- 337

Obsidian has attracted the attention of Pacific archaeologists in recent years because of its widespread occurrence on early sites in island Melanesia. Of particular interest has been the possibility of outlining prehistoric exchange networks (e.g. Smith, Ward and Ambrose 1977, White, Downie and Ambrose 1978; Ambrose 1976). Element analyses have shown that most of the archaeological specimens originated from sources in the Manus and West New Britain Provinces of Papua New Guinea (Key 1968, 1969; Smith 1974; Bird et al. 1978; Duerden et al. 1979). Obsidian from the West New Britain source, Talasea, was transported several thousand kilometres about 3,000-4,000 years ago (Ambrose and Green 1972), was imported to New Ireland 6,500-7,000 years ago (White, Downie and Ambrose 1978; Downie and White 1979), and within New Britain itself may have been exploited for more than 11,000 years (Specht, Lilley and Normu 1981). Talasea obsidian thus has a position of some importance in Pacific prehistory.

While much attention has been directed to the archaeological occurrences of Talasea obsidian, little has been published about the source area itself or the exploitation and use of the obsidian in recent times. This paper seeks to redress this imbalance. I review first the likelihood that Talasea is the only area in New Britain where obsidian occurs naturally. I then describe specific source localities which I recorded during visits to Talasea in 1973 and 1974. This is followed by a summary of the uses and exchange of obsidian in recent times, drawing upon published accounts and my own unpublished field notes made in several parts of New Britain between 1965 and 1980. 1 I then consider the possible relation between obsidian distribution, settlement location and local oral history. Finally, I review some of the factors which may have contributed to the longstanding and widespread demand for Talasea obsidian.

The field data were collected from male informants, many of whom had used obsidian at some time, or who had seen it in use by others. These data and those on local oral history are not exhaustive. Obsidian - 338 tools have now been supplemented by metal forms and, in many cases, the practices which required their use have ceased. What follows, then, essentially belongs to the past.

New Britain is crescent-shaped, about 590km long and variable in width. Its area is about 37,300km2 (Figure 1). Administratively the island is divided into the two provinces of East and West New Britain within the independent state of Papua New Guinea.

Whereas the southern side of the island is formed mainly from limestones, the northern side is dominated by products from volcanoes of Pleistocene and Recent age. These volcanoes belong to the Bismarck Volcanic Arc which runs through the islands off the north coast of New Guinea and through the north coast of New Britain (Ryburn 1975: 13).

FIGURE 1A

The New Guinea region

- 339

FIGURE 1B

The Talasea area, showing locations of obsidian sources and major settlements: 1: voko-e-Balive; 2: Pilu; 3: Polongela Hill; 4: Mt. Bao; 5: Kelepu; 6: Polosiviriri—Matanavoko; 7: Bitokara Mission; 8: Bamba; 9: Schaumann Island.

This arc can be divided into eastern and western sections. The Bali-Vitu Islands, the Willaumez Peninsula and the volcanoes of the north coast between the Willaumez and Gazelle Peninsulas belong to the eastern section (Johnson 1976: 102). These volcanoes form the Kimbe Volcanics, - 340 which consist of lava flows, high-level intrusives, pyroclastics and reworked pyroclastics. They are formed over the New Britain Benioff Zone, which dips steeply northwards, and hence their rock types vary systematically across the arc (Johnson 1976: 107-8).

Within this eastern section, Ulawun volcano at the eastern end appears to be the most active one today (Johnson, Davies and White 1972). Pago, on Cape Hoskins Peninsula, was active between 1914-1918 (Blake and McDougall 1973), and Makalia in Dakataua caldera at the northern end of the Willaumez Peninsula seems to have been active about 1880-1890 (Branch 1967: 22). There are extensive areas of thermal activity on Mt Garbuna and near Garu village at the southern end of the Willaumez Peninsula, and at Pangalu and Talasea Station on Garua Harbour (Lowder and Carmichael 1970: 23).

Locally along the north coast there are small areas of uplifted coral limestones, and on each side of Willaumez Peninsula there are alluvial deposits derived from the volcanic products.

Obsidian is a general term for volcanic glasses of high silica content formed by the supercooling of high viscosity magma which prevents crystallisation. Within New Britain volcanic glasses have been reported from several areas and the possibility of additional occurrences cannot be dismissed at this stage.

A volcanic glass of distinctive composition has been reported from the Rabaul caldera (Key 1968: 1969), but because this occurs in small cobbles it is unsuited to artefact manufacture. On the Cape Hoskins Peninsula, Blake and Ewart (1974: 323) have reported two volcanic glasses. That from Witori volcano is a rhyolitic welded tuff, up to 10cm thick, which is a colourless to pale brown clear glass. The other is a clear, colourless pitchstone from Buru volcano. Neither of these appears to have been used for artefacts at any time. During a study of the volcanoes of Cape Hoskins, obsidian flakes were found in a soil buried under more than 20m of tephra on the south-west flank of Witori. This horizon has been dated (Gak-3076) to 2,590 yrs ± 250bp (Blake and McDougall 1973: 202; Blake 1976: 193). Analysis of some of these flakes has shown them to differ from obsidians of the Talasea area (Duerden et al. 1979: Table 1, under 'Waisisi'). Ambrose (1979) has raised the possibility of an obsidian occurrence on Cape Hoskins which may have been buried by substantial deposits of volcanic products since the first millennium B.C. In 1979 a man from the Gasmata area on the south coast of New Britain informed me that his father obtained obsidian formerly from a locality somewhere in the middle of the island, between Kimbe and Gasmata, and not from - 341 Talasea. This account would be compatible with an additional source as suggested by Ambrose.

A possible obsidian source may exist in the Hixon Bay-Ulamona-Lolobau Island area. In 1879 the explorer Wilfred Powell visited this area and reported about the eastern side of Hixon Bay that "The whole of this part of the country seems to be a field of obsidian, huge blocks of it lying on the shores like lumps of glass" (Powell 1883: 217). 2 Today, this area has tephras on the coast, with only a small section of rocks exposed in the Rangambol Point area. These rocks belong to the Merai Volcanics Unit and are most unlikely to contain obsidians (Davies 1973: 10). Just west of Hixon Bay, in the Cape Deschamp area, Powell reported that "some natives brought off great lumps of grey obsidian" (Powell 1883: 219). This area is at the base of Likuruanga volcano (North Son); this, too, is unlikely to have produced obsidian (Davies 1973; Johnson 1970). A source for obsidian in the area of Ulawun volcano, slightly further south and west, is improbable, since dacitic and rhyolitic rocks do not occur there (Johnson, Davies and White 1972: 4).

In 1979 I visited Sule village near Ulamona, just west of Cape Deschamp, to obtain local views on Powell's reference; time and circumstances prevented a visit to Hixon Bay itself. Older men of Sule unanimously agreed that there is no naturally occurring obsidian either in the Cape Deschamp area or to the east of it. Formerly, they obtained obsidian from the Talasea area. A young man stated that he had seen obsidian inland from Sule, but the older men declared that he had only seen old village sites with flakes from Talasea. Powell may have seen "a field of obsidian" which has since been concealed by tephra or alluvium. If this were so, it is surprising that the event is not remembered, since it would have occurred within the last 100 years and historical details older than that are well remembered.

Although a mainland origin for obsidian in the Hixon Bay-Ulamona area seems improbable, a source in the adjacent islands cannot be dismissed. The rocks of Lolobau and neighbouring islands range from olivine-bearing basic lavas to "dominantly glassy pitchstones of dacitic and possibly rhyolitic composition" (Johnson 1970, and cited by Davies 1973: 16). In 1979 I visited briefly the south shore of Lolobau and collected four obsidian flakes on the beach of Mauga Plantation. Elemental analysis suggests that these fit the range of obsidians known from the Talasea area (P. Duerden and W. Ambrose, personal communication).

Present indications are, therefore, that the obsidian at Talasea is the main, but possible not the sole, occurrence within New Britain of material suitable for artefact manufacture.

- 342

The peninsula lies along an approximately north-south line on the north coast of New Britain, just west of the provincial centre of Kimbe, and covers about 2,700km2. In 1973 it had a population of about 6,000 (Papua New Guinea 1973: 263), substantially more than the figure of about 2,700 recorded in 1928 (Kroll 1938: 371). Most of the population speak the Bola language; to avoid confusion with Bola village, these people will be referred to as the Bakovi, a name freely used among themselves. At the northern end of the peninsula are the Bulu speakers, who live in the two main villages of Bulu Daba and Bulu Muri. The Bola and Bulu languages, together with those of the Bali-Vitu Islands to the north-west and those of the north coast to the east, as far as the base of the Gazelle Peninsula, belong to the Kimbe family of Austronesian languages (Chowning 1976). According to Chowning (1978: 297), this linguistic unity is reflected in a broader cultural unity, though this has been much modified in recent years by extensive contact with Bariai family speakers to the west.

There are 11 major volcanoes on the Willaumez Peninsula, in addition to Dakataua caldera and several cinder cones and rhyolitic extrusions (Lowder and Carmichael 1970: 18). The highest peak is Wangore (also known as Bola) at 1200m, about 20km north from Talasea Station. Oral history records an eruption of Wangore which closed a passage between Wangore and Dakataua; this story is known widely throughout the peninsula (Goodenough 1951; Branch 1967: 22; Specht 1981).

The volcanic rocks of the peninsula are mainly andesites, with lesser quantities of basaltic andesite, basalt and dacite, and a small percentage of rhyolite (Lowder and Carmichael 1970: 24). The rhyolite occurs mainly around Talasea, mainly as a buff pumiceous rock. Less common is black obsidian which occurs either as a thick carapace on lava flows, or as thin bands 1mm to 5cm thick (Lowder and Carmichael 1970: 19). It is this obsidian which is of interest to archaeologists and with which this paper is concerned.

Obsidian is not confined to a single occurrence at Talasea, but is available at several points over an area of many square kilometres. In 1974 I attempted to visit all localities which were remembered by the local Bakovi people to have been sources of obsidian for artefacts, but did not try to plot the overall distribution of obsidian around Talasea. Six main areas were remembered, including thirteen specific source localities. Two other areas were recorded, from my personal observations, at Bitokara Mission and on Schaumann Island.

- 343

Obsidian appears to be distributed discontinuously from Voganakai in the north to Kumeraki in the south on both the coast and inland. This distribution can be divided into two groups based on the nature of the sources. The northern group includes Voganakai, Bamba and Schaumann Island, where obsidian occurs as thick banded flows at beach level. The southern group includes Bitokara Mission, Volupai-Liapo-Buluwara, Waru-Dire, Kumeraki and Kumuvavo; each of these sources is inland in gullies where obsidian occurs as boulders of various sizes and is usually obtained by digging through a topsoil of volcanic ash.

(a) Voganakai: Voganakai village is a beach-side settlement situated on the west coast of the peninsula. About 1km south of the village are two named areas from which obsidian was obtained. The first, moving south from the village, is a flow into which a small mine has been excavated. This is known as voko e Balive, 'Balive's obsidian' (archaeological site FEF) and is owned by the Duku lineage of Voganakai. The mine penetrates the flow horizontally; at the entrance it is 90cm high and 30-60cm wide. Inside it runs back 160cm, with a maximum internal width of 100cm. The mine is said to have been worked by lighting a fire inside to heat the obsidian. Cold water was then thrown on to the obsidian, causing small blocks to break away. The obsidian is black with pink spherules. This is the only source recorded in 1974 for which a special extractive technique was used.

Several hundred metres south from voko e Balive is the area of Pilu, where large boulders of obsidian occur for about 300m along the beach. This obsidian is banded and varies in colour from black to dark grey. Flakes and chunks have been struck from some of these boulders. Samples from these sources have been analysed (Duerden et al. 1979: 2352, table 1).

(b) Bamba: Bamba village stands on a broad beach flat 2 to 3km west from Talasea Station. On the eastern edge of the village, a hill slope falls directly on to the inter-tidal zone (the present road along this section has been built up on the inter-tidal flat). At the base of the slope is exposed part of a thick, banded obsidian flow. The obsidian is basically grey, with some black and (rarely) brown bands; part of the flow has hydrated to form perlite, a soft and friable form quite unsuited to artefact manufacture. The inter-tidal zone in front of the exposure is thickly covered with blocks and large flakes of obsidian, among which was found a sherd of Lapita pottery. The exposure and inter-tidal area have been allocated archaeological site code FDK.

(c) Schaumann Island: This is a small island to the east of Garua Island, - 344 owned by Pangalu village. The island is about 250m long and 100m wide at the maximum point, with a maximum height of about 20m. At the northern end an obsidian flow of about 50m length is exposed horizontally on the beach. The flow is brecciated, with grey to black cobbles of obsidian. While this limits the size of flakes and blocks which can be removed, pieces have been removed by percussion.

This flow was not located through informant testimony; at no time during the 1974 season was Schaumann Island cited as an obsidian source. Perhaps the easy availability of better quality obsidian on the peninsula meant that this source was little used. Elemental analysis has shown this obsidian to differ markedly in some elements from other analysed samples from Talasea, though it is closest to that exposed at Bamba (Duerden et al. 1979: 2352, table 1, where Schaumann Island is listed as 'Garua').

(a)Bitokara Mission: The mission is situated on the hilltop above the Bamba source. On the southern side of the mission is a path to the villages of Dire and Waru. Obsidian pieces can be found at many points along this path near the mission. Obsidian flakes and blocks are also common on the north side of the mission, especially near the crest of the hill. Many of the flakes have been struck.

As the distribution of obsidian on the ground surface is discontinuous, the four areas over which flakes and pieces were found have been allocated separate site codes (FDV, FDW, FDX, FDY). The quantity of obsidian, however, suggests that it may occur naturally on the hill top, though now partially concealed by volcanic ashes, rather than having been brought up from beach level as blocks. Part of a flow is exposed on the surface of the road from the beach to the mission.

(b) Volupai-Liapo: Volupai is a coastal village on the western side of the peninsula. Liapo is situated several kilometres inland from Volupai and south of Humgari Mountain. The two villages formerly obtained obsidian from several localities between Humgari and Bao Mountain. According to one Volupai informant, both villages obtained obsidian from Bao, though it is not clear whether they held separate collection areas.

Liapo had two main sources, one on Bao and the other at the southern base of Humgari. The Humgari source was not visited, but it was described as an area with boulders of obsidian on or under the ground surface. Bao mountain was visited primarily to inspect an old settlement site and a stone important in local history. In the course of the visit my guides indicated several areas on the northern flank where obsidian might be obtained by digging.

- 345

Although Volupai village is said to have used a Bao source, the main collection areas are on Polongela hill, several kilometres inland and south-east from Volupai. The hill has two crests from which several gullies run down towards Volupai and Buluwara villages. Obsidian blocks are found in these gullies and on the slopes above them.

At the northern end of Polongela is a narrow, steep-sided gully, Ulemono, which runs down to Volupai. Blocks and flakes occur on the gully sides and floor, but preferred material is obtained by digging in the gully floor. The first obsidian to be traded north to the Bulu villages is said to have come from this gully. This source, allocated site code FDE, is controlled by Hipamanga lineage of Volupai.

South of Gulemono, on the main hillslope of Polongela, obsidian flakes occur on the ground surface and blocks can be found in the volcanic ash covering the hill. Below this area is another gully, Matangalovekea, which runs south-west to join the main obsidian-bearing gully of Matanavoko (site FDU). The main hill slope area is controlled by the Bagele lineage, and Matanavoko and Matangalovekea by the Ngalagundi lineage of Volupai.

Matanavoko gully runs down from the southern crest of the hill (Kulavu) to Buluwara village on the coast. The gully is 20 to 30m wide and 8 to 10m deep; in its base are blocks of obsidian weighing 12 kilograms and more. This gully is reputed to have been the main obsidian source for Volupai (site FDF).

(c) Waru-Dire: These are inland villages south and east of Bao at approximately 200m altitude. The two villages include several smaller settlements, one of which is Polosiviriri in the Dire area. Near Polosiviriri is a hillside, known as Matanavoko, with small obsidian blocks and flakes on its surface; larger obsidian blocks are found below ground surface (site FEE). The area is controlled by the Marenge, Tenge and Pake lineages of Waru. The name, the same as that at Volupai, can be loosely translated as "the primary source of obsidian".

(d) Kumeraki: This is another inland group of settlements, south of Waru, at approximately 250m above sea level. Of three named obsidian sources, one only, Kelepu, was visited. The original source was at Matane Ningekira. When obsidian was first found there it was named after the blackened teeth of a leading big-man, Kira. This source seems to have passed out of use a long time ago. Another source, Kaputu, is said to be on the north bank of the Lalundupi River, to the west of Narunegeru settlement.

The most important source today is at Kelepu, near Kulagia hamlet. This source was formerly controlled by the Mota lineage, but is now owned by the Takale lineage. The leader of the Takale lineage in 1974- 346 was Anton Gala, of Kulagia, who stated that Kelepu obsidian flakes were better and yielded sharper edges than that of the Bitokara-Bamba area, because it is pure black. Waru informants expressed a similar view.

Kelepu (site FED) is about 10 minutes' walk north from Kulagia. The source is a gully about 4m deep and 3m wide at the base, with almost vertical sides which are said to be natural, but their straightness suggests that they may have been cut by man. Boulders of obsidian up to 50cm long lie on the gully floor, but most pieces are 20cm or less in length. The best obsidian is said to be found by digging down a metre or so.

(e) Kumuvavo: This is a beach settlement on the west coast, south from Buluwara. It is said to have its obsidian source at Kalogo, a gully about 1.5km inland from the beach. The gully varies in width from 3 to 8m, and is up to 1.5m deep; starting in the Kumeraki area, it runs down to the coast near Kumuvavo. Small boulders of obsidian are visible on the gully floor; larger ones are found by digging.

Little specific information is available about techniques of extraction or reduction. With the exception of the fire and water technique at the voko e Balive source at Voganakai, the flow sources appear to have been worked simply by striking off pieces where suitable material was visible. At sources where obsidian was collected from the ground surface or by digging for suitable pieces, no special extractive techniques were necessary. Pieces were selected according to their size and general suitability for flaking.

The collection of obsidian appears to have been mainly the responsibility of males, though females were not banned from the sources at Volupai. Access to a source appears to have been relatively uncontrolled; a man could exploit the source owned by his lineage, and did not require the special permission of his lineage head.

Reduction of the selected pieces took place at villages. Informants agreed that the objective was to remove a sharp-edged flake from the core, which might be a block or large flake of obsidian. Core preparation was limited to the removal of weathered surfaces; striking platforms were not specially prepared, nor were there attempts to control the shape of flakes removed. Flake size and shape were probably less important than the sharpness of the edge for the task in hand. Retouch was not practised, since this would have added little, if anything, to edge sharpness (cf. Parkinson, 1907: 240). In areas that imported obsidian, especially as distance from the sources increased, efficient core flaking was probably of greater importance. In east New Britain and on the New Guinea mainland only small flakes generally less than 20mm long were - 347 removed from the core, presumably to maximise the number of flakes from each piece of obsidian.

In the Talasea area, the lack of specific forms in the recent past is in marked contrast to finds from several archaeological sites in that area. In 1972, J. Kamminga identified an extensive obsidian workshop on the inter-tidal zone in front of Bamba village (site FCH) with many retouched pieces, including tanged points and bifacially retouched points (Specht 1973; Ambrose 1976: 364; cf. Casey 1939: 149). I recorded other sites with similar tools in 1973 and 1974 (Specht 1974: 305). Such forms are lacking from the ethnographic record; their age and function are not known.

Information on the use of artefacts made from Talasea obsidian, both in the Talasea area and beyond, is generally brief and lacking in detail. Two main groups of uses can be defined: those relating to the human body, and those concerned with artefact production. A third, minor group includes miscellaneous uses.

These required a thin, sharp cutting edge. The group can be further divided into uses of a medical nature, and those of a decorative or initiatory nature. Some of these were very widespread, while others were of very limited distribution.

Blood-letting by skin cutting with an obsidian flake was mentioned at several Bakovi villages from Voganakai to Waru. Further afield, obsidian was used for this purpose among the Sengseng near Kandrian (Chowning 1978: 217). Among the Bakovi, obsidian flakes were used to cut away the flesh around a spear wound to permit removal of the spear point. Unspecified forms of "surgery" are mentioned for the Maenge (Panoff 1969: 7) and for the Tolai (Powell 1883: 165). Among the latter, however, a major use of obsidian was in the process of trephination of the skull.

Trephination is an operation in which part of the skull is cut away to relieve local pressure on the brain. Among the Tolai, the operation was usually carried out in cases of compressed fracture caused by a blow from a slingstone; damage caused by a blow from a stone club was usually regarded as fatal and no operation was carried out (Crump 1901: 167; Parkinson 1907: 109; Brodsky 1939: 2; Ford 1937: 471). The operation, carried out while the patient was unconscious, involved cutting away and removing the damaged bone with obsidian flakes. A similar operation was carried out on New Ireland to cure conditions, such as headaches, not caused by violence (Crump 1901).

Among some societies of the north coast of New Britain, such as the - 348 Bariai, and at the village of Sio on the New Guinea mainland, a baby's umbilical cord was severed with an obsidian flake (Friederici 1912: 44, 139; Harding 1967: 42-3).

The decorative or initiatory uses included the cutting of hair among the Bakovi, the Bariai (Friederici 1912: 139), and at Sio (Harding 1967: 42). Beard shaving with an obsidian flake appears to have been the most common and widespread application of obsidian, being recorded throughout the Tolai, Bakovi, Bulu and Sengseng areas; among the east Nakanai (Powell 1883: 217), the Maenge (Bühler 1946-49: 240), the Kaulong (Goodale 1966: 25), the Bariai (Friederici 1912: 139), the Bali-Vitu Islands (Parkinson 1907: 218), and on the New Guinea mainland at Sio (Harding 1967: 42).

The "initiatory" group of uses includes circumcision, the cutting of ear lobes, tattooing and cicatrisation. Circumcision was widely practised, and the use of an obsidian flake is recorded among the Bakovi, the Maenge (Panoff 1969: 7), the Sulka (Parkinson 1907: 178, 181), the east Nakanai (Powell 1883: 217), the Bariai (Friederici 1912: 139), and the Sio (Harding 1967: 42). The cutting of an ear lobe, presumably as a mark of status, is recorded for the Bakovi, and the Sio (Harding 1967: 42). Skin-cutting with an obsidian flake for tattooing and cicatrisation is reported among the Bakovi, the Sengseng (Chowning 1978: 300), the Sulka (Parkinson 1907: 178, 181), and the Kilenge (Dark 1974: 36).

Harding (1967: 42) suggests that on New Guinea obsidian was not used in artefact production because of its rarity, but on New Britain obsidian tools were used to make artefacts by several societies. The Bakovi used obsidian for cutting turtle shell and for scraping spear shafts, though in the latter case shells were usually employed. Unspecified carving uses are recorded for the Maenge (Panoff 1969: 7), in the Kandrian and Gasmata areas (Bühler 1946-49: 240), and among the east Nakanai (Powell 1883: 217). Parkinson (1907: 240) states that obsidian was important throughout New Britain for wood-working, especially for carving low relief designs on canoe prows and on wooden bowls. On Watom Island, in east New Britain, canoe paddles were scraped with an obsidian tool (Meyer 1911: 265). The only reports of obsidian being hafted are an unsubstantiated statement by a Bakovi informant, referring to another Bakovi village, and a reference by Friederici (1912: 139), who states that occasionally the Bariai would haft a suitable piece of obsidian as an adze.

Unmodified flakes were used for butchering pigs in the Waru area of the Bakovi. Chowning (1978: 300) mentions the inclusion of obsidian in "wealth bundles" among the Lakalai, and in the Bali-Vitu Islands obsid- - 349 ian was one of the several items in bride-price exchanges.

Obsidian was one of many goods exchanged within and beyond the Willaumez Peninsula in recent times. Figure 2 summarises the movement of some goods between groups on the Willaumez Peninsula, based on published sources and field informants. The figure does not show exchange equivalences, and the apparently central position of Waru does not necessarily indicate its importance.

Within Bakovi, exchanges involved raw materials as much as manufactured goods. Garu and Patanga were the main, if not sole, sources for red and white pigments respectively which, like obsidian, would be source-specific. In exchanges between non-Bakovi and Bakovi, manufactured goods of less source-specific nature moved against obsidian and pigments. In the Volupai-Vitu exchanges, the manufactured goods, woven mats and shell money, moving out of Volupai may have been obtained initially from the Kove.

In the Kumeraki area a block of obsidian 15-20cm in diameter was

FIGURE 2

The movement of goods through exchange within the Willaumez Peninsula, West New Britain.

- 350

apparently the standard exchange unit, similar to that of the Maenge (Panoff 1969: 7). Such blocks eventually reached the mainland of New Guinea, apparently losing little of their bulk en route. According to Friederici (1912: 138-9), obsidian blocks of one hand-span length and width were worth one clay pot on the mainland, and a cube of that size could attract three pots. In more recent times such blocks could be worth 12 pots at Sio (Harding 1967: 2-3, 129-30). In the exchanges across Vitiaz Strait obsidian was an important link in the conversion of low-value goods from one side of the Strait into high-value goods from the other side (Harding 1967: 139).

Obsidian did not simply retain its value as distance from Talasea increased, but the value increased greatly and "several small and roughly shaped flakes" could be traded by the Sio to interior people for small pigs (Harding 1967: 42-3).

On New Britain obsidian from Talasea was carried eastwards along the north coast, then across the island to the Maenge on the south side (Panoff 1969: 7). Among the Maenge, a block of obsidian about 20cm long was worth a large bundle of tobacco leaves.

According to Tolai on Watom Island in 1965-66, the islanders sailed to Talasea itself, but Chowning (1978: 299) states that the Tolai obtained obsidian from the east Nakanai. Direct visits to Talasea appear to have started only since pacification. The Kove did visit the Talasea area. Depending on whether relations with the Bakovi were hostile or friendly, they would either wait on the beach and send a message to the Kumeraki area villages, or walk inland to them. The specifics of the transactions are not known.

Chowning (1978: 298) observed that the Willaumez Peninsula "seems to have acted almost as a barrier" between the Kove and Nakanai, with only disc-shaped shell money passing across or around the peninsula. From another point of view, the peninsula could be regarded as the starting point of two exchange chains with many links extending east and west, with unique products of the peninsula—obsidian and certain pigments—serving as major elements in the maintenance of those chains. The Bakovi enjoyed a privileged position through their monopolistic control of these raw materials.

I recorded oral traditions around Talasea in 1973-4 which suggest a complex history of population movements along the Willaumez Peninsula, from south to north, in the recent past.

Histories of some founding lineages of nine present-day villages trace a common descent from an ancestral place known as Gamai (Figure 3).

- 351

FIGURE 3

Family Tree. Settlement movements implied by oral traditions from the Talasea area, West New Britain. GAMAI, 'NAKANAI', GANEBOKU, NGALOLO, GARU, BUNGA, KAPOGARA, KOVE-N.G.MAINLAND (KALAPIAI), BILAPURAGI, GANEBOKU, KAMBAKA, KUMERAKI, WARU, WARU/DIRE, BAGUM, VOLUPAI, BAO, LIAPO, BULU, VENGO, KAVUTU, WARI, BAGUM KENDE, VITU, BAGUM, BAOMELI, LANGEA, VOLUPAI, KAMBILI, BUNGA, LANGEA, BULUWARA, BAIBAKO, GALAU, BAMBA, DIRE, KUMERAKI, SIRALO, RALA, KULUKAVUVU (WARU), WARU, KUMERAKI, TANGARAGALUPE, PANGALU, GALAU, BILO, BAMBA, WUNDI, TOLOKAMBA, BILO, TARAKULU, VOGANAKAI, BAMBA, BANGO, PANGALU

- 352

Gamai appears to have been situated in the south-central part of the Willaumez Peninsula. From Gamai, some groups moved eastwards towards, perhaps into, the Nakanai area; others moved westwards, some joining the Kove, others finally reaching the New Guinea mainland. The main thrust appears to have been northwards towards the Talasea area to occupy the areas of the major obsidian sources known today. The lineage histories are not always clear whether the movements were into previously unoccupied areas; some are couched in mythological terms and describe prior populations of women only, while others imply entry into unoccupied terrain. Whatever the context, the diagram suggests that the present-day settlement distribution is of relatively recent origin, with seven of the movement histories terminating at specific obsidian sources: Voganakai, Bamba, Volupai-Liapo, Waru-Dire, Kumeraki. The lineage histories do not cite the control of the obsidian or pigment sources as an objective for these movements, but the settlement pattern which today bestows upon the Bakovi an absolute monopoly over the sources may have been created or reinforced at that time.

The relationship between resource distribution and social structure is complex (e.g. Finney 1966 on Tahiti). The Bakovi trace patrilineal descent (Goodenough 1951: 13) and apparently follow nominal patrivirilocal post-marital residence rules. Such a system seems well suited to continuance of lineage control over specific fixed resources such as obsidian and pigments. This would have been particularly important in a social universe in which males controlled the exploitation and use of the resources, which were themselves prominent in exchanges yielding male access to wealth and status items.

The long period of exploitation and widespread prehistoric distribution of Talasea obsidian suggest it may have had certain qualities which made it a desirable good, even though, for each of the uses listed above, other materials would have served equally well. Obsidian, however, offered several special advantages to the Bakovi and their predecessors in the Talasea area. Even after millennia of exploitation, there remain thousands of tonnes of obsidian around Talasea which can be gained with comparatively little effort and without specialised extraction techniques. 3 A block could be exchanged immediately, without elaborate preparation. Unlike most manufactured goods, a piece of obsidian did not embody a heavy investment of labour when exchanged; as such it was probably the most profitable good that could be exchanged.

Obsidian had several other qualities which gave it an advantage in exchanges over manufactured goods and may have assisted its wide - 353 transport to communities beyond New Britain. It can be stored indefinitely without danger of deterioration or damage which would render it useless. Most foodstuffs and goods made from organic materials are susceptible to degradation or damage; to a lesser degree, even clay pots, shell artefacts and ground stone tools may be rendered useless if damaged during transport or use. Obsidian can be destroyed only by reduction to the stage where it is impossible to remove any more flakes. A block of obsidian weighing several kilograms, about the size of the blocks traded from Talasea in recent times, would yield hundreds, perhaps thousands, of flakes each with one or more usable edges. A single block might have a life of several years, and could be worked without specialised reduction techniques but with minimal wastage. Furthermore, a large block could be split and part exchanged for other goods.

Although some of these qualities might seem disadvantageous to those controlling the sources, the transportability of obsidian meant that it could enter distant exchange networks, thus maintaining and perhaps increasing overall demand. Control of the sources was equivalent to the best of "blue chip" investments, perhaps endangered only by the termination or redirecting of exchange networks, or by the introduction of metal tools.

- 354

• AMBROSE, W. R., 1976. "Obsidian and its prehistoric distribution in Melanesia", in N. Barnard (ed.), Scientific Methods of Research in the Study of Ancient Chinese Bronzes and Southeast Asian Metal and other Archaeological Artifacts, pp. 351-78. Melbourne, National Gallery of Victoria.
• —— 1979. The Impermanence of Obsidian Sources in Melanesia. Paper presented to the XIV Pacific Science Congress, Kharbarovsk, U.S.S.R., August, 1979.
• —— and R. C. GREEN, 1972. "First Millennium B.C. transport of obsidian from New Britain to the Solomon Islands." Nature, 237: 31.
• BIRD, J. R., L. H. RUSSELL, M. D. SCOTT and W. R. AMBROSE, 1978. "Obsidian Characterization with Elemental Analysis by Proton Induced gamma-Ray Emission." Analytical Chemistry, 50: 2082-4.
• BLAKE, D. H., 1976. "Pumiceous pyroclastic deposits of Witori volcano, New Britain, Papua New Guinea", in R. W. Johnson (ed.), Volcanism in Australasia, pp. 191-200. Amsterdam, Elsevier.
• —— and A. EWART, 1974. "Petrography and Geochemistry of the Cape Hoskins volcanoes, New Britain, Papua New Guinea." Journal of the Geological Society of Australia, 21: 319-31.
• —— I. MCDOUGALL, 1973. "Ages of the Cape Hoskins volcanoes, New Britain, Papua New Guinea." Journal of the Geological Society of Australia, 20: 199-204.
• BRANCH, C. D., 1967. "Volcanic activity at Lake Dakataua caldera, New Britain", in C. D. Branch, Short Papers from the Volcanological Observatory, Rabaul, New Britain. Bureau of Mineral Resources, Geology and Geophysics of Australia, Report 107, pp. 21-25.
• BRODSKY, I. I., 1938. "The trephiners of Blanche Bay, New Britain, their instruments and methods." British Journal of Surgery, 26: 1-9.
• BÜHLER, A., 1946-9. "Steingeräte, Steinskulpturen und Felszeichnungen aus Melanesien und Polynesien." Anthropos, 41-44: 225-74, 557-606.
• CASEY, D. A., 1939. "Some prehistoric artifacts from the Territory of New Guinea." Memoirs of the National Museum of Victoria, 11: 143-50.
• CHOWNING, A., 1976. "Austronesian Languages: New Britain", in S. A. Wurm (ed.), Austronesian Languages, New Guinea Area Languages and Language Study, pp. 365-86. Canberra, Department of Linguistics, Australian National University.
• —— 1978. "Changes in West New Britain Trading Systems in the Twentieth Century." Mankind, 11: 296-307.
• CRUMP, H. J., 1901. "Trephining in the South Seas." Journal of the Anthropological Institute, 31: 167-73.
• DARK, P. J. C., 1974. Kilenge Art and Life. London, Academy Editions.
• DAVIES, H. L., 1973. Gazelle Peninsula, New Britain—1: 250,000 Geological Series. Bureau of Mineral Resources, Geology and Geophysics of Australia, Explanatory Notes, sheet SB/56-2.
• DOWNIE, J. E. and J. P. WHITE, 1979. "Balof Shelter, New Ireland—Report on a small excavation." Records of The Australian Museum, 31: 763-802.

- 355

• DUERDEN, P., D. D. COHEN, E. CLAYTON. J. R. BIRD, W. R. AMBROSE and B. F. LEACH, 1979. "Elemental Analysis of Thick Obsidian Samples by Proton Induced X-ray Emission Spectrometry." Analytical Chemistry, 51: 2350-4.
• FINNEY, B., 1966. "Resource Distribution and Social Structure in Tahiti." Ethnology, 5: 80-6.
• FINSCH, O., 1888. Samoafahrten. Reisen in Kaiser Wilhelmsland und Englisch-Neu-Guinea in den Jahren 1884 and 1885. Leipzig. Hirt und Sohn.
• FORD, E., 1937. "Trephining in Melanesia." Medical Journal of Australia, 2: 471-7.
• FRIEDERICI, G., 1912. Wissenschaftliche Ergebnisse einer amtlichen Forschungsreise nach dem Bismarck-Archipel im Jahre 1908. II. Beitrage zur Volker-und Sprachenkunde von Deutsch Neuguinea. Mitteilungen aus den Deutschen Schutzgebieten, Erganzungsheft No. 5, Berlin.
• GOODALE, J. C., 1966. "Imlohe and the Mysteries of the Passismanua, Southwest New Britain." Expedition, 8: 20-31.
• GOODENOUGH, W. H., 1951. "Ethnological Reconnaissance in New Guinea."Bulletin of the University Museum, University of Pennsylvania, 17: 5-38.
• HARDING, T. G., 1967. Voyagers of the Vitiaz Strait: A Study of a New Guinea trade system. Seattle, University of Washington Press.
• JOHNSON, R. W., 1970. "Likuruango Volcano, Lolobau Island, and associated volcanic centres, New Britain: Geology and Petrology. Bureau of Mineral Resources, Geology and Geophysics of Australia, Records, 1970/42 (unpub.).
• ——1976. "Late Cainozoic Volcanism and Plate Tectonics at the Southern Margin of the Bismarck Sea", in R. W. Johnson (ed.), Volcanism in Australasia, pp. 101-16. Amsterdam, Elsevier.
• ——, R. A. DAVIES and A. J. R. WHITE, 1972. Ulawun Volcano, New Britain. Bureau of Mineral Resources, Geology and Geophysics of Australia, Bulletin, 142.
• KEY, C. A., 1968. "Trace element identification of the sources of obsidian in an archaeological site in New Guinea." Nature, 219: 360.
• —— 1969. "The identification of New Guinea obsidians." Archaeology and Physical Anthropology in Oceania, 4: 47-55.
• KROLL, H., 1938. "Sagen und Märchen der Bola." Zeitschrift für Ethnologie, 70: 371-425.
• LABILLARDIÈRE, J., 1800. Relation du voyage à la recherche de la Perouse . . .1791, 1792, (1793-4). Paris, Jansen.
• LOWDER, G. G. and I. S. E. CARMICHAEL, 1970. "The Volcanoes and Caldera of Talasea, New Britain: geology and petrology." Bulletin of the Geological Society of America, 81: 17-38.
• MEYER, O., 1911. "Die Schiffahrt bei den Bewohnern von Vuatom (Neupommern, Südsee)." Baessler-Archiv, 1: 257-68.
• PANOFF, M., 1969. Inter-tribal relations of the Maenge people of New Britain. New Guinea Research Bulletin, 30.
• PAPUA NEW GUINEA, 1973. Papua New Guinea— Village Directory 1973. Konedobu, Department of the Chief Minister and Development Administration.

- 356

• PARKINSON, R. R. H., 1907. Dreissig Jahre in der Südsee. Stuttgart, Strecker und Schroeder.
• POWELL, W., 1883. Wanderings in a Wild Country. London, Sampson Low, Marston, Searle and Rivington.
• RYBURN, R. J., 1975. Talasea-Gasmata, New Britain—1: 250,000 Geological Series. Bureau of Mineral Resources, Geology and Geophysics of Australia, Explanatory Notes, sheet SB/56-5 and SB/56-9.
• von SCHLEINITZ, Freiherr, 1896. "Begleitworte zur Karte der Nordküste des westlichen Teils der Insel Neu-Pommern." Zeitschrift der Gesellschaft für Erdkunde zu Berlin, 31: 137-54.
• SMITH, I. E., 1974. Obsidian sources in Papua New Guinea. Archaeology and Physical Anthropology in Oceania, 9: 18-25.
• ——, G. WARD and W. R. AMBROSE, 1977. "Geographic distribution and the characterization of volcanic glasses in Oceania." Archaeology and Physical Anthropology in Oceania, 12: 173-201.
• SPECHT, J., 1973. "Prehistory poses many problems." Australian Natural History, 17: 445-51.
• —— 1974. "Lapita pottery at Talasea, West New Britain, Papua New Guinea." Antiquity, 48: 302-6.
• —— 1981. "Aspects of the oral history of the Bakovi people, West New Britain Province." Oral History.
• ——, I. LILLEY and J. NORMU, 1981. "Radiocarbon dates from West New Britain, Papua New Guinea." Australian Archaeology, 12: 13-15.
• WHITE, J. P., J. E. DOWNIE and W. R. AMBROSE, 1978. "Mid-Recent Human Occupation and Resource Exploitation in the Bismarck Archipelago."Science, 199: 877-79.