New Zealand contains many sequences of tectonically uplifted Miocene/Pliocene marine sediments, which have been central to the development of a biostratigraphic framework for temperate southern latitudes. Chronological control is almost invariably dependent on magnetostratigraphy. Interpretation of the natural remanent magnetization (NRM) of these sediments is, however, notoriously difficult because of the weak magnetization, secondary components that resist demagnetization, and thermal instability of the clay minerals. A new paleomagnetic study of Wanganui Basin sediments has yielded unusually well defined demagnetization data from which systematic characterization, explanation, and interpretation of the multicomponent nature of the NRM have been possible. In general, the magnetization consists of three components: (1) a low (250 ¿C) Tb component that is considered to be diagenetic in origin. This interpretation of the intermediate and high Tb components is supported by rock magnetic data which show that specimens carrying strong high Tb components have a distinctly different spectrum of ferrimagnetic grains, specifically a population with much higher coercivities than is seen in specimens which lack a high Tb component. It is argued that these high-coercivity, high Tb grains are secondary in origin, leading to the conclusion that the intermediate Tb component is the primary detrital magnetization. Consistency of the paleomagnetic data within and between sites also supports this conclusion. It is suggested that failure to recognize these three components has, in some cases in the past, led to misidentification of the primary magnetization. Using this hypothesis of the multi-component magnetization, the Turakina River data of McGuire <1989> have been reinterpreted, yielding a high-resolution magnetostratigraphy from the upper Gilbert, Chron 3n, (circa 4 Ma) to the lower Matuyama, Chron 2r. The correlation of this magnetostratigraphy with the goemagnetic polarity timescale, is supported by biostratigraphic constraints. ¿ 2001 American Geophysical Union