Some pa were constructed in the 19th century. The almost constant warfare in the 1820s and 1830s meant people needed to protect themselves. A pa built at Ongare Point by Ngai Te Rangi was attacked by Taraia of Ngati Tamatera in 1842 (Shawcross 1964), and Opita on the Waihou River was re-built in 1842 in anticipation of reprisals for Taraia's actions against Ngai Te Rangi at Ongare Point (Phillips 1994: 412). There is mention in the Land Court records of other pa constructed at this time, for example, on Waiheke (reported in Monin 1992), and Te Kari on the Waihou River re-fortified in 1831 (Phillips
1994:377)
It is apparent from the Land Court records and from European eye-witness accounts that the pa was often the focus of settlement in an area, with an undefended village nearby and gardens scattered through the area. John Nicholas (18171: 394) described several villages on the Wharekawa coast (on the western side of Firth of Thames) in 1815, one of which may be Whakatiwai, with the palisaded pa on the ridge above and gardens on hillslopes in the vicinity At Matatoki, Oruarangi was identified in the Land Court records as the principal pa with Paterangi a satellite pa, and kainga and gardens in the vicinity (Hauraki MB 63:133). By contrast, Captain Cook in Mercury Bay described the well organised Wharetaewa pa with numerous houses, and very little evidence of occupation outside of the pa, although there were people in different parts of the wider bay. (B eaglehole 1955:198-99).
Middens
Middens are the rubbish heaps of a settlement. There are approximately 4300 recorded sites containing midden (Fig. 5). They are made up mainly of shell, but may also contain small amounts of fish and other bone, charcoal, oven stones and other debris such as seeds. Shell deposits are highly visible in disturbed ground and are often the only indication that an occupation site is present. The amount of shell can vary from a few kitfuls to several cubic metres. For instance the middens at Kauri Point pa and Ongare Point pa were just under two metres in depth.
By sorting the midden into its various components and identifying shell and bone material, archaeologists can interpret diet and preferences for food although the information has many limitations.
Because shell middens are highly visible, the absence of this site type from some areas appears significant but can be explained by environmental conditions. Thus a comparison of Fig. 5 with the overall site distribution map (Fig. 3)shows a general absence of shell midden at the northern end of the Coromandel Peninsula and along the upper west coast north of Colville. By contrast almost all of the Coromandel east coast sites contain shell. On the Wharekawa coast middens are generally only present in the coastal sites. Some of these absences can be explained by distance from sources of estuarine and sandy shore shellfish (such as the rocky Moehau coastline on Coromandel) but shellfish have been carried some distance in other places, for example, to Raupa near Paeroa, and cockles are present in middens on Slipper Island.
16