Stone and earthen architecture is nearly ubiquitous in the archaeological record of Pacific islands. The construction of this architecture is tied to a range of socio-political processes, and the temporal patterning of these features is useful for understanding the rate at which populations grew, innovation occurred, and social inequality emerged. Unfortunately, this temporal patterning is poorly understood for many areas of the region, including the Sāmoan archipelago. Here, we describe a project directed toward establishing a robust chronology for the construction of these earthen and stone terraces and linear mounds on Ta‘ū Island. Using recent methodological improvements, we highlight the tempo at which different architectural types were constructed on the island and the implications for understanding demographic expansion and changing land tenure practices in the last 1500 years. This research suggests the construction of architecture was largely confined to the 2nd millennium AD with a small number of terraces plausibly built in the 1st millennium AD. This temporal patterning suggests that a reconfiguration of settlement patterns occurred within West Polynesia as people there moved into other regions of Oceania.

Japanese archaeology benefits from the large number of rescue excavations conducted during recent decades that have led to an unparalleled record of archaeological sites. That record is here put to use to interrogate changing settlement patterns in the north-eastern corner of Tokyo Bay during several millennia of the Jomon period (Early, Middle and Late Jomon: 7000–3220 cal BP). Jomon hunter-gatherer occupation is characterised by large numbers of settlements, some of them substantial in size, containing hundreds of individual pit-house residential units. Detailed analysis of the rank-size distribution of these settlements reveals a pattern in which periods of settlement clumping, with few large settlements, alternate with more dispersed settlement patterns on a regular cycle of approximately 600 years. The regularity of this cycle might suggest a correlation with cycles of climatic change, such as Bond events. Closer scrutiny shows, however, that such a correlation is unconvincing and suggests that cyclical change in Jomon settlement patterns may instead be due to other factors.