been used include Waihi-Waimata, Whangamata, Cooks Beach-Purangi, Awana-Te Ahumata on Great Barrier Island, and Fanal Island. Other places where obsidian has been found (for example, Tairua, Maratoto and Hahei) may also have been exploited but no evidence has so far been found in sites.
In the majority of excavated sites where the obsidian assemblage has been analysed, the Mayor Island source is represented. This is not surprising—the stone is consistently of high quality and flakes easily. However the proportion of Mayor Island to Coromandel sources varies considerably depending on the age of the site. All of the early sites (with the exception of Whitipirorua at Onemana) had a predominance of Mayor Island obsidian, while later sites received their obsidian from a wider range of sources. For instance Kauri Point pa at Birkenhead had obsidian from Mayor Island, inland North Island, Fanal Island, Northland and Coromandel (Prickett 1989, Davidson 1990:11–12). Whitipirorua went against the trend, with an estimated 80% from the source found in the clay behind the beach, and exposed in stream beds on the Whangamata peninsula. By contrast, the Whangamata Wharf site, which was also occupied in the early period of settlement, had a predominance of Mayor Island obsidian.
Basalt
Fine grained basalt used for making adzes is found only at Tahanga, Opito, on the eastern side of the Coromandel Peninsula. It was a very important stone resource and adzes made of the distinctive stone have been found over the upper half of the North Island (Fig. 13). Only the metasomatised argillites from the Nelson-d'Urville Island area had a distribution exceeding that of Tahanga basalt.
The Tahanga source covers an area of approximately 6o hectares, and consists of a number of outcrops of weathered boulder piles on the slopes of the hill (Moore 1982:35). Exposures also occur on the beach. It is possible to determine which locality some finished adzes have come from: for instance the stone from the beach is of a consistently darker colour than that on the hill itself.
In several locations there are depressions in the ground surface adjacent to the stone piles suggesting there may have been some digging or quarrying to obtain less weathered boulders. At T10/166 a large outcrop may have had blocks prised off by taking advantage of natural fracturing planes in the rock (Moore 1982:36, 38).
Dense scatters of flakes of a large size are to be found within the boulder piles, evidence of the splitting open of the boulders and initial shaping of the adze (called a preform), removing the weathered outer surface in the process. The preforms were taken away to be further shaped and there are a number of places on the east coast of the Coromandel Peninsula where the final stages of adze making took place. Thousands and thousands of smaller flakes are found at these sites extending as far south as Mt Maunganui (and possibly further but there has been no concerted effort to locate early sites in the eastern Bay of Plenty). One site in Whitianga Estuary must represent the fine trimming and
44