Twenty-nine samples collected from an 80-m section of the Trenton Limestone give almost vertical directions of natural remanent magnetization (NRM), but these become nearly horizontally directed after magnetic cleaning (D = 173°, I = 18°, α95 = 5°). The result is confirmed by thermal demagnetization experiments, and the consequent low latitude suggested for the time of deposition is supported by the paleoclimatic evidence. Samples collected from a penecontemporaneous slump bed demonstrate that both the NRM and cleaned directions are postdepositional. Magnetic-anisotropy measurements show good agreement between the mean principal axis of maximum susceptibility and the cleaned remanence direction, suggesting that the alignment of the long axes of magnetic grains was controlled by the ancient magnetic field direction. This can only have arisen through a postdepositional precompaction in situ rotation of the magnetic grains, a process originally suggested by Irving and Major (1964) from laboratory deposition experiments. This study represents the first positive demonstration that such a process occurs naturally.
The Ordovician paleomagnetic pole position for North America deduced from these rocks (lat 36° S., long 66° W.) is in complete conflict with the predictions of the remagnetization hypothesis, which requires near-polar conditions at that time. The result, however, is in excellent agreement with that obtained from igneous rocks of similar age from the Front Range of Colorado and shows that the Ordovician pole position lies close to the late Paleozoic position. A consequence of this conclusion is the existence of an Ordovician proto-Atlantic Ocean, as has been suggested independently from geological evidence.