A geochemical and biostratigraphic approach has been applied to investigate the spatial and stratigraphic variability of Palaeogene sandstones from key wells in Taranaki Basin, New Zealand. Chronostratigraphic control is predominantly based on miospore zonation, while differences in the composition of Paleocene and Eocene sandstones are supported by geochemical evidence. Stratigraphic changes are manifested by a significant decrease in Na2O across the New Zealand miospore PM3b/MH1 early Eocene zonal boundary, at approximately 53.5 Ma. The change in Na2O is associated with a decrease in baseline concentrations of many other major (MnO, CaO, TiO2) and trace elements, and is interpreted to reflect a significant change in sandstone maturity. Paleocene sandstones are characterized by abundant plagioclase (albite and locally Na–Ca plagioclase), significant biotite and a range of heavy minerals, while Eocene sandstones are typically quartzose, with K-feldspar dominant over plagioclase, low mica contents and rare heavy minerals comprising a resistant suite. This change could reflect a change in provenance from local plutonic basement during the Paleocene Epoch to relatively quartz- and K-feldspar-rich granitic sources during Eocene time. However, significant quartz enrichment of Eocene sediment was also likely due to transportation reworking/winnowing along the palaeoshoreface and enhanced chemical weathering, driven in part by long-term global warming associated with the Early Eocene Climatic Optimum. The broad-ranging changes in major-element composition overprint local variations in sediment provenance, which are only detectable from the immobile trace-element geochemistry.