damara belt, followed by delamination of the subcrustal lithosphere. Hot asthenospheric material rises to take the place of the detached and sinking lithospheric base, inducing subduction and interstacking of continental crust. The much thickened continental crust is partially melted at depth, intruded by synorogenic and postorogenic granites and finally uplifted and eroded to its present level of exposure. The model is compatible with plate tectonics in that the development of the Damara belt can be broadly compared with modern marginal seas, with the exception that stretching of the lithosphere was not induced by secondary convection above a downgoing slab. Possible cause for stretching are rising mantle plumes or intracontinental distortion within the pre-Damara African plate.