Volume 65 1956 > Volume 65, No. 4 > Notes on Maori sites in Eastern Wellington Harbour, by J. B. Palmer, p 342-355
                                                                                             Previous | Next   

- 342




  • (a) Main Camp Site, Paraoanui.
  • (b) Middens 1-3, Nga Hu and Nga Rerenga.
  • (c) Hut Site, Mt. Cameron.
  • (d) Middens 4-5, Pencarrow Head.
  • (e) Burial, Camp Bay.



THE REGION fringing the eastern side of Wellington Harbour has been neglected for many years and numbers of people have assumed that investigation of probable Maori sites would not be worthwhile. Two reasons seem to be responsible for this attitude to what is really an interesting region, where some of Wellington's earliest inhabitants lived, or at least camped, for seasonal periods.

These reasons are firstly, that the sites are unimportant as they would probably be those of recent occupation and secondly, that erosion and man have destroyed all signs of likely sites, making recovery of any material evidence most unlikely. Although the first of these is partly true, evidence has been found of early occupation. In regard to the second, erosion has to some degree protected the sites by covering them with a layer of talus, while man has both destroyed and exposed part of the sites, making inspection possible.

There has been little recorded about the section of harbour coastline under consideration and in his map of Maori place-names, Elsdon Best 1 showed two sites only in this region. Both of these appear to be incorrectly located but this fault may be due either to his informants or the draughtsman of the time making an error. Best used information taken from the now suspect Te Matorohanga source and much of it must be questioned. He also based his map on observations made by H. N. McLeod who may have been at fault in plotting the sites. These factors make it necessary to give a more accurate account of the region and these notes are presented in the hope that they will make a small contribution to a better appreciation of Wellington's past.

- 343
FIG. 1.
Occupation sites in East Harbour, Wellington. Karaka groves are marked by crosses.
- 344

The particular length of coastline described extends from just south of Point Arthur (map reference 436197) 2 to the headland marking the limit of the little bay immediately to the south of Pencarrow Head (map reference 403132). In this region alone eight sites have been located and some of them are previously unrecorded ones. Other possible sites exist but they will not be described in this paper.


Wellington Harbour is a drowned or submerged area which gives it the typical indented appearance with small bays and coves between numerous headlands. Within the period of human, and even European, settlement uplifts have taken place totalling several feet. They are evidenced by raised beaches, hanging valleys and modification of the shoreline by wave-cutting and beach-building. In such a region former settlement was limited by the extent of relatively flat land in the pockets and it could be said that discontinuous, littoral occupation resulted. Only in the more favourable places would settlement be possible as the high, steep hills must have been a limiting factor in earlier times.

At present the hills on this side of the harbour are grassed, with gorse and tauhinu as a secondary cover. Nearer Pencarrow the grass is more of the tussock type, especially on wind-swept faces, and flax grows on the steep slopes. Groves of ngaio fringe the coast in sheltered bays, together with occasional karaka groves near old sites. In some of the hanging valleys and larger gullies there are fairly dense clumps of coastal trees. Beyond the hill crestline remnants of the original beech and rimu cover persist but this is absent on the seaward side until one goes further north to Eastbourne. In 1839 Heaphy 3 observed that “from the mouth of the Hutt River to outside of Ward Island, the forest was uninterrupted, and the trees overhung the water, giving shelter to great numbers of wild fowl.” Heaphy's description is broadly true today and it shows that the original character of the vegetation has not been completely destroyed. But the region dealt with in this paper could not have been a friendly one as there was not much shelter from the winds.

Yet the mention of wild fowl recalls the fact that the hinterland forest was an important source of supply for food as it was “teeming with birds.” 4 Shore and wading birds were also evidently common and Maori fowlers had ample opportunity to practise their art. Sea-foods were rich in variety and quantity while the deeper water provided scope for catching fish as tarakihi, snapper, barracouta and kahawai.

Unsuitable for cultivation, the coastal fringe must have supported small groups of people on fowling or fishing excursions during the warmer months. These people may have crossed the harbour from such places as Seatoun or Somes Island, where kumara cultivation was said to have taken place. 5 On the other hand, it is faintly possible that small - 345 numbers of people lived permanently at some of the sites, as a hut site has been located on the hillcrest above one of the middens. There may be others which have been covered by talus.

Investigation reveals that erosion took place preceding and during Maori settlement because some of the oven and midden areas are built on old talus slopes and, in one or two cases, extend uphill a little distance. The ancient underlying talus is seen clearly in Fig. 6. Subsequent talus of a later cycle of erosion has protected these areas by a top-cover which varies in depth. The history of many of the sites is complicated by probable earthquake disturbance, the building of a quarry railway around a section of the coastal fringe in 1905 and lately, by road-building operations destroying undetermined parts of them. As a result, some sites have been sliced through, scattered to form the new road-bed and the remnants exposed on unstable banks. This record is based on the remaining small areas which were originally larger. Before they became completely destroyed it was decided to place on record the present fragmentary information. No systematic digging was possible and the terminal sections are given despite their limitations.


(a) Main Camp Site, Paraoanui.

This site has been named Main Camp for the convenience of description. It is identified with the occupation area shown on Best's map as Paraoa-nui which is incorrectly marked on that map as being in Camp Bay. Any observer would find it hard to visualise settlement in Camp Bay prior to the uplifts of last century when there could only have been a narrow rock strip against the faceted cliffs. Best stated that Paraoa-nui was a place on the beach just north of Nga Hu and Nga Rerenga. 6 The Main Camp is certainly located on an old beach and is just north of those two places which were “at or near” Section 63, a description which fits the locality today as Middens 1, 2 and/or 3 appear to be Nga Hu and Nga Rerenga. An important point is the fact that the Main Camp is situated on the largest area of flat land along this part of the coast and as such, it was a more favourable place for settlement than the exposed, steep shore of Camp Bay. One may confidently assume that the Main Camp is actually Best's incorrectly plotted Paraoa-nui.

The site (Fig. 2) is bounded by swamp and talus from the hill rim inland and by rocks and beach seawards, with the stream rising on Mt. Cameron cutting through the flat at the north end. North of the stream is a separate, thin oven area but the more important oven area is south of it. This extends for about seventy yards and the width varies from twelve to fifteen yards. There are no signs of whare sites although there are small ledges which could be mistaken for former hut sites.

- 346
FIG. 2.
Main Camp, or Paraoanui, situated in a harbour pocket.
FIG. 3.
Vertical sections at Points A, B, and C, Main Camp or Paraoanui.
- 347

Three sections (Fig. 3 A, B, C) show the existing stratification. Overlying a ground mass of clay, rubble and alluvial wash is a deposit of ovenstones and charcoal which measures 8-9 in. in each lense. Above this layer is a deposit of fine, stained pebbles, shells, bones and charcoal which is 12 in. at its greatest depth and, like the layer below, thins out and slopes upwards at the south end. A deposit of recent talus covers the area and this is the extension of the fans shown in Fig. 2. It was not possible to dig transverse trenches but the stratification was recorded at the limit of the site inland (Point D), this being thirty-six feet from Point B. At D the layers are much thinner and beneath a 9 in. cover of clay there are 2 in. of shell, 4 in. of ovenstones and underlying these is a 2 in. layer of charcoal.

Material has been recovered from the exposed faces, most of it from the layer of shell, pebbles and charcoal but some was found at various places on the surface. The main species recovered are:—

  • Maori dog: canine teeth.
  • Dolphin: teeth of Bottle-nosed dolphin (Tursiops truncatus).
  • Fish: bones of barracouta (Thyrsites atun) and Parrot fish.
  • Birds: weka and other unidentified bird bones.
  • Mollusca: paua (Haliotis iris, H. virginea), limpet (Patella), Turban shell (Cookia sulcata), Mussel (Mytilus sp.), common pipi and cockle.

The number of artefacts found has not been large. An interesting fragment of a quadrangular adze shows a distinct ridge of transverse type across the back of the blade which is slightly concave longitudinally. At the time of finding this was thought to be similar to Duff's Type 2, Variety A, but the fragment was not large enough to form a definite conclusion.

Since then a complete specimen (Fig. 4) confirms the type and it was found only a few feet away from the fragment at Point D. It was at a depth of twelve inches in a layer of pebbles and charcoal. The adze is quadrangular, the back of the blade is concave longitudinally with crescent-shaped transverse ridge and the butt shows some reduction on the back. It is well ground and the profile is so similar to the fragment that it could have been produced by the same craftsman. The butt itself was not ground and it was marked by bruising. The cross-section shows the front wider than the back which was also the case with the fragment. Material: dark grey baked argillite. Length: 160 mm. (6 5/16 in.); depth: 32 mm. (1 4/16 in.); width: 56 mm. (2 3/16 in.).

That this type of adze was not the only one typical of the site is shown by another butt of an adze whose width is not much greater than the depth. It is quadrangular with rounded edges and one surface shows some reduction. The fragment is partly ground.

Amongst the debris slipping down from the bank was an unusual black, argillite chisel (Fig. 5). It has a triangular sectioned blade still attached to what may have been part of an adze as there is a polished surface near the butt end. The longitudinal ridge formed by the intersection of the sides is ground to form a cutting-edge bevel and grinding

- 348
FIG. 4., FIG. 5., FIG. 8.,
Adze from Main Camp., Chisel from Main Camp. Length 4⅛ ins., Bone minnow shank. Midden 2. Length 3 3/16 ins.
- 349

has been started on the reverse surface. This chisel would appear to be related to Duff's Type 3, Variety D. Material: baked argillite. Length: 105 mm. (4 2/16 in.); width: 17 mm. (11/16 in.).

Other material from this site comprised a pumice rubber, a sandstone grindstone showing signs of considerable use, two small fragments of adzes, several beachstone spawls or cutters, kokowai, flakes of flint and the usual obsidian.

The thin oven area on the north side of the stream is exposed in places but elsewhere it is covered with light vegetation. Ovenstones, shells, and fine, stained pebbles cover what is now only a small area. The only material found here consisted of a pumice top (potaka), a badly decayed part of a bone fish-hook, obsidian, and a section of clay pipe stem. In addition to mollusca, which were similar to those of the main site, the spine of a stingray was found.

It is known that Ngati-Ira lived at Paraoanui until 1825-26, when they were expelled by Ngati-Awa 7 so that it was occupied until historically recent times. Ngati-Awa probably left it at some date after 1840 when the infant settlements in the Hutt Valley and Wellington were springing up. If Best's genealogies 8 are accepted then it would appear that the earliest date of occupation by Ngati-Ira would be somewhere about 1625. This must be regarded as a tentative date only as Paraoanui may have been occupied before or after that, but C14 dating will provide a firmer horizon.

(b) Middens 1-3, Nga Hu and Nga Rerenga.

A short distance south of the Main Camp are three middens which appear to be on the site of Nga Hu and Nga Rerenga. These two places were said by Best 9 to be “on the beach, at or near Section 63, north of Pencarrow Head.” They were evidently named from an incident whereby two women narrowly escaped drowning there. The actual locations of the two places seem to be those of Middens 1 and 2 as recorded on the map (Fig. 1) because Midden 3 may simply be a now isolated extension of Midden 2.

All three middens show evidence of being above old talus slopes. Although they are shown at varying depths from the present surface (Fig. 6) they are approximately the same height above the road and the variation in depth could be accounted for by differential erosion at a later date.

Midden 1 is located just south of the old stone wharf (map reference 416166) 10 and lies beneath three feet of rock, soil mixed with rock, and much larger rock. These can be defined as three separate deposits and it would seem that stratification can be detected to some degree even in such unpromising fields as talus slopes. The section illustrated (Fig. 6) is a transverse one drawn to give some idea of the way in which oven areas and middens on this coast were made on old talus slopes forming the bedrock at the time of occupation. There is a 10 in.

- 350
FIG. 6.
Vertical sections of Middens 1, 2 and 3.
FIG. 7.
Butt of an adze, Midden 2.
- 351

layer of small pebbles, ovenstones, shell and charcoal but even this remnant is poor. It extends 45 feet uphill and probably marks the edge of a small site about which nothing is remembered but the vague circumstance giving rise to the name of either this site or its neighbouring one. The only tools found here were a few spawls.

Fish: Parrot fish (pseudolabrus).

Mollusca: paua (Haliotis iris), sea-snail (Amphibola crenata), limpet (Patella), cat's -eye (Lunella smaragda), whelk (Comminella adspersa).

Midden 2 lies across the harbour from Seatoun (Fig. 1) where the important Te Whetukairangi pa was situated. It was at Seatoun that moa remains were found associated with human middens 11 and H. M. Christie thought that a moa-hunter camp could be placed there. In recent years, G. Leslie Adkin has confirmed this by figuring an adze from Seatoun which is of accepted moa-hunter type. 12 With these facts in mind, it had been puzzling not to find further evidence of occupation in the eastern harbour region as moa-hunters must have moved about most of the harbour shoreline on periodic excursions.

When this particular midden (Midden 2) was examined it did not look very promising but later finds, however, showed it to be an early site with a few artefacts typical of moa-hunter culture. Until C14 dating provides an answer, caution is necessary as J. Golson has pointed out that “save within the broadest of limits a New Zealand archaeological site does not by the material culture it discloses proclaim its relative and still less its absolute age.” 13 Other specimens of similar type were found at a later date further round the coast but finds should perhaps be greater when one considers the proximity of Wellington to the great moa-hunter site at Wairau. Sites and artefacts in the Wellington area may in time throw a little more light on the moa-hunter culture period and it is hoped that greater interest may be stimulated by readers recording available data.

Midden 2 lies seventy-five yards south of Midden 1 and extends south for forty-five feet. At a depth of 6 inches below the surface, it consists of fine, stained pebbles, charcoal, ovenstones, shells, bones, and some rubble. Talus again forms the bedrock. The main species recovered are:—

  • Dog: two mandibles.
  • Fish: Parrot fish (pseudolabrus), Frost fish (Lepidopus cordatus), barracouta (Thyrsites atun).
  • Mollusca: cockle (Chione stutchburyi), sea-snail (Amphibola crenata), pipi (Amphidesma australe [?]), whelk (Comminella adspersa), limpet (Patella), mussel (Mytilus sp.), Turban shell (Cookia sulcata), cat's-eye (Lunella smaragda) and paua (Haliotis iris and H. virginea).
  • Bird and other bones not yet identified.
- 352

The butt of an adze (Fig. 7) was recovered from this site and it seems to be either Variety A or B of Type 4 in Duff's classification but it is more likely to be the former. The fragment is long and narrow with a marked triangular section at both the poll and fractured end. The ridge shows reduction towards the poll to form a tang which is marked by hammer-dressing. The colour is unusual, being creamy and it gives the appearance of baked limestone. Length: 9.5 mm. (3¾ in.); width: 4 mm. (1½ in.); depth: 3.5 mm. 1 4/10 in.).

A bone minnow-shank was also recovered from Midden 2 by Derek Ritchie. Brown and aged with natural cavities forming striations, it is oval in section. There is a platform at the distal end with a scarf and four propecting lugs. The platform itself is slightly concave or grooved and the end is quite straight while there is a transverse hole drilled from both sides at the proximal end of the shank. It is illustrated in Fig. 8. Length: 3 3/16 in.; width: 8/16 in.; depth: 6/16 in.; diameter of hole: 1/16 in.

Apart from these two pieces there has been little found at the site. A piece of rhomboid-shaped argillite has been polished for some unknown purpose and obsidian has been rare, only a few small pieces being found.

Midden 3 may be an extension of the previous one which has perhaps been cut off from it by roading operations. Located some twenty yards from Midden 2, it extends for a distance of twelve feet and is bedded a similar distance above the road at its lowest point. The topsoil is 4 inches deep and the main over-burden, varying between 18-24 inches in depth, is soil mixed with rubble. Beneath this is a 6-inch layer of stained pebbles, ovenstones, shells, bones and rubble overlying the familiar bed of old talus.

Nothing of value was recovered from this site but in comparison with the other middens a point of interest arises. The apparent greater depth of Middens 1 and 3 need not mean greater age. Indeed, Midden 2 lying near the old surface, has yielded some material of a supposed early culture period. This accords with J. Golson's statement that thickness of deposit and depth below the surface may be completely irrelevant. 14

(c) Hut Site, Mt. Cameron.

The hillslopes above the Main Camp and the three middens were examined for traces of occupation as the possibility existed that the middens were partly the result of deposition from higher points. There were a few terraces of doubtful origin and two dying karaka groves marked by crosses on the locality map.

The only reliable clue was a report from Oliver Burdan of Gollans Valley that an “oven” area was to be seen near Mt. Cameron. This was found to be down the ridge south of Mt. Cameron, about twenty-five yards west of the telegraph line. Situated just over the lip of the crest above the middens it was at the head of a gully leading down to Kohangapiripiri lagoon and close to a spring.

- 353

This site turned out to be not an oven area but a hut-site, comprising a sub-rectangular depression with rounded corners and a raised, flattened rim. The inside measurements were twelve feet by nine feet while the rim was approximately two and a half feet high and three feet wide at the top before it sloped down and out. The site was orientated east-west with the rim absent at the western end where one would suppose the doorway to be.

The area was tested for evidence of the site's nature and some slight charcoal traces were found beneath the clay cover about the centre of the site. No further charcoal was found anywhere, nor any shells, stones or bones so that negative evidence confirmed the view that it was a hut-site.

The reason for its existence was soon apparent when looking out to sea. It commanded views of the two critical headlands, Baring Head and Sinclair Head, besides a section of Cook Strait so that it was undoubtedly a look-out site. Being well hidden from the sea it would be an ideal and logical site for the purpose. Shallow, circular depressions in the grassy flat close by suggest traces of occupation, particularly as the grass in them is a different shade.

(d) Middens 4-5, Pencarrow Head.

The site named Midden 4 was reported by Mr. G. Leslie Adkin as being in the little bay to the south of the track leading from the beach to the lighthouse entrance. No details of this midden are available as it was exposed and subsequently covered over by the road-making machines.

The last site, Midden 5, was only recently located by the writer. It lies close by the jetty, south of Pencarrow Lighthouse, in a very bleak, exposed spot. There is a top-cover of 15 inches of clay and rubble below which is a 4-inch layer of shell, ovenstones and bones. Human remains including segments of a skull were found mixed with the usual midden material.

These two middens are probably linked with the headland which separates them. H. N. McLeod 15 reported “Indistinct traces of native occupation have been noted on the hill near the lighthouse.” Unfortunately, he gives no clue about the type of site there but it would seem to be a natural place for a headland pa. Middens 4 and 5 may have marked the cooking sites for that place. Best 16 has stated that Tautoki's fort at Parangarehu was on that headland but he appears to be in error as Colenso 17 makes it quite clear that Parangarehu was the village further round in Fitzroy Bay, to the east of the Paiaka Stream. Colenso 18 writes about a canoe trip across the harbour and round Pencarrow to Parangarehu where he spent the night. The inhabitants of Parangarehu also grew wheat according to the same authority 19 and - 354 one finds it hard to see how eighty bushels of wheat could be harvested at Pencarrow. At the correct site of Parangarehu, however, there is a cultivated flat and two stone walls so that Best's information about Parangarehu appears to be unreliable and should be disregarded. There are traces of a site further north on the ridge near the lighthouse and it is possible that Tautoki or some other person did have a fort on or near the headland and that its name has been confused with the village name.

Midden 5 marks a suitable southern limit for the region as most of the sites have features in common. In Fitzroy Bay the nature of the sites change and, being similar in character, they form another distinctive region for that part of the coast. A later paper will describe that region.

(e) Burial, Camp Bay.

Mr. Oliver Burdan, Sen., states that about the year 1921 two “Maori” skulls were found in Camp Bay near the present grove of ngaio trees. They were later placed on the gateposts and remained there for a considerable time. Nothing is recalled about the circumstances of their discovery and one cannot exclude the possibility that they were brought and left there from another part of the coast. If Camp Bay was not a habitation area it may perhaps have been a local burial site. No reports of other remains are known and one hesitates to suggest secondary burial on so little evidence.

As there is an absence of sand-dunes which would provide the incentive for beach burials, it could be concluded that caves were used by the people along this part of the coast. If so, this would help to account for human remains in this bay. An unlocated burial cave called Te Anakopiro 20 was said to exist in the Wainui Valley but this was apparently linked with the more recent inhabitants of Wellington. Earlier people may have used caves closer at hand along the immediate coastline.


Although the information given in these notes is rather fragmentary, it is clear that much work remains to be done around the neglected coast east of Wellington. It is important to have absolute dating by C14 methods so that the pattern of migration and settlement becomes clearer. There is also a need for a re-examination of all known sites where such still exist and for them to be interpreted in the light of present knowledge. The notes suggest that early sites can still be found, but caution is necessary until the results are confirmed by scientific dating. The possibility of post-Wairau resettlement must be kept in mind and such dating will show how much ebb and flow took place about Cook Strait.

- 355

I want to thank Mr. Oliver Burdan and his son Oliver, Jun., for their companionship, information and help during many searching hours, also Mr. G. Leslie Adkin for his advice and assistance and Messrs. Y. M. C. McCann and J. Moreland of the Dominion Museum for their identification work.

  • ADKIN, G. Leslie, 1950. “Supplementary Data Relating to the Ancient Waitaha in the Horowhenua - Te Whanga-nui-a-Tara Area, North Island, New Zealand.” Journal of the Polynesian Society, 59:1-34.
  • BAGNALL, A. G. AND PETERSEN, G. C., 1948. William Colenso, Printer, Missionary, Botanist, Explorer, Politician. His Life and Journeys. Wellington, Reed.
  • BEST, Elsdon, 1917-1919. “The Land of Tara and They who Settled it.” Journal of the Polynesian Society, 26:144-169; 27:1-25, 49-71, 99-114, 165-177; 28:1-17, 79-96, 123-133.
  • DUFF, Dr. Roger S., 1950. The Moa-hunter Period of Maori Culture. Wellington, Department of Internal Affairs. Canterbury Museum Bulletin No. 1.
  • GOLSON, J., 1955. “Dating New Zealand's Prehistory.” Journal of the Polynesian Society, 64:113-136.
  • HEAPHY, Charles, V.C., 1879. “Notes on Port Nicholson and the Natives in 1839.” Transactions and Proceedings of the New Zealand Institute, XII: 32-39.
1   Best 1918: frontispiece.
2   N.Z.M.S. Series No. 1, 1:63360, Sheet N. 164.
3   Heaphy 1879:33.
4   Heaphy 1879:33.
5   Best 1917:154-156.
6   Best 1918:166.
7   Best 1918:103.
8   Best 1918:12-13.
9   Best 1918:166.
10   New Zealand 1:25,000, Sheet N. 164/6.
11   Duff 1950:285.
12   Adkin 1950:30.
13   Golson 1955:115.
14   Golson 1955:115.
15   McLeod's notes are contained in Best 1919:1.
16   Best 1918:166.
17   Bagnall and Petersen 1948:215.
18   op. cit., 215.
19   op. cit., 227.
20   Best 1918:10.