The paleomagnetic reliability of the remanent magnetism of the Pleistocene Seminary Till is examined. The lack of correlation in remanence direction between two profiles 0.7 m apart in the till, as well as the large within-core variation of 13° between specimen pairs, suggests the possibility that the sampling procedures cause significant errors. Laboratory experiments on several kinds of un-consolidated glacial sediments show significant thixotropic remanence resetting toward the ambient magnetic field caused by hammer shocks during sampling. Significantly different and significantly higher precision results are obtained from cores that have been drilled by diamond bit from oriented blocks of till than are obtained by driving polycarbonate plastic tubes into the till. Plastic-tube cores also have a distorted magnetic fabric caused by flow deformation. The six tills studied to date in Ontario, including the Seminary Till, yield far-sided pole positions reflecting an average 13° shallowing of the remanence inclination. For normally polarized till, the primary DRM inclination error substantially exceeds 13° but is reduced to this figure by secondary thixotropic resetting. For reversely polarized till, the primary DRM inclination error is further increased by the addition of the resetting error.