Ngā Tohuwhenua Mai Te Rangi: A New Zealand Archeology in Aerial Photographs
4 — Horticulture
Primarily focused on the cultivation of rootcrops such as kūmara or taro, Māori horticulture survives as only one element of horticultural practice which the first Polynesians attempted to bring to New Zealand. The staple foods in Central Polynesia, derived from tree crops such as breadfruit or coconut, 1 were from trees that could not survive in New Zealand. However, the processing techniques applied to the Polynesian tree crops stayed on in the treatment of New Zealand native counterparts of Polynesian trees, such as the fruiting of karaka (examples of which appear in chapter 13) or the fibres of the epiphytic kiekie. 2 The crops that survived in New Zealand were the plants that were well adapted, through sophisticated cultural practices, to seasonal cold: the root crops, yam, taro and kūmara (sweet potato). 3 Māori had first to seek soil conditions which matched the temperature and physical requirements of their crops, which had evolved and developed in cultivation in the tropics. High mean temperatures, at least over 18°C, over a five-month growing season, were therefore critical. Even where physical conditions were adequate, good or moderate soil fertility was also needed for gardening to be sustained for more than a few years.
Physical conditions and soil fertility
The most important physical conditions were sunshine, a degree of shelter and the need to avoid frost in the growing season. Moisture was also important, but critical only for taro, which is a species adapted to water. The New Zealand species is known as a 'dryland' variety but it still needed a good supply of water to flourish. It has survived by natural propagation in some localities from the East Coast northwards, but only where water is present, for example, near springs.
Sunshine, its effects enhanced by careful choice of site-aspect, cultivation and other modifications of the soil, provided the warmth needed for the tropical crops to thrive and the radiant energy needed for photosynthesis. Apart from the obvious digging and weeding, the most important cultivation techniques were mounding or the formation of raised rows, and the selection of areas that had stony or gravelly soil. Where there were nearby sources of gravel or coarse sand, these could be applied to the surface as a mulch, where they appear to have raised the temperature of the soil by a few critical degrees. Eventually, the gravels came to be mixed into the soils, improving friability where this was otherwise lacking. On river terraces or other terrace landforms and older dunes, it is not uncommon to see borrow-pits, irregular depressions ranged along the edge of the terrace 4 or on ridgelines. 5 These pits were dug to take sand or gravel from lower strata in the terrace, to add to the surface of the garden plots on the nearby terrace. The illustration here is from near Lake Karapiro on the course of the Waikato River, and Aotea Harbour examples are in chapter 8. The application of gravel or sand, and the blackening of the soil as the result of fire, meant that the soils would warm very rapidly. This would not be so important late in the planting season, but early in the season it initiated the sprouting of the leaves and roots of the kūmara. The overall effect was a lengthening of the growing season for crops that had not evolved or been bred for the cool, short growing season of southern latitudes.
Borrow-pits on the edge of terraces on the Waikato River near Tirau
The borrow-pits show as irregular depressions up to 20 m across, aligned along the edges of the terraces. On the hill above where the road cuts through the terraces, a pā shows (but not very clearly). On the furthest (righthand) extent of the terrace there is a ditch and bank, dog-legged in plan. The ditch was about 4 m across and it is likely to be a late pre-European or nineteenth-century fortification, exploiting an excellent tactical and strategic position on the terrace. The valley floor is now the bed of Lake Karapiro. The road is State Highway 1 where it leaves the Waikato River (flowing downstream from east to west) some 10 km from Tirau.
Swamps and river valleys generally were highly fertile and important in gardening. However, with the exception of taro, Māori root crops did not flourish with page 63 'wet feet'. Swamps therefore had to be drained, 6 and even where swamps were not drained deliberately, they may have been seasonally dry in some places, for example, around the broad fringes of infilled lakes. As the drains were dug and renewed over a span of time, the fertile organic matter was thrown up into beds, raising the surface slightly and increasing the thickness of the topsoil. Layers of sand may have been encountered and these would be mixed by this process with the organic matter. Once exposed to the air, the preserved organic matter in the swamp would rot, creating another source of some nutrients and improving physical conditions. The organic matter and manuka would also burn if fired, adding nutrients from the ash. Such beds would have been ideal for kūmara, and the bases and sides of the drains themselves would have suited taro.
This process was labour-intensive and appears to feature mainly in the far north around Kaitaia and the Bay of Islands, and in places on the lowlands of the Bay of Plenty. In these localities climate, particularly temperature, was not so much a limiting factor. However, there were other limitations on gardening in the wider region. In Northland the soils on the country surrounding the swamps are of very poor fertility. They had long had a cover of kauri forest, and the combination of warm temperatures, rainfall and the acidity of the soils under such forest had chemically removed the nutrients from the soil. The swamps on the other hand, particularly those formed on estuarine soils, had a cover of kahikatea, which demands good soil fertility. As a result, once the kahikatea was cleared, the fertility was high and gardening on areas of swampland was particularly highly developed.
Another important place for gardening was on dunelands. These gardens were on soils created on the sheets of sand, disturbed initially by human firing, driven by wind inland from the coast. Such soils were distributed over much of the North Island west coast from the Manawatū to Ninety Mile Beach. Other smaller areas were on the east coast of the North Island. Within a couple of hundred years of the soil's formation, they would have had good covers of nutrient-rich manuka, kanuka or other coastal forest. The nutrients could be released by further firing. What may now appear to be a 'desert', the coastal strip of Ninety Mile Beach from about 1 to 5 km inland to the east, was once a complex, if patchily used, gardening system. 7
The importance of gravelly or stony soils has already been stressed. The gravels could be added, or naturally stony ground could be selected, and enhanced by removing the larger stones and placing them in rows or mounds. Such techniques were extensively used in the volcanic soils of Auckland and Northland, with smaller areas of volcanic soils at Cape Runaway in the eastern Bay of Plenty. Of course, these soils were both friable (because there had not been sufficient weathering to create clays) and of high natural fertility. Similar horticultural practice was common on the volcanic soils of Central Polynesia and Hawaii, 8 and the potential of such soils would have been readily recognised by the first Polynesians entering New Zealand. 9
Gardens on stony soils of non-volcanic origin are common on the coastal strip in Hawke's Bay, the Wai-rarapa and the north-east of the South Island. 10 The coastal strip had no frosts, and had readily warmed and friable soils. Stone rows are rare inland, but they do exist, for example, 15 km inland in the Tukituki River valley, central Hawke's Bay. 11 In these cases, Māori built extensive lines and areas of mounds and stone rows. At Palliser Bay, such soils appear to have been gardened early in pre-European history. The Palliser Bay case is particularly interesting because here the mapping of the stone rows over a coastal strip some 20 km long and up to 500 m wide was done using purpose-flown, low-level aerial photography. The photographs were used as a map base and supplemented by Helen Leach's detailed field observations, 12 although none of the photographs have been published.
Radiocarbon dates at Palliser Bay go back as early as the twelfth century, but most are in the fifteenth century. 13 The abandonment in pre-European times of the garden areas in Palliser Bay that were evidently once valued has been much discussed. This area approaches the climatic margins of kūmara horticulture. Another 2° of latitude (or 400 km) south, a drop in mean temperature with the corresponding shortening of growing season, and horticulture was not possible at all. 14 It has been argued by Foss and Helen Leach, the original investigators, that a fire-induced reduction in forest and increased erosion, exacerbated by a long-term deterioration in climate, led to climatic marginality for gardening and the destruction of fishing grounds. 15
Stone mounds, part of the gardening system at Pouērua
The many small mounds are about 4 m across at their base. The functions of such mounds are difficult to determine. They may have served to clear the stony soils of the old lava flow so that the friable and fertile soils that resulted were more easily tilled. Alternatively, they may originally have consisted of heaped-up mounds of soils, organic compost and stones, on which plants were grown. The topsoils have subsequently been eroded away from the surface leaving a mound of stones which does not look as if it could have been gardened. A further possibility is that the vines of plants such as kūmara or yam were trained over the mounds. At top are some larger lava domes (about 25 m across). These have had stone stripped from their surfaces and placed in walls on their perimeter. At bottom and bottom right, a heavier fall of ash has reduced the need to create stone mounds. Here there are barely perceptible lines in the small valley floors which are either trench garden boundaries or track ways. Also showing are lines of stone walls constructed in the nineteenth century.
Stone rows delineating garden boundaries near Okoropunga, on the Wairarapa east coast
The individual patches are 10 to 30 m across and 150 m from the foot of the steep slope to the sea. This view is taken looking to the south-west. The appearance of rows running parallel to the beach, further dividing the gardens into rectangular plots, is somewhat misleading. The lines parallel to the beach are ridges of stones thrown up by Storms at the back of former beaches. Over thousands of years, these beaches have been lifted by earthquakes out of the reach of wave-action. The main stone row boundaries have been constructed at right angles over the line of the beach ridges, although some of the ridges themselves may have been enhanced and also served to mark the boundaries of gardens. Some of the stones in the garden rows were taken from the depressions (borrow-pits) which show along the top edge of the principal beach ridge. Throughout the Wairarapa, the original topsoils that existed in these areas have been removed by fierce squally winds, thus causing the abandonment of the gardens. In their original form the gardens would have been fringed with coastal forest and protected from wind.
In stony areas the presence of mounds, rows and other features offers the enticing prospect of furthering the understanding of Māori gardening practices. We know that Polynesians were very particular about the arrangement of plots and the paths through them. One reason was the tapu that attached to growing plants at certain stages, and there may have been more general considerations of what would now be called 'property'. The rows and mounds therefore had functions other than just the placement of stone waste removed from the plot itself. The stones are often arranged in definite patterns, with the rows running down the slope. In this pattern they are thought to reflect a typical Polynesian subdivision of the landscape into individual or whānau plots. 17 These garden plots come close to offering a sense of property, however ambiguous and difficult to interpret, in pre-European society. Helen Leach has estimated that, for the Auckland area, plots ranged from 80 to 300 m long and 25 to 60 m wide, i.e., from 0.2 to 1.8 ha. These larger areas were subdivided into smaller areas of about 10 by 20 m (0.02 ha). At Anaura Bay on the East Coast the plots recorded by the Endeavour observers were up to 2 ha in area and subdivided into plots of 0.04 to 0.06 ha in area. 18 Some of the aerial photographs presented here show plots of just these sizes, for example, those at Okoropunga on the Wairarapa coast.
Other purposes that have been suggested for the stones set aside from fields are: field shelters or houses; 19 'shelter belts' for the crops; markers for the sides of paths (important in avoiding breaches of tapu); 20 areas of particularly high ground temperatures for the propagation of young plants in spring; 21 and finally, elevated, well-drained places over which the running vines of yam or kūmara could be placed.
Drains and trenches
Mainly in Northland, but also as far south as Kawerau in the Bay of Plenty, drains or trenches are a feature of gardening areas. The functions of these are also controversial. Were they indeed drains, pure and simple, or trenches which served to mark boundaries where stone or scrub was not available? 22 Did the trenches offer unique gardening conditions, planted in their own right, for example, with taro planted at the bottom? A drainage function may seem obvious, but it was probably only one of several functions. In Northland drains or trenches occur in open swamps, at the foot of hill slopes and in small basins in volcanic stonefields. In the far North near the Awanui River, they lie in two broad patterns: one is a deliberately artificial, more or less rectangular pattern of some intricacy, delineating plot boundaries as well as serving to drain. The other is the simple modification of natural drainage features such as rivulets or erosion runnels. 23
Drainage pattern for gardening on the alluvial flats of the inner Rangaunu Harbour, near Kaitaia
The lines that show in the soil at this time (August 1944) were originally trenches or drains constructed by Maori, probably in the seventeenth or eighteenth century. They appear to follow and enhance a natural drainage pattern, showing clearly where natural streams enter the sinuous inner reaches of the estuary. The ditches were probably used for taro while the beds between were for kūmara. The ditches would have filled with dark, organic-rich mud over time, which would be periodically thrown on the gardens. When the ground was ploughed deeply, perhaps for the first time in the 1940s, the pattern of the channels has shown against the lighter coloured subsoil. Modern European drains have been cut over the older drains, and form a different and more sharply defined pattern. The river is about 16 m wide, and the typical trenches are 1 to 2 m wide.
The area shown is on the lower tidal reaches of the Waipapakauri and Waimanoni Creeks, about 2 km west of the Awanui River; mangrove trees show on the edges of the river. The airfield was a wartime measure and no longer exists.
At least two pā defended by double ditches and banks may be seen. One is at top left; the other is at the eastern end of the light-coloured debris from the creation of the prominent west to east drain. Other possible pā lie on each side of the river just to the south of the prominent drain.
Trenches on a slope near Marsden Cross, Bay of Islands
A particularly fine example of a complex of trenches. The lines of the trenches run about 25 m up the slope; the trenches are about 1 m across and lie at 5 m spacing across the slope. The trenches may be a type of crude irrigation, i.e., irrigation without elaborate water reticulation to the garden plot. Intercepting groundwater towards the foot of the slope, they may form artificial springs which would assist the growth of taro in summer and late summer dry spells. The taro would be planted actually in the trench, hence the close spacing. This particular example has a perimeter ditch running around the top of the trenches showing clearly at right.
One of the often-cited achievements of Māori horticulture was the development of kūmara storage. Kūmara was stored in pits dug into slopes, in ridge crests, or the flat surfaces of areas in or close to sites of occupation. 27 The function of storage pits was to provide a storage space of even temperature (never less than about 4°C), waterproof, well drained and rat-proof. Kūmara will be destroyed if exposed to temperatures less than 4°C, and would not survive if left in the open or in raised-house page 71 stores throughout the New Zealand winter. No doubt general security in the event of warfare was also a factor. Indeed, open rectangular storage pits are often in apparently awkward positions among the ditch and bank defences of pā. They were perhaps placed there to amplify the defences and not just to have them out of the way of the living spaces within the pā.
Storage pits are very distinctive from the air. On the East Coast and in Taranaki, the Waikato and Hawke's Bay, they appear as open depressions, rectangular in shape, as much as 8 m long and 4 m wide. Typically they measure about 4 by 2 m in plan and are 1 m deep. The depths are fairly consistently at just over 1 m, suggesting that the tubers were stacked one on top of the other rather than on racks; any deeper than 1 m and the lower tubers would have been crushed or bruised leading to rapid decay. Pits may have an earthed-up rim around the outside edge of the depression, with a drain outside the rim leading away to the edge of the ridge. The raised rim may be the remnants of the earthed-over roof of the pit (now decayed and collapsed into the pit cavity) or a bank placed under the eaves, while the drain served a function similar to that of a gutter at the lower edge of the roof. When excavated, the bottoms of pits are commonly found to have a central row of postholes at their base to support a ridge pole for a roof. There may also be an interior drain, at the foot of the earth walls, designed to carry away water from the stored crop inside the pit. Sometimes pits were cut from clay topsoils down into porous sandy subsoils, again offering drainage at the base. 28
In regions with deep, easily worked subsoils, such as Taranaki or the Bay of Plenty, the completely subterranean cave pit was more common. In these, the roof of the pit (the immediate subsurface of the ground) is dug in such a fashion that it curves over like the top of a bell. This makes the roof of the pit very strong. Underground, the pit was circular or oval in plan and could be as much as 2 to 3 m wide by 2 m deep. The vertical passage into the top of the pit was about 30-60 cm in diameter and sealed by a tight-fitting lid, so that the whole structure did not lie above the general level of the ground. 29 This type of pit, known as a 'rua' by archaeologists, was also cut into the defensive banks of pā, in which case the entrance passage was horizontal. These pits were very economical of surface space within a constricted area such as a pā. Some pits also appear to have been used to gather and store water. Today, such pits are generally not found open at the top, except in areas where the forest has grown back over the site and prevented erosion and stock tracking. Typically the whole of the pit collapses in under the weight of stock and leaves a distinct more or less circular depression up to 2 m in diameter. In some places in Taranaki, the Waikato or the Bay of Plenty, the whole surface of a pā may be pockmarked with such depressions.
There is now general agreement that pits were not for houses, 30 as was once thought. The storage function was generally recognised as late as the 1920s, 31 but as pits fell out of widespread use a 'pit dwelling' theory took hold, originating particularly among scholars working in the South Island. 32 The accepted interpretation, kūmara storage, nevertheless raises some unresolved questions. Were pits in use from earliest times, as the existence of their Polynesian precursors would suggest? No pit found is as old as the earliest settlement contemplated in New Zealand. The oldest examples are about 700 years old at Skippers Ridge and Sarah's Gully in the Coromandel and 600 years old in inland Tolaga Bay on the East Coast. 33 Most pits when excavated and dated by radiocarbon prove to be younger. The late age for pits has been used to suggest that horticulture was introduced to New Zealand some time after first settlement, after large animals like the moa had been hunted to extinction. 34 Such a view is not generally accepted today.
There are several further points that may be made. First, the regional distribution of pits follows very closely the distribution expected for gardening on soil and climatic grounds. This in itself points to their function. Second, when excavated, pits are often found intricately cut into one another, showing that pits were deliberately filled back in, and new ones cut at a slightly different location but overlapping the former one. Estimates have been made that only a small proportion of the total number of storage pits can be seen from the surface. 35 The purpose of cutting pits again and again was to remove the risk of pathogenic fungi establishing on the walls, and passing the infection from one year's crop to the next.
Whenuanui, a pā with many raised-rim storage pits on the high terrace of the Waipāoa River near Pūhā, East Coast
In this oblique view, raised-rim pits show clearly within the defended area of the pā. A Pleistocene (ice-age) river-terrace surface has been cut by gully erosion (foreground) and the main river (in distance), leaving a narrow neck at the right of the picture. This neck has been defended by two sections of ditch and bank, only one of which shows here. The view is to the south-west in afternoon light. The largest of the pits is about 7 m square. The main defended area is about 100 by 200 m in extent.
Overall then, the value of aerial photography in documenting horticulture has been considerable, with perhaps better established results than in any other area of application of the technique in New Zealand archaeology. Because of gardening's extensive nature, the difficulty of seeing patterns from a ground view, and indeed the destruction wrought by modern horticultural practice, aerial photography has given insights to a level of detail that is only matched by the first written observations from the Endeavour in 1769. 37
1 Other species were coconut, banana and tī, see H.M. Leach (1984: 17-32).
2 Corynocarpus laevigatas; Freycinetia baueriana, banksii (the latter genus of the same family as the tropical pandanus).
3 H.M. Leach (1984); E. Best (1925b).
4 More rarely, on the surface of the terrace.
5 E. Best (1925b: 123); Walton (1983); Fox and Cassels (1983).
6 Barber (1989).
7 K. Jones (1984a; 1988b); Walton (1983).
8 Kirch (1985: 215-231).
9 For gardening on stony soils, see H.M. Leach (1979; 1984); Bulmer (1987); K. Jones (1989a).
10 McFadgen (1980a, b).
11 My field observation.
12 H.M. Leach (1979).
13 H.M. Leach (1979: 144-157).
14 Compare the annual mean temperatures for representative coastal locations such as Wellington Airport (13.3°), Para-paraumu Airport (12.9°), Nelson Airport (12.1°), Kaikoura (12.1°), Onāwe (Banks Peninsula) (12.4°), on the one hand, and Timaru (11.2°), Ashburton (11.2°) and Oamaru Airport (10.6°) on the other (New Zealand Meteorological Service n.d.).
15 H.M. and B.F. Leach (1979); B.F. Leach (1981).
16 K. Jones and Law (1987).
17 H.M. Leach (1979).
18 K. Jones (1989c: 60).
19 Bulmer (1987).
20 H.M. Leach (1979; 1984).
21 McFadgen (1980a).
22 I. Barber (1989).
23 Johnson (1986: 163-176) distinguishes between 'modified' or 'extended natural hydraulics' and 'artificial channel hydraulics'.
24 The soils were a dark colour and had gravel, sand, shell and charcoal mixed in (Peters, 1975).
25 Photograph in chapter 7.
26 Yen in Nicholls (1965); Peters (1975).
27 E. Best (1916: 77-108); Golson (1959); Green (1970); Parker (1962); Fox (1974).
28 For example, K. Jones (1984a).
29 See rua kopiha, E. Best (1925b: 227).
30 Davidson (1984: 121-127).
31 For example, E. Best (1916; 1925).
32 For example, Duff (1961).
33 Davidson (1984: 121-127); K. Jones (1989b: 251-255).
34 For example, Green (1970: 12-15).
35 K. Jones and Law (1987: 86).
36 For example, K. Jones (1986).
37 K. Jones (1988b).