Other formats

    TEI XML file   ePub eBook file  


    mail icontwitter iconBlogspot iconrss icon

Ngā Tohuwhenua Mai Te Rangi: A New Zealand Archeology in Aerial Photographs

14 — South Island

page break

South Island

The terrace country at Clarence River, seaward Kai Kōura Range in distance

The terrace country at Clarence River, seaward Kai Kōura Range in distance

The best researched gardening sites at Clarence River lie on the low coastal strip inland of the road and on the margins and south of the prominent bush-filled gully (centre right) in the high terrace. They do not show in this photograph, but are depicted in a photograph elsewhere in this chapter.

page 195

A brief description of the physical setting of sites in the South Island is difficult because of its-size and topographical complexity. The island is characterised by a complex spine of high ranges in the west and low-lying plains in the central eastern portions. The latitudinal range of the island, coupled with the mountainous topography, is such that there are distinct climatic differences between northern and southern parts, and western and eastern. To the west, rainfall is abundant; in the east, seasonally dry conditions prevail and areas of grassland occupy the broad areas of plains. In the south, mean temperatures are lower than in the north. As in the North Island, the coastal strip and river valleys were dominant in determining settlement pattern. Earth ovens from the earliest periods have been frequently recorded and excavated on the outer reaches of estuaries and on river terraces inland. Inland trips to hunt moa or gather stone also appear to have been common.

In this chapter the principal themes to be covered are the restrictions on population growth in the South Island and its effects on the visible archaeological landscape; moa and moa hunting, particularly at the major river mouths such as Wairau Bar and Waitaki River mouth; horticulture in gardens defined by stone rows; page 196
The Wairau estuary and bar

The Wairau estuary and bar

The moa-hunter site complex lies at the inner edge of the bar, by the prominent trees surrounding the homestead just above centre at the modern entrance to the estuary. Old distributary channels, intersected by modern river channels, are prominent above the estuary at lower left. The view is from the south.

page 197

A closer view of the modern entrance to the Wairau estuary. The site lay over the broad evenly grassed area on the inner side of the bar, nearest the camera. The initial finds were made in the course of ploughing this area in the 1940s. The later excavations of Roger Duff and Jim Eyles in the band of ovens show as a rectangular patch with interior mounds near the edge of the estuary. There appears to have been long and continuing erosion of the inner edge of the bar, cutting into the full extent of the site, and leaving a tiny island (middle foreground).

page 198
Wairau Bar

Wairau Bar

The site of Roger Duff and Jim Eyles' excavations at Wairau Bar. The original excavation squares created the distinct rectangular outline while the undulations result from backfilling from adjacent squares.

page 199 the relative lack of substantial pre-European earthwork pā; and the Ngāi Tahu pā of the mid-Canterbury region, attacked by Ngāti Toa under Te Rauparaha in 1829-30.

The principal tribe associated with the South Island is Ngāi Tahu, descended from the great canoes of the eastern North Island, Horouta and Tākitimu. The descent lines are linked with those of Ngāti Porou, Ngāti Kahungūnu and Rangitāne. 1 In the nineteenth century, there were incursions of Taranaki and Waikato iwi into the northern South Island, part of the general southwards movement into Cook Strait. The earliest traditions can be read in the naming of Banks Peninsula, Te Pātaka a Te Rākaihautū, the storehouse of Rākaihautū. Lake Ellesmere immediately to the south was Te Kete Ika ō Rākaihautū, the fish basket of Rākaihautū. Rākaihautū was also the creator and namer of the great southern lakes. 2

The South Island occupies only a single chapter of this book because it has far fewer archaeological sites than the North Island. Sites in the South Island tend to have less visibility in the landscape than in the North Island, and the main reason is the paucity of highly visible pā and horticultural sites of the late pre-European period. I discussed in chapter 2 why the readily visible sites in the north tend to be late in age. In the North Island, horticulture was important because it was essential to sustain large numbers of people. The essential factor making horticulture possible in the north was climate, especially the length of the growing season needed for the pre-European root crops. Production was virtually limitless, provided the effort was made to cultivate. When James Cook and his fellow observers commented on the eastern North Island, they noted the numbers of people. It seems that the horticultural capacity of the North Island made it possible to sustain large numbers of people there late in the pre-European period. At the time of early settlement, when number of immigrants and not natural increase was the relevant factor, the population in the South Island was probably the same as, or even outnumbered, that in the North Island. 3 As time went by the numbers in the north far outgrew those in the south, as the latter reached (and may indeed have lowered) its carrying capacity.

When people first visit a new environment, they may take some time to adjust to its advantages and difficulties. For tropical Polynesians, the difficulties of New Zealand after an arduous voyage must have been great. On arrival, the obvious response is to take food from the easiest, most productive and most accessible sources. In New Zealand, that food was meat, and it came in the form of moa and other land birds, fish and sea mammals. Moa had never been hunted by any other animal, and they were probably naive in their attitude to aggressive human predators. 4 Moa were also abundant in the South Island, particularly in their preferred habitat of grassland plains and open forest—a habitat not as common in the North Island as it was in the South Island. Sea mammal colonies, such as that illustrated in chapter 15, were distributed around the New Zealand coast, but with greater numbers in the south, since these are animals that range in the sub-polar oceans for the most part. 5 The South Island was therefore a very attractive initial habitat for the small numbers of people who first arrived. Even at the low level of population prevailing in the South Island, however, the first Polynesians hunted and reduced the animal populations, leading eventually to the extinction of moa and a loss of breeding colonies of seals and other species. The overall result was that in the South Island, with at best only marginal horticultural opportunities, there was never the longer-term potential for population numbers that existed in the North Island. A greater proportion of the South Island sites, compared with the North Island, may therefore be earlier.

This leads to an important consequence for illustration in this book. None of these activities, seal and moa butchery and consumption, have left much that is as accessible to aerial photography as the abundant pā and pits of the North Island. However, both the general ecological setting of important sites, and the potential abundance of prey such as seals or fledgling birds (such as mutton-birds) can be illustrated, as we saw in chapter 5. Wairau Bar, a locality for moa butchery and consumption, is illustrated here. The site itself is complex, covering some 3 ha, and lies on the inner edge of the southern point of the Wairau River entrance. 6 Ploughing in the early 1940s revealed many artefacts and brought the site to the attention of the Canterbury Museum. Further hand excavation by Roger Duff, Jim Eyles, Owen Wilkes, Michael Trotter and others was carried out up until about page 200 1964. 7 This work, reviewed recently by Atholl Anderson, showed that a band of ovens lay near the lagoon edge while closer to the sea there was an extensive area of huts and other habitation evidence. The site contained much moa bone and shell: 25 tonnes of moa bone, and some 1,600 tonnes of marine shell, so the proportion of moa meat in the diet was not necessarily very high. At the southern end of the site was an extensive area of human burials. 8 Aerial photographs show the estuarine setting of the site, and the dimpled low relief of the backfilled area of hand excavation conducted by Roger Duff and Jim Eyles still shows clearly.

Further south, at the mouth of the Waitaki River, ovens used for cooking moa showed clearly in oblique photographs taken in the early 1960s. The wider site covered an area of 10 ha on the river terraces 1 km south of the point where the river now enters the sea. Because of the extreme rate of coastal erosion, there is no estuary at the Waitaki River mouth, and access to the site must have been over a bar directly up a fast-flowing river. The most valued parts of moa were probably transported down river, perhaps down the now abandoned river channels which show in the aerial photographs. Unlike Wairau Bar, the Waitaki River mouth may not have been permanently settled; there is insufficient evidence of substantial houses or the artefacts associated with permanent settlement such as adzes. 9 The ovens show very clearly in the photograph in chapter 5 because the area has a thin topsoil on a stony subsoil in which the ovens had been built. The area of the ovens had been little cultivated, 10 and the ploughing of 1960 in effect exposed their image in the soil.

With the exception of a site at the Heaphy River mouth, pre-European settlement on the West Coast of the South Island has been very little investigated until recently. Ray Hooker has now established the existence of sites on the many estuarine banks and river mouths along the coast. These are in natural environments almost unaltered since the time of occupation, with extensive, scrub-covered shorelines, many freshwater lakes and, not far inland from the settlements, the fringe of productive lowland podocarp forests. A photograph of a not untypical setting, north of the Haast River, is in chapter 2.


Previous sections of this chapter dwelt on the precariousness of animal populations under human predation. Horticulture offered a potentially sustainable subsistence base, but in the south it was limited by climate. Once the difficulties of horticulture in a temperate climate had been solved in the North Island, it remained only to put in the much greater amounts of work needed to grow and store the crop, and Māori had a renewable food source limited only by the quantity of suitable soils. As we have already seen in Northland and Palliser Bay, ecological limits appear to have been reached in some areas, although novel methods of exploiting soils were being attempted.

In the South Island, horticulture was possible on the coast, from the Banks Peninsula north, 11 and in the lower reaches of the major northern rivers. Areas known or inferred to have supported horticulture, though much smaller in extent than in the North Island, are Banks Peninsula, Woodend near Kaiapoi pā, Claverley in North Canterbury, the Kai Kōura Peninsula, Clarence River, the Blenheim coastal plains, the inner and outer Marlborough Sounds and, probably most important of all, the rivers and coastal strip of inner Tasman Bay and Golden Bay. 12 In the last areas, however, stone rows are rare, and some of this inferred distribution is based on the existence of kūmara storage pits. The best-preserved examples of stone rows are on the coastal strip north and south of the Clarence River where they are in forms similar to those of the Wairarapa coast. 13

Throughout this broad area, the average number of pits in any one site is small, seldom more than two or three, although there are rare single sites with as many as 15 or 20, for example at Pari Whakatau in North Canterbury, or Pariwhakaoho in Golden Bay. The pits are also weathered with the sides greatly eroded, so that they do not make satisfactory photographic subjects. 14 However, an exceptional group of pits lies on the terraces of the Seventeen Valley Stream at the southern end of the Blenheim plains, not far from Wairau Bar. These pits are in an ideal locality for marginal horticulture: an open river valley facing north, on the edge of a terrace raised slightly above the ponding of frosty air. Below the pits, on the former flood plain, lie stone rows marking the area of the original gardens.

On the open plains between this valley and Wairau Bar to the north-east were Māori 'canals', a feature of Blake-Palmer's pioneering paper on archaeology and aerial photography in New Zealand, and thus among the first aerial photographs published in New Zealand. 15 They page 201

Cooks Cove and Pourewa Island (foreground), Tolaga Bay. The Endeavour anchored to the right (north) of the picture.

page 202

Garden stone rows on the Pōtikirua coastal strip, near Cape Runaway. The view is to the south.

page break

Pukemaire, a pre-European pā on terrace land at Tikitiki, East Coast. The pā was occupied by Pai Mārire forces in 1865.

page break

Ngātapa, inland Gisborne, scene of Te Kooti Arikirangi's defeat in January 1869. Some six transverse rifle trenches or breastworks cross the slope facing the camera. The saps are obscured by the low scrub on the natural scarp across the centre. A vertical aerial photograph and interpretative drawing are on page 152.

page 205

Pukerangiora and Te Arei, on the Waitara River, Taranaki. Te Arei (the Maori position in 1861 and the redoubt of 1864) is at centre with the sap constructed by Pratt's troops at left obscured by the distinct line of trees. At right is the main defended area of the pre-European fortification. Further photographs and an interpretative drawing are on pages 174-177.

page 206

Pre-European pā near Otautu, Pātea River, south Taranaki. A further photograph is on page 170.

page 207

Oika, a pā of the period 1865-68, above the Whenuakura River, south Taranaki. At bottom right is the Whenuakura Marae.

page 208

Putake, a pre-European pā on the Tāngāhoe River, south Taranaki.

page 209

The pā (left of centre) by the State Highway 3 crossing of the Whangaehu River, near Wanganui. Storage pits on the old dune surfaces nearer the river are less distinct. Further photographs are on page 184.

page 210

Pā on ridges above the Whangaehu River. Pits, probably nineteenth century in age, show on low hillocks above the flood plain (in middle distance at left).

page 211

Storage pits at Paekākariki, Wellington. The larger trees on the slope below are karaka.

page 212

Wairau Bar, Blenheim. The main part of the moa-hunter site is in the middle distance by the lagoon edge, but there are significant archaeological deposits on the lines of beach ridge in the foreground.

page 213

Kaiapoi and the Kaiapoi monument, north of Christchurch. Part of the defensive perimeter of the pa is clearly defined by the defensive bank or breastwork of the pā. The defended area extended to the left of the road in the vicinity of the buildings.

page 214

Gum-digging trenches, dating from the 1930s, on the Ahipara plateau. The white patches are the leached sandy subsoil exposed by stripping the gum-rich peat.

page 215

Northburn herring-bone tailings, a gold-mining site near Cromwell, central Otago. Large stones from the sluice face (in shadow nearest viewpoint) were stacked in these rows where they continued to guide waste water flow. Further photographs are on pages 257 and 263.

page 216

Nukutaurua, Māhia Peninsula. At centre is the pā, Waipuna, and at top right is Maungakahia. The pits discussed in the final chapter are by the stream, right of centre, and above Waipuna. A vertical aerial photograph and interpretative drawing are on pages 252-254.

page break
Pits and stone rows north of the Clarence River mouth

Pits and stone rows north of the Clarence River mouth

Stone rows similar to those at Palliser Bay run down from the foot of the steep slopes (partly out of the image, left) to an uplifted eroded beach line running top left to bottom right. Between the stone rows are darkened, gravelly soils, well suited to kūmara horticulture. There are pits on the terrace beyond the prominent gully. The view is to the north.

were supposed to have been excavated to catch 'flappers' (unfledged ducks). These canals are among the great sports of New Zealand archaeology, ranking along with house pits as unsustained interpretations. The 'canals' are natural distributaries, 16 characteristic of the outflow plains of major rivers. As they accumulate over time, with the changing course of the river and river gravels covering over parts of the earlier ones, their natural origin may have been difficult to detect. They would, of course, have been used for eeling and fowling, but their origin is natural.
page break

Storage pits and garden stone rows in Seventeen Valley Stream, southern Wairau Plains

Storage pits and garden stone rows in Seventeen Valley Stream, southern Wairau Plains

Three groups of kūmara storage pits lie on the edge of a high river terrace. The individual pits are up to 6 m in their greatest dimension. On the river flat below the pits there are a number of stone rows, marking the edge of garden plots. The river is actively cutting away the bank just by the southernmost group of pits. An area of gardens may have been eroded away here. The high terrace on which the pits lie appears not have been gardened, to judge from the lack of stone rows. However, it may be that the natural soils here were silty rather than stony. An area of ploughing shows clearly at the foot of the eastern hills but does not extend over all of the high terrace at the time of this photograph (May 1947).

page 219

As we have seen, the marginality of horticulture remains the only credible explanation for the relatively low population numbers in the South Island late in pre-European history. That small population is reflected in the limited numbers and size of pā. In Nelson, pā on coastal headlands are not uncommon but small in size compared with their North Island counterparts. The largest single pā, and also the greatest concentration of individual pā in the South Island, are on the Kai Kōura Peninsula. On the edge of the high coastal terrace at South Bay, there are five well-preserved pā, the largest of them with a complex set of interlocking defensive ditches and banks. Nearby, on the northern edge of the terrace, are several other pā. 17 The pā were lived in from the pre-European period up until the raids of Te Rauparaha in the period 1822-28. 18

South of Banks Peninsula, pā with earthworks are rare. Pā a Te Wera, on the Huriawa Peninsula near Kari-tane, some 20 km north of Dunedin, has little to show by way of defensive earthworks. The Huriawa Peninsula is more a complex of small settlements, some with unambiguous terraces, others apparently defended by narrow natural razorback ridges and steep coastal cliffs. 19 Mapoutahi, another pā on a headland north of Dunedin, has very little constructed fortification 20 compared to the scale of such fortification in most parts of the North Island. Other pā, both in the Banks Peninsula region and to the south, are more usefully reserved for discussion in the section of this chapter on nineteenth-century settlement.

South Island stone resources

Throughout pre-European Māori history, the South Island remained an important source of stone. There was no obsidian but there was an apparent abundance of stone for adzes. The main sources were along the alpine fault and its northern and southern extensions in Southland and Nelson. Argillite for adzes was found from Rangitoto (D'Urville Island) to the mountain range lying just east of Nelson. 21 D'Urville Island lies in the western reaches of Cook Strait, with its northern tip almost as close to Pātea in Taranaki as it is to Nelson in inner Tasman Bay. Tribes based there, therefore, had good access to the North Island for trade; much stone went north, and east through Cook Strait, to judge from the amount of Nel- page break
Pā on the Kai Kōura Peninsula

Pā on the Kai Kōura Peninsula

Five or six pā, several linked in pairs or groups, show on the edge of the high terrace in this 1942 vertical aerial photograph. They are Ngā Niho, above the township at the top of the photograph; and the various pā (unnamed) above South Bay. Ngā Niho has had its ditch ploughed in and the bank height 'enhanced' in recent times. The pā are all built on various levels of old marine terraces. The two pā to the south-east (bottom right, north-east of the isolated house and hedges) lie adjacent to water (in the gully). The easternmost one has a second ditch and bank which shows faintly. There is a dogleg section of ditch and bank defending the terrace edge at its westernmost extremity. The pā with the largest defended area, one of the largest in the South Island, is on the prominent point at left. It has an interior and exterior defensive line, with the exterior line not fully closed. However, the apparently incomplete defence takes advantage both of the upper slopes of the gully and the proximity of the adjacent pā. The defended area of the large unit is 280 m long by about 50 m (width varies); the prominent section of ditch and bank at left is 90 m long. The defensive form of these pā indicates pre-European occupation, but they were also occupied into the earlier part of the nineteenth century until the raids of Te Rauparaha in the late 1820s.

page 221 son argillite found in sites in the western and southern North Island. A photograph of a quarry at Samson Bay, not far from D'Urville Island and also in the Nelson mineral belt, is in chapter 5. In Southland, similar stone is available in inland and coastal localities but was only exploited on the coast; in the vicinity of Riverton and Bluff Harbour are extensive workings. 22 It appears in the same distinctive landscape as in Nelson, with hard rock protrusions lying above the general surface of the landscape. Nephrite (pounamu) was less used than argillite in the earlier periods for reasons that are not entirely clear. 23 It is tougher than argillite, 24 and was used late in pre-European history as a prestige trade good. Greenstone occurs in remote areas in geological formations similar to that of argillite, with the vegetation also stunted in the vicinity of the primary outcrops. Greenstone was also recovered from river-beds but the landscape imprint of this activity is not remarkable.

Nineteenth-century settlement

The South Island did not see widespread military conflict between Māori and European forces, such as the North Island suffered. However, earlier in the century the general north-to-south imbalance in the availability of muskets, and the subsequent raiding of the south by northern tribes, caused devastation in the main Ngāi Tahu villages. The principal sites date to the period when Te Rauparaha of Ngāti Toa was trying to establish his mana whenua over the South Island. This was after the initial raids of the early 1820s when Ngāti Koata, allied with Ngāti Toa, occupied much of Nelson-Marlborough.

The pā at Kaiapoi (north of Christchurch) and Onāwe (Akaroa Harbour), attacked by Te Rauparaha in separate actions in 1829-30 and 1832, 25 illustrate quite different kinds of ecological and defensive settings. Kaiapoi is a large example of a swamp pā, i.e., the defensive perimeter is strengthened by a swampy lagoon, the Taerutu swamp, which lies inland of the dunes of Waikuku Beach. The pā itself covered an area of 2 ha with a perimeter ditch and bank and double palisade. The pā was divided internally by a ditch and bank surviving today parallel to and north of the line of the present-day road. 26 At a late stage in the engagement in 1832, which lasted several months, Te Rauparaha appears to have dug a series of saps in the form of connected parallels at the south-west entrance of the site. At this point the pā joined firm ground to the south-west, page break
Kaiapoi, site of the defeat of Ngāi Tahu by Te Rauparaha in 1832

Kaiapoi, site of the defeat of Ngāi Tahu by Te Rauparaha in 1832

The perimeter bank by the edge of the swamp shows clearly following recent clearance of willows from the site. The swamp at rear still remains while, on this side, the original strip of swamp and stream is marked by the open drain in the foreground. The pā originally extended in another complete segment over the area where the house and sheep are at right. The bank casting a strong shadow by the road is 90 m long. One of the original gateways through the bank by the road shows clearly. Te Rauparaha's saps occupied the lefthand edge of the rectangular paddock on this side of the road and led in towards this gateway. The view is from the south-west. A colour photograph with interpretative drawing is on page 213.

difficult ground for the attacking force since it was flanked by the Ngāi Tahu defenders from both north and east. These particular areas have unfortunately been destroyed by the modern road, 27 but the banks north of the road, the original entrances through them and some other interior features are still evident.
Onāwe lies on a peninsula in Akaroa Harbour which, because of its very narrow neck, is more like an island. Te Rauparaha turned to an attack on the pā here after his sacking of Kaiapoi. The nearly flat slope of the surface of the ridge had been divided into several enclosures by defensive lines of ditch and breastwork. The exact placement and configuration of the fortification appear to be a response to the size of the peninsula: it was too long to be defended as a single perimeter. A simple ditch and bank across the ridge was inadequate and a fully con- page break
Onāwe pā, Banks Peninsula, attacked by Te Rauparaha after the sacking of Kaiapoi

Onāwe pā, Banks Peninsula, attacked by Te Rauparaha after the sacking of Kaiapoi

The defences of the pā are in part natural, using the steep cliffs of the peninsula (at top, in scrub). A ditch and bank, partly obscured by the scrub, extends along this cliff edge with a short field of close fire against any attack from that quarter. The distinct ribbon-like line is a path through long grass. On the near side of the main rectangular block of fortified perimeter are trenches which give access to water and the shoreline (foreground). The total defended area on the peninsula is about 450 m long by 40 m (width varies). The rectangular enclosure seen here (ditch and bank out of view at left) is 130 by 40 m. The view is to the north-east. For a wider view of the pā, see the illustration of the fish trap (here out of picture, bottom left) in chapter 5.

structed enclosure had to be made, as indeed we have seen at Kaiapoi. On the northern edge there is a short field of defensive fire (although the area may have been reduced by slumping since the 1830s). The rectangular defended areas on the broad slope are also overlooked by the steep stony hill to the east. It may be speculated also that the defensive perimeter was originally a ditch and bank fence for a village, internally partitioned, perhaps between hapū. This pattern is apparent in many early nineteenth-century pā on river terraces in the eastern North Island.

Onāwe appears on first sight to have been a good defensive position, and it was not taken by a direct assault. Te Rauparaha had gathered at Onāwe with his own forces and many prisoners from his earlier defeat of Kaiapoi. The prisoners were Ngāi Tahu and closely re- page break lated to the occupants of Onāwe. They were brought forward to open a discussion with the defenders. In the confusion and reluctance of the defenders to fire on their own people, the Ngāti Toa were able to enter Onāwe. Once inside, they turned on the defenders and many Ngāi Tahu were killed. Following these two incidents, Ngāi Tahu abandoned their other central Canterbury pā, such as those at Taumutu, and fled further south. 28

The early nineteenth-century pā at Taumutu (near Lake Ellesmere), Wairewa (Lake Forsythe) and Waiteruaiti (near Timaru), have massive defensive ditches and banks. The defensive significance of the position of one of the sites at Taumutu, Orariki, is not obvious at first sight. From the air, however, it can be seen to lie across a narrow neck of dry land formed where the mudflats of the margins of Lake Ellesmere (Waihora) finger into the land. 29 This locality was first settled by the chiefs Te Ruahikihiki and Kaweriri.

Orariki, a pā near Taumutu at the southern end and outlet of Waihora (Lake Ellesmere)

Orariki, a pā near Taumutu at the southern end and outlet of Waihora (Lake Ellesmere)

The church and graveyard at top right are within the bounds of the defensive ditch and bank, immediately left of the road at top. Built in the early nineteenth century by the Ngāi Tahu hapū, Ngāti Ruahikihiki, the pā originally occupied the neck of dry land that extended out into the lake between mudflats. The Waikewai Stream bed and mudflat show clearly on the near side of the church. The defences may originally have been dogleg in plan, with the modern road obscuring a large part of the original defences. The view is to the north-west.

page break
Oruaka, a pre-European pā above Wairewa (Lake Forsythe), Banks Peninsula

Oruaka, a pre-European pā above Wairewa (Lake Forsythe), Banks Peninsula

Running up through the centre is the pre-European ditch and bank, with the pits and terraces of the settlement on the slope top right. Outside the major ditch and bank, and postdating it, are irregular ditch and bank enclosures. They may relate to early or mid-nineteenth century Maori settlement; they may be of European origin. The view is to the east.

page break

1 Tau et al (1990: 3/6).

2 Te A. Davis and Wilson (1990: 91); Tau et al (1990: 5/25).

3 For discussion, see Davidson (1984:30-59). Anderson (1989) discusses many South Island sites.

4 Anderson (1989).

5 I. Smith (1989).

6 Duff (1977); Anderson (1989: 122-125).

7 Trotter in Duff (1977: 349-354).

8 Anderson (1989: 122-125).

9 Anderson (1989: 131-134); Teviotdale (1939); Knight and Gathercole (1961).

10 Knight and Gathercole (1961: 133-136).

11 A. Jones (1962). Records of pits to the south of the peninsula are not considered to be for horticulture.

12 Trotter (1977); Trotter and McCulloch (1979); Brailsford (1981: 96-176); Rigg and Bruce (1923); H.M. Leach (1984: 33-53).

13 Trotter (1982: 97-99); McFadgen (1980b: 9-12).

14 Aerial photographs in Brailsford (1981).

15 Blake-Palmer (1947).

16 The main river channel enters a fan and breaks into several smaller channels or distributaries.

17 Brailsford (1981: 119-129).

18 Evison (1993: 49-50).

19 McKay and Trotter (1961).

20 Anderson and Sutton (1973).

21 K. Jones (1984b).

22 H.M. Leach and B.F. Leach (1980); Bristow et al (1985).

23 Davidson (1984:195-200).

24 Toughness was important in maintaining the cutting qualities during hard use.

25 Evison (1993: 51).

26 Brailsford (1981: 177-181).

27 They show only in original plans, see Brailsford (1981:177).

28 Brailsford (1981: 184-189); Evison (1993: 61-62).

29 Brailsford (1981: 154-156).