Paleomagnetic data for nearly all geologic periods are now available from North America and Europe. The more reliable Paleozoic and Mesozoic evidence is analyzed in terms of some North Atlantic paleogeographic models, but the usefulness of paleomagnetism in testing such reconstructions is not obvious and depends critically upon the geometry of continental drift.

The divergence of the North American and European apparent polar-wandering curves is systematic only over their Carboniferous, Permian, and Triassic parts. The data are simultaneously compatible with two major rotations since the Permian: (1) separation of the two continents initially about the polar pivot proposed by Schuchert and Bullard et al. and the northward movement of North America relative to Europe, and continuing separation since the Triassic about a new pivot, perhaps near Wegener's Alaskan one; and (2) a chiefly northward 50° displacement of the two continents, imagined to be "rigidly connected," equivalent to their joint clockwise rotation about a pivot near the equator. Preliminary paleomagnetic data suggest that Greenland largely shared the second rotation. If thos displacements are assumed to be real, the east-west component of the first was recorded paleomagnetically only because the second rotation occurred as well. The latter could be due to either polar wandering or continental drift; the two mechanisms ultimately may become distinguishable by geometric checks from ocean-floor spreading. Cretaceous data from Europe are lacking, but the North American Cretaceous pole is displaced 32°, mainly eastward, from the Triassic pole. Paleomagnetism also supports qualitatively Carey's post-Permian counterclockwise rotation of Spain.

Twenty-five samples of ignimbrite from three or four distinct bands of Ordovician (Llanvirnian?) ignimbrite exposed near Killary Harbour, Eire, gave a mean direction, D = 135°, 1 = +36°, ₉₅ = 9°, relative to bedding, after demagnetization to 540 oe (peak); a fold test was positive, but a residual secular variation component may be present. The corresponding south pole (7° N, 147° W) relative to Ireland is in the present Pacific Ocean, as are other European and North American early Paleozoic poles. However, pre-Carboniferous paleomagnetism is still inconclusive in testing "pre-Wegenerian" North Atlantic reconstructions, including the rotation of Newfoundland, and as much as a tenfold incr ase of data may be required. Pending that, a paleomagnetic scheme for testing Wilson's concept of a "proto-Atlantic" is proposed. Far from being premature, such schemes could guide systematic paleomagnetic sampling programs in the future.