Sediments have proved irresistible targets for attempts at determining the relative variations in the Earth's magnetic field because of the possibility of long and continuous sequences that are well dated and have a reasonable global distribution. The assumption underlying paleointensity studies using sedimentary sequences is that sediments retain a record reflecting the strength of the magnetic field when they were deposited. Early theoretical work suggested that because the time required for an assemblage of magnetic particles in water to come into equilibrium with the ambient magnetic field was quite short, no dependence on magnetic field was expected. Nonetheless, a number of experiments showed that sedimentary magnetizations varied in accordance with the field, albeit not always in a simple, linear fashion. Experiments in which the sediments were stirred in the presence of a field (to simulate bioturbation) showed a reasonably linear relationship with the applied field, and these results spurred the hope that variations in the Earth's magnetic field might indeed be recoverable from appropriate sedimentary sequences. Examination of existing paleointensity data sets allows a few general conclusions to be drawn. It appears that sedimentary sequences can and do provide a great deal of information about the variations in relative paleointensity of the Earth's magnetic field. The dynamic range of sedimentary data sets is comparable to those acquired from thermal remanences. Moreover, when compared directly with such independent measures of magnetic field variations as beryllium isotopic ratios and thermally blocked remanences, there is considerable agreement among the various records. When viewed over timescales of hundreds to thousands of years, relative paleointensity data sets from more than a few thousand kilometers apart bear little resemblance to one another, suggesting that they are dominated by nondipole field behavior. When viewed over timescales of a few tens of thousands to hundreds of thousands of years, however, the records show coherence over large distances (at least thousands of kilometers) and may reflect changes in the dipole field. On the basis of a sequence spanning the Brunhes and Matuyama chrons, the magnetic field has oscillated with a period of about 40 ka for the last few hundred thousand years, but these oscillations are not clear in the record prior to about 300 ka; thus they are probably not an inherent feature in the geomagnetic field, and the correspondence of the period of oscillation to that of obliquity is probably coincidence.