Paleomagnetic studies of volcanic terranes in the Appalachians provide quantitative data on the Ordovician paleogeography of the Iapetus ocean. New paleomagnetic results from submarine volcanics of the Middle to Late Ordovician Winterville Formation of Maine further constrain the evolution of Iapetus. Ten sites yield a tilt corrected direction of D/I = 3277/−21 (α95 = 9.3, k = 27.9); the corresponding paleomagnetic pole calculated for the average site location (46.8°N, 291.3°E) is located at 26°N, 148°E (dp = 5°, dm = 10°, A95 = 7.7). A Silurian conglomerate (Frenchville Formation) test is inconclusive due to a strong present‐day field overprint which obscures any primary remanence. The presence of a positive tilt test, however, supports the conclusion that the characteristic magnetization of the Winterville Formation is a prefolding, primary magnetization. An Ordovician paleolatitude of 11±5° for the Pennington Mountain terrane of northern Maine is indistinguishable from that of the Laurentian margin (15–20°) during the Middle to Late Ordovician. The paleolatitude of the Pennington Mountain terrane is also similar to, but slightly more equatorial than that of the previously studied Bluffer Pond Formation of the nearby Munsungun terrane (18±9°). We conclude that the Ordovician Pennington Mountain and Munsungun volcanic terranes of Maine were formed and acquired their characteristic magnetization near the Laurentian margin. The paleolatitudes of these terranes are also similar to that obtained from the Stacyville volcanics of the more outboard Lunksoos terrane (20±8°), but they contrast strongly with the Ordovician Miramichi terrane of northern New Brunswick that yielded a paleolatitude of 51° (+21, −16°). These results from Maine support the presence of one or more backarc basins adjacent to the Laurentian margin during the Middle to Late Ordovician.

Recent palaeomagnetic results from coeval Silurian sedimentary and volcanic rocks of Newfoundland yield contradictory results, with shallow characteristic directions recorded by the redbeds and steeper characteristic directions recorded by the volcanics. In their recent paper on the Springdale Group redbed, Hodych & Buchan (1994a) argue that the redbed magnetization is more reliable based on a fold test, dual-polarity magnetization, conglomerate test and an IRM acquisition experiment specifically designed to test for compaction or inclination error. Accordingly, these authors dismiss the volcanic results as well as the Silurian reference poles for Laurentia, citing possible remagnetization or other complications. However, we argue that the IRM experiments of Hodych & Buchan (1994a) are flawed and do not necessarily resolve the question of inclination error. Moreover, we point out that the conglomerate test is incomplete and that the proposed tectonic reconstruction based solely on the redbed magnetization implies a unique tectonic scenario that is incompatible with out present understanding of Silurian palaeogeography.

Ancient remanences are retained by the Early Silurian (429 + 6/−5 Ma) mafic volcanics of the Springdale Caldera (five sites) and the overlying red beds (seven sites). Dual polarity magnetizations are obtained by thermal demagnetization of samples from the red beds, whereas single polarity directions are observed in the volcanics. High unblocking temperatures indicate hematite as the remanence carrier in both the volcanics and sediments. These high-temperature, characteristic remanences are easily isolated and pass both the tilt and conglomerate tests; they are likely to be of primary Silurian age. Characteristic declinations are predominantly northerly and northeasterly, and indicate significant structural rotations on a local scale. When the results of the red beds and the volcanics are combined they show characteristic inclinations that are shallower than those of the correlative Botwood Group (ca. 36° vs. 43°) but not nearly as shallow as those reported from the King George IV Lake area (0.5°). Mean inclinations obtained from the Springdale red beds are, however, significantly shallower than those of the Springdale volcanics. The same difference can be seen in other previous Silurian paleomagnetic studies of central Newfoundland. We infer that an inclination error affects the red bed magnetizations of the Springdale Group, Botwood Group (Wigwam Formation) and rocks of the King George IV Lake area. Therefore, the results from Silurian red beds should not be used to determine paleolatitudes for central Newfoundland. The mean paleolatitude of the Springdale Group volcanics is 30°. The mean paleolatitudes for both the Springdale volcanics and Botwood volcanics (Lawrenceton Formation) are indistinguishable within paleomagnetic error limits from the predicted paleolatitude of Newfoundland on the northeast-trending North American margin. Thus, no detectable post-Silurian displacement is shown by the volcanics of the Springdale Group with respect to cratonic North America.

The Munsungun area of northern Maine is an early Paleozoic terrane within the Central Mobile Belt (CMB) of the northern Appalachians. The Bluffer Pond Formation of the Munsungun Terrane is composed primarily of pillow basalt and gabbro with minor intercalated sediments and has a maximum thickness of 1230 m. This unit is of early‐Late Ordovician (Caradocian) age, based on paleontological data, and folding most likely occurred during the Devonian (Acadian Orogeny). This region has been metamorphosed to subgreenschist facies. Samples for paleomagnetic study have been obtained from 15 sites of pillow basalt and seven sites of gabbro within the Bluffer Pond Formation. Natural remanent magnetization (NRM) directions are similar to the expected present‐day field direction and are interpreted to be strongly affected by a viscous overprint. Detailed stepwise thermal demagnetization fails to adequately resolve a partial overprint direction with intermediate blocking temperature, but stable endpoints combined with great circles do isolate a characteristic high‐temperature component carried by magnetite as indicated by unblocking temperatures and rock magnetic experiments. After tilt correction the formation mean calculated from pillow basalt sites only is D = 339°, I = −33° (k = 12, α95 = 14°, n = 11), with a corresponding paleopole of 23°N, 133°E (δp = 9°, δm = 16°), calculated for an average site location of 46.5°N, 291°E. When both pillow basalt and gabbro sites are used to calculate a tilt‐corrected formation mean, the result is D =339°, I = −28° (k = 12, α95 = 12°, n = 14), and this characteristic component passes a tectonic tilt test. The calculated paleopole (26°N, 134°E; δp = 7°, δm = 13°) from pillows and gabbros falls near (but not on) the Ordovician apparent polar wander path segment for North America and yields a paleolatitude (18°± 9°S) similar to that of the Lunksoos Terrane which is located farther to the southeast within the CMB of Maine. The Laurentian margin at this time was located at paleolatitudes of 15–20°S. The paleogeographic implication of this result is that the Munsungun Terrane was located near the Laurentian margin within the Iapetus Ocean and this near‐continent location is consistent with a marginal or back arc basin tectonic environment for the extrusion of the Bluffer Pond volcanics.