Paleomagnetic study of the shale to slate transition in the Ordovician Martinsburg Formation at Lehigh Gap, Pennsylvania reveals that the remanence of the relatively undeformed shales predates cleavage formation, and that the slates were remagnetized during cleavage development. The shales display two components of remanence. Component 1 (C1) was the only magnetization direction in the majority of the shales, and the first-removed component in 10 shale specimens. Component 2 (C2) is observed as a single-component remanence with a low (300--400¿C) unblocking temperature range in samples from one shale station and as a second-removed component in 10 shale specimens. The mean directions are (declination/inclination) 347¿/-04¿ (346¿/-47¿) in situ (tilt-corrected) for C1 and 80¿/-63¿ (135¿/-40¿) for C2. The pencil slates and well-cleaved slates both display single component remanences.

The mean directions are 198¿/-12¿ (201¿/26¿) for the pencil slates and 199¿/24¿ (225¿/54¿) for the well-cleaved slates. Using the relationship between cleavage development and remagnetization in the slates, the timing of cleavage development and folding, and the regional geology, we can constrain the relative ages of the remanences as follows: (1) the shale C1 direction is most likely prefolding and represents a primary upper Ordovician direction, (2) the slate directions are younger in age than the shale directions and are probably of Permo-Carboniferous (Alleghenian) age. These results show that remagnetization occurred during development of slaty cleavage in these rocks. We interpret the mechanism for this remagnetization to have been strain-induced dissolution and neocrystallization of magnetite. The additional difference in the characteristic directions of the pencil slates and well-cleaved slates may be attributed to strain reorientation of the pencil slate remanence after new growth of magnetite. The unknown temporal relationship between dissolution and neocrystallization and strain precludes any quantitative attempt to correct remanence for strain in these samples. Our study shows that paleomagnetic studies of deformed sediments must consider the possibility of dissolution and neocrystallization of remanence carriers, in addition to strain-induced rotation. ¿ American Geophysical Union 1993