Volume 75 1966 > Volume 75, No. 3 > Umu-ti, by Hardwicke Knight, p 332 - 347
                                                                                             Previous | Next   

- 332

Large pit ovens such as those discussed in this article are in evidence in Otago and South Canterbury. An archaeological and ecological description of examples of these from a selected area on the Otago Peninsula is first given, showing that they serve a special purpose and distinguishing them from other types of oven; they are then compared with large Maori ovens for cooking the stems and tap-roots of the Cordyline described by Shortland, Stack, Tikao (quoted by Best), Hay, and others. Finally they are compared with the umu-ti of Eastern Polynesian and Samoan cultures.


OVEN S.164/305778 (Grid reference to 1: 63,360 map, New Zealand).

This oven is situated on a hillside 350 feet above sea level. The plan is a perfect ring with an outside diameter of approximately 6.3 m. The diameter at the highest part of the rim is 4 m. A NE-SW diametric cross section showed a raised rim, with a maximum height of approximately 50 cm. above the original soil horizon. Charcoal and red fired clay appeared in the rim. At 25 cm. below the surface at the highest part of the rim is a hard clay, disturbed but free of charcoal and fired clay, and apparently forming the throw out from the oven pit. The original horizon was detectable as a gritty layer at a depth of another 25 cm.

The oven pit contained 400 stones of an average weight of 10 kg. (minimum 5 kg.- maximum 15 kg.) The stones are set flat against the sides and bottom. In the centre, the stones were found to be packed on edge, tightly. Charcoal was found behind the flat stones and against the wall of the pit. The lower part of the wall retained digging tool marks in the form of grooves 2.5 cm. wide and right-triangular in section. The diameter of the floor of the pit is 1.5 m. The floor is flat, and natural clay was found at a depth of 1.6 m. below the top of the rim. The sides of the oven pit are almost vertical.

- 333

The raised rim was excavated in one quadrant and the material removed was sifted. Neither bone nor shell was traced, nor were faunal remains found in the oven pit. The stones removed from the pit were examined after rain had washed them and they had dried in the sun, and observed to be bright red, and pitted.

OVEN S.164/305775

The oven is situated on a saddle, 700 feet above sea level. The plan is a perfect ring with two dimples (which are described later in this article) adjoining it. The rim has an outside diameter of approximately 9.8 m., and the diameter at the highest part of the rim is 5.5 m. A NE-SW diametric cross section showed a raised rim, with a maximum height of approximately 50 cm. above the original soil horizon. The sectioned rim showed layered material that had been thrown up when the oven pit had been re-used and dug deeper to increase its capacity.

At its first use, the oven appears to have a rim with a diameter of 3.25 m. at its highest part. On cleaning out the pit for re-use, the charcoal and burnt clay appear to have been thrown behind this first rim. The floor would then appear to have been dug deeper and the clean clay was thrown on top of the charcoal and burnt clay to form a wider rim which has at its highest part a diameter of 5.5m. Over this clean clay in the rim there is a further spread of charcoal and burnt clay. This may result from uncovering the oven on its re-use, or it may represent a third use of the oven.

The oven pit contained 2,200 stones of an average weight of 6 kg. (manimum 4 kg.-maximum 16 kg.). Among the uppermost stones, lumps of red fired clay, looking like brick, were found, suggesting the use of clay to clam up the space between the stones forming the top layer. The lumps bear long, shallow impressions, 1.5 cm. wide, that fit the fingers. (Microscopical examination shows no trace of finger-prints). The depth of the pit is 1.2 m. below the surface at the centre of the oven and 1.7 m. below the highest part of the rim. The bottom of the oven pit is flat, and 2 m. diameter. The pit walls are almost vertical.

Excavation of the adjoining dimples showed that the charcoal and red fired clay from the oven formed a layer across the depressions compatible with the depressions being earlier than the oven.

OVEN S.164/304776

The oven is situated on a shelf on a rocky hillside at an altitude of 650 feet above sea level. The plan is a perfect ring with an outside diameter of approximately 9 m., and a diameter at the highest part of the rim of 5 m. A N-S diametric cross section of the raised rim showed a more complex stratigraphy than the previously described ovens. However, the bottom of the oven pit is dug into a distinctive light-coloured volcanic stone, and the occurrence of this in the raised rim facilitates the interpretation of the stratigraphy.

The pit is approximately the same size as that of oven S.164/305775. The number of stones is estimated, on the quadrant excavated, to be 2,000. Lumps of red fired clay are present both among the upper stones - 334 of the pit and in the top layers of the rim. The walls of this oven pit have been markedly affected by heat to a varying depth of from 15 cm. at the top of the wall to 2 cm. at the bottom of the wall. The floor of the pit is similarly affected to a depth of 2 cm. This wedge of fired clay of the pit walls is of two different textures and colours. At its deepest it is a reddened clay that feels wet. The surface exposed nearest to the oven fire and towards the top of the wall is a deep plum colour and feels gritty. The pit walls are almost vertical.

OVEN S.164/307773

The oven is situated not far from the trigonometric station on the hilltop, on almost level ground at an altitude of 750 feet above sea level. The plan is a perfect ring with two dimples. The ring has an outside diameter of 9 m., and the diameter of the rim at its highest part is 7 m. Excavation showed that the raised rim was composed of the same soil as that of the surrounding terrain and was distinguishable only by being less compact. There was no charcoal or burnt clay on the outside of the raised rim, but charcoal was found on the inside of the rim and among the stones of the oven pit. Although not completely excavated on account of water seepage (July) the pit and its stone content appeared similar to ovens S.164/305775 and S.164/304776 previously described.

It would appear that this oven was used once, and the culinary contents removed without spreading much charcoal or burnt clay.

Comment on the Excavations

In addition to the four excavations described, quadrant excavation of a number of others has been undertaken. The following list shows the principal dimensions of the four fully described ovens together with four others from the selected area.

Reference N.Z. Map 1:63360 S.164 Altitude a.m.s.l. feet Over-all diameter metres Diameter at top of rim metres Diameter of floor of pit metres Depth of pit from top of rim metres
305 778 350 6.3 4.0 1.5 1.6
305 775 700 9.8 5.5 2.0 1.7
304 776 650 9.0 5.0 2.0 1.5
307 773 750 9.0 7.0 2.5 1.5
314 778 350 7.5 4.5 2.0 2.0
310 780 100 8.0 4.5 1.5 1.8
308 773 750 6.5 4.5 1.8 1.6
323 785 650 9.0 5.0 2.0 1.5

Examination of the stratigraphy of the rims of twenty-two ovens of this general type (the raised rim being a corollary, as it were, of the deeply dug pit) on the Otago Peninsula suggests that some ovens were used once, some twice, and some more than twice. Site records of these have been filed with the New Zealand Archaeological Association. It has been found that all of these ovens so far examined conform nearly to one

- i
An umu-ti on the Otago Peninsula.
Air photograph showing an umu-ti in proximity to a creek and in association with dimpling and fallen broadleaf.

- ii Page is blank

- 335

or another of the four examples fully described. All have characteristics in common, the most marked being:—

  • An almost perfectly circular raised rim and deep oven pit.
  • The rim raised 50 to 80 cm. above the surface at the centre.
  • The pit walls almost vertical (95°-100°.)
  • The clay of the pit walls shows in varying degrees the action of heat.
  • The stones in the oven pit are covered with soil. (If stones are visible it is usually due to the oven pit having been used for cattle or sheep burial).
  • Lengths of unbroken charcoal (including Griselinia littoralis) found at sides and at the bottom of the oven pit.
  • Fragments of burnt clay in the open pit and in the rim.
  • Charcoal embedded in the thrown up clay in a manner that suggests the charcoal was thrown up on the clay rim while the clay was still lying in lumps.
  • Complete absence of faunal remains from pit, rim, and vicinity.

In two instances has a farmer reported finding a greenstone adze within a stone's throw of an oven, but during the excavations not a single artefact or shell or piece of bone has been found.

The appearance of the stones found in the ovens suggests that very high temperatures have been reached. The stones are selected, none is abnormally large or small. The stone is the same as that which outcrops in the vicinity, but it is not now found in such convenient sizes. It does not appear that the stones have fractured in the oven as a result of the firing. Assuming a vegetable cooking purpose of these ovens, the conversion of the Cordyline root and stem into an assimilable carbohydrate can be achieved at a low oven temperature, but it is found that prolonged heating increases the sweetness. The pitting of the oven stones is taken as evidence of the use of water, the action is therefore comparable to the prolonged boiling of the sugar juice in both cane and beet-root sugar technology. The placing of the stones in these ovens is as left after the last use. The stones are closely packed on edge. They may have been packed by hand, or more probably shuffled into this pattern by the movement of a heavy stick in the centre. The stones are found to be so tightly wedged in the oven pits that a crowbar is required to remove them. Ovens constructed for as specialized a purpose as cooking the Cordyline or Ti, the generic Polynesian term for it, would be unlikely to be suitable for cooking anything else. In Otago, at all events, there would have been no other vegetable supplying essential carbohydrate as economically as the Cordyline australis. Anderson 1 was making use of a familiar tradition when he made one of his characters say: “You will remember, too, the tale of Rongo', how he came to the lands more towards the south, and found a people who had no kumara, but ate only aruhe and ti-kouka.

The only positive archaeological identification of the ovens of this type is excavation, the primary criterion being a steep sided pit, the walls of which show the effect of heat. The raised rim and its stratigraphy and the large selected oven stones are important criteria when they exist, but the rim may be removed by ploughing, and the stones to make room for a cattle burial.

- 336

The Environment

A study of what would have been the ecological environment of the human activity associated with these ovens has to be based on surface remains and on reports of descendants of settlers of the 1860's, particularly those who are today farming their fathers' and grandfathers' land on the Otago Peninsula.

It has been ascertained that the location of the ovens is in every case in proximity to a former growth of cabbage trees (Cordyline australis); Broadleaf (Griselinia littoralis); Kowhai (Sophora microphylla); Lancewood (Pseudopanax crassifolium), and Kotukutuku (Fuchsia excorticata). A few straggling cabbage trees among outcrops of stone remain. The broadleaf can be traced in the fencing posts, and is also found laid end to end where it was dragged to form fences by the settlers who cleared the ground. The kowhai turns up in the form of tenoned fencing rails, and still grows in the vestigial bush. Many of the ovens are at the edge of areas of bush, and all are associated with fallen broadleaf trees still lying on the ground. The proportion of these trees that had been deliberately felled, as evidenced by sawn stumps and trunks, to those that have fallen naturally, varies in different areas of the Otago Peninsula. In the selected area, the majority of recumbent trees appear to have fallen naturally. It is interesting that users of the land state that the high ground to the south-west of Weirs Road, on the Otago Peninsula, grows totara (Podocarpus totara) and that ovens are rare, while the high ground to the north-east of Weirs Road grows broadleaf but not totara, and is the area where the greatest number of ovens have been found. Areas where pine was felled for milling during the 1860's show no evidence of ovens.

The Ti Plant

Ti is a generic term for the Cordyline. Each of the species of the Cordyline has up to three Maori names, so that a total of nineteen native names is listed by Best. 2 There are a dozen species throughout India, Malaya, Polynesia, including New Zealand. Cordyline australis is the common species in New Zealand, but is rare in Stewart Island. It is found from the sea level to about 2,500 feet. Maori and other names are: Ti, Ti-kouka, Ti-kauka, Ti-whanake, Ti-rahau, Palm Lily, and Cabbage Tree or Palm. It flowers from November to January. The seeds ripen by April. If the stem only or a portion of the tap root of young C. australis were taken, the plant would not be destroyed, the remaining parts of the plant would regenerate, and the tree would be ready for plundering again after four years.


All the deep pit ovens observed from the air and found during surface exploration are in areas where there is dimpling. Botanical opinion is that although these are circular and not crescentic as is usual - 337 with tree dimples, they could be caused by fallen trees. They do not appear to be sinks, and there is no Karst topography. Neither ovens nor dimples occur in the present areas covered with bush. Distribution in open areas can best be studied on air photographs. It can be shown on air photographs taken obliquely from 1,000 feet with the sun at a low angle that on areas that have been ploughed adjacent to dimpled areas, the dimples can still be traced, although they are not detectable from the ground.

Excavation of dimples within the area of charcoal spread from an oven (undertaken in connection with the examination of ovens S.164/ 305775 and S.164/307773) has shown that the dimples can antedate the ovens. The dimples vary considerably in diameter and depth and may have a slightly raised rim. In a sample hectare (10,000 sq.m.) in the selected area, including one oven, there are 30 distinct dimples, varying from 1 m. to 3 m. in diameter, from 10 cm. to 30 cm. deep, and with rims up to 20 cm. high. They are fairly evenly distributed in the sample area, as they are throughout the larger selected area, and occur singly, in two's and three's. In addition there were perhaps as many indistinct dimples.

Since the special type oven is always found associated with a peculiar dimpling, and vice versa that wherever the peculiar dimpling is noted an oven can be detected, it was postulated that there must be a probability of cultural association. An experimental area was found near Roxburgh, Central Otago, where C. australis had propagated on ground that was known to be clear thirty years ago. Here the cabbage trees with stems of from 3 feet to 6 feet high were dominant among low scrub, and an area of one hectare was chosen where there were near enough 30 trees. A photograph was taken, and the trees plotted, each tree being represented by a circle of 2 m. diameter on the ground. The trees were growing singly, and in two's and three's, and the similarity between the graphic representation of the experimental area of living cabbage trees and the sample area of dimpling was close.

An attempt was made to dig out the tap-root of a tree with a stem of about 5 feet high. It was found that, with a steel spade, it was necessary to dig a large hole on one side to enable a 2 ft. length of taproot to be removed undamaged. It is possible that with Maori digging tools it might be more effective to open out the ground on all sides of the desired part of the root, in which case the displacement of the soil would, after settling, resemble the dimpling noted.

An overall statistical count of the proportion of dimples to ovens in the selected area has not been attempted because of the weak archaeological evidence as to the number of times certain ovens in a definite area have been used.

European methods of bush clearance have been considered. The larger trees, mostly broadleaf, that grew in the bush were saw felled and the stumps left in the ground for a number of years. They were eventually dug out. This work was done by contract on the Otago Peninsula during the first intensive period of land clearing. Sawn and lifted stumps are to be found in certain areas, although the process might be expected to leave dimples, no resulting dimpling can positively be demonstrated.

- 338

On land where the broadleaf trees lie most thickly, they are found to be complete with their roots, and characteristic crescentic dimples are present.

In general, the peculiar dimpling, and the ovens that are associated with this, are on land that has had minimal European interference. The dimples are frequently seen too close together to be left by large trees. However, the almost universal opinion of the farmers is that the dimples, both circular with raised rims and crescentic, are caused by fallen trees.

Topography and Surface Archaeology

Topographical features shown on air photogarphs, taken from an altitude of 1,000 feet, and confirmed by subsequent examination of the ground, are:

Spurs and raised ground. The ovens are seen to be located in these selected areas and to be dug in clay.
Outcrops of stone. The ovens are found close to exposed stone, some of the outcrops appearing from the air as eroded cones. At lower altitudes where ovens are situated beside creeks, stone is exposed in the bed of the stream.
Creeks. The ovens are close to creeks which run in deeply worn gullies. Although air reconnaissance was undertaken in the winter, it appeared that the ovens occurred only where the hillside was sunlit.

Apart from an occasional artefact found during ploughing or fencing, and burials, there is no archaeological evidence of occupation of the elevated country of the Otago Peninsula. The ovens stand out as isolated earthworks. There are shell, fish, and moa middens on the flats bordering the sea at the foot of the hills on which ovens occur. On Okia Flat there is an occupied cave site. On Allans Beach there is a spit site with moa cooking ovens. On the cliff-top at Tarewai there is the remains of a village which may be one of those noted by Brees 3 on his plan of the Otago Harbour made in the 1840's, and excavated by Teviotdale. 4

On the Pacific coast of the Otago Peninsula there is a series of deep pit ovens on raised spurs along the course of a creek at an altitude of between 200 and 300 feet above sea level, and 500 m. from the sea. On the cliff-top where this creek falls into the sea, there is an occupation site on wind-blown sand. There are also deep pit ovens on high ground also within 500 m. of the occupation site. These features are related only in that there are no other occupation sites or deep pit ovens, along the coast within 1.5 km. in one direction and 2 km. in the other, or within a much greater distance inland. The occupation site reveals conventional cooking ovens, fish and shell middens, small quartzite flakes, but no controlled excavation of the site has been made.

At Portobello, an area from which some of the most refined green-stone artefacts have been recovered, there are, in addition to cooking ovens with fish and shell middens, deep pit ovens which are in some respects miniature editions of the large ovens of higher altitudes. The - 339 diameter of the rims of these smaller ovens approximates to 1.5 m. and the stones number not more than one hundred. It is possible these may be family ovens used for cooking the stems of tap-roots of the ti, and that the large ones located where the crop was abundant may be regarded as tribal or communal undertakings.

Considerable labour must have been involved in digging the large pit ovens, collecting the stones, firing the oven, and harvesting the stems or tap-roots. The oven-stone capacity of one of the larger ovens is calculated to be 13.2 metric tons.

Water Sources

The sources of water for use in the process of cooking the ti roots are creeks and springs. A study of these sources in association with the large number of umu-ti examined on the Otago Peninsula shows that in many cases there has been damming. In the case of springs, it is not often possible to distinguish Maori work from that of the European settler. It would appear that a higher than normal expectation number of dammed springs occurs in close association with umu-ti. The damming of creeks, however, where it occurs at the closest point to the umu-ti, is almost certainly Maori work. The damming consists in putting across the creek earth cut away from the banks in order to form a reservoir. The cutting out of the bank remains, but the earth used to form the dam is in all but a few instances completely washed away. Where large stones have been used to form the foundations of the dam, these remain, and at one place are found carefully laid.

These operations suggest that a considerable quantity of water was required. It has been noticed that where the creek is today a small one with little water, the umu-ti is often found to be placed between two water sources.

New Zealand References

Some of the early writers mention ovens similar to those described in this article, and appear to be familiar with the Maori practice of cooking the stems and tap-roots of the Cordyline. Shortland, 5 in his Southern Districts of New Zealand, mentions a Maori bringing him a “basket of kauru or baked root of the ti” and observes that “the natives have learned to dig it at the season when it contains the greatest amount of saccharine matter . . . just before the flowering . . . they then . . . steam it in their ovens.” Hochstetter 6 notes that “there are several varieties of this tree (Cordyline) all of which have long tap-roots, which the natives cook,” and Stack, who is quoted by Best, 7 gives a detailed description: “The kauru was prepared in the summer months from the cabbage palms . . . young trees about five feet high were selected . . . the stems were cut into about two feet lengths . . . an oblong pit was dug . . . from four to twelve feet in length and about five or six feet in depth . . . a quantity of stones - 340 were placed in the bottom and firewood piled upon them, which was afterwards lit, and, when consumed, the pit was filled in with the prepared ti palm stems, which were covered with matting and soil . . . a quantity of water was then procured in buckets formed with flax leaves and pourned into the pit, the bottom of which was covered with the heated stones . . . the steam generated was prevented from escaping by a sufficient quantity of soil being heaped upon the mat covering the pit.”

If the clay found during excavations was used to clam up the oven to retain the steam (taopakatia) this would explain why the baked clay is also found in the top layer of the rim where it would have been thrown out when the cooked food was removed.

Best 8 also quotes Hone Tikao, who informed him that “the ti was not cultivated by man, but grew spontaneously in many places”. Best agrees with his informant, so far as the South Island and the C. australis are concerned, but thought it probable that some species were introduced by the Maoris into New Zealand. 9 Best was informed by Tikao that plants about four feet in height were selected, the outer parts of the stems were chipped off, and the stems were cut into lengths of about two feet. He is also informed that when the tap-root of the ti para (one of the most valued species of ti) was taken, the lower part of the root was left in the ground that it might grow again, 10 and he quotes Colenso that young seedlings were carefully selected and planted out.. 11

According to Tikao “this kind of food was prepared twice a year, in the months of November and February. The ovens of each family were prepared at dawn, the trenches used as steam ovens being very long. The cooking occupied about twenty-four hours. There was a considerable amount of tapu pertaining to the cooking of this food . . . the stumps (of the Cordyline) again grew and, in four years, another cutting would result.” Best 12 adds that “native evidence tends to show that in no case was it left in the steam pit oven for less than twelve hours . . . some state that forty-eight was the time allowed”. Best concludes that the steaming pits in which the ti was cooked would be called umu-ti.

James Hay, one of the older colonists who formed a committee in 1909 to put down valuable and authentic data for the historian, 13 writes that when he was a boy he was much interested in the customs of the Maoris of the Banks Peninsula, and that ‘in the forties’ the Maoris had a method of extracting sugar from young cabbage trees, which, I fear, is now lost. They began operations by digging a hole 8 ft. long, 4 ft. wide, and from 5 ft. to 6 ft. deep. A layer of stones was placed in the bottom, and on them an enormous fire was built. When this had burned down the young cabbage tree was stripped and laid on the stones. Water was then poured over them, and all was quickly covered over with earth and left for many days . . . This native-made sugar was troublesome to produce, and when ordinary sugar could be procured they gave up preparing it, and I believe the art is now lost.”

- 341

Colenso 14 notes that “the large tap-root of this plant (C. australis) was dug up and split and cooked for food; it was very fibrous, yet contained a large amount of both saccharine and farinaceous substance. It took very long in cooking . . .” Rogers 15 commenting on the occurrence of umu in the Warepa district of Otago, in 1922, found the ovens were “nearly all 6 to 8 feet in diameter and paved with volcanic stones reddened by heat” and that they were “all in sheltered positions, within 20 feet of water . . . a few on hillsides and two on top of a hill . . . occasionally groups of three, the largest being 12 feet in diameter and the smallest 6 feet across, well paved with large stones, smaller lining the wall, and a lip round the outside, varying from 9 inches to 1 foot in height.” On the authority of Hoani Kaahu of Arowhenua, quoted by Anderson, 16 ti trees and umu-ti were in great abundance near Temuka, South Canterbury, the Ngai-Tahu people claiming that the name of the place derives from Te umu kaha, place of many huge earth ovens for cooking the ti. Similarly, the Honolulu suburb of Kaimuki (ka imu ki) is said to have obtained its name from the large oven (imu) for cooking the ti (ki) root prepared there by the community.

Analysis of the N.Z. References

The information contained in the foregoing references which is particularly relevant to the study may be itemized as follows:—

root of the ti (Shortland)
tap-roots (Hochstetter)
tap-root (Colenso)
ti palm stems (Stack)
stumps grew again—implying the stems only were taken (Tikao)
young cabbage tree (Hay)
dug at season when sugar content highest (Shortland)
prepared in November and February (Tikao)
young trees about five feet high were selected (Stack)
in four years another cutting (Tikao)
plants about four feet in height being selected (Tikao)
split and cooked (Colenso)
cut into two feet lengths (Stack)
outer part of stems chipped off, stems cut into lengths of about two feet (Tikao)
stripped and laid on the stones (Hay)
steamed in ovens (Shortland)
water poured into pit and steam retained by coverings (Stack)
ovens within twenty feet of water (Rogers)
steam ovens (Tikao)
steam pit oven (Best)
water poured over the hot stones (Hay)
long time to cook (Colenso)
cooking occupied about twenty-four hours (Tikao)
twelve to forty-eight hours (Best)
left for many days (Hay)
- 342
oven oblong, 4 to 12 feet long by 5 to 6 feet deep (Stack)
6 to 8 feet in diameter, one noted 12 feet in diameter (Rogers)
very long trenches used as steam oven (Tikao)
hole 8 feet by 4 feet by 5 to 6 feet deep (Hay)

The following points may also be singled out as being of particular interest:

on hillsides and hilltops (Rogers)
rim 9 inches to 1 foot high (Rogers)
groups of three (Rogers)
ovens of each family (Tikao)
considerable amount of tapu involved (Tikao)

The principal details of the references are compatible with the archaeological findings, and the use of water and the protracted cooking time are corroborated, but it should be noted that all the ovens examined on the Otago Peninsula, as well as those on the hills north of Waikouaiti, on Mount Royal, and on the hills south of Balclutha, are circular, whereas Stack and Hay mention oblong ovens, and Tikao a trench. Oblong pits have been noted on the Otago Peninsula, but these are thought by the present occupiers of the land on which they occur to have been saw pits.

Rogers noted groups of three, but does not state actual proximity. On the Otago Peninsula ovens occur singly and in groups. The highest concentration of ovens is nine in the 1,000 yards Grid Square S.164/ 300770. The closest proximity of any two ovens in this group is 70 m. The ground where these nine ovens occur is now open, with outcropping stones, and fallen broadleaf trees. The land was first farmed in 1860, and considerable clearing of bush occupied the settlers, but it is possible that the high ground and spurs on which the ovens are situated were already thinning and that the broadleaf trees were on the way out, some standing isolated, and many fallen. Even at lower altitudes it is possible there were open areas and that the bush was receding on all fronts into the gullies.

Assuming such a distribution of the bush, the ovens are seen to occur on relatively level ground above, and on spurs marginal to pockets of bush in the gullies. On nearby hillsides pine was felled for timber during the 1860's and although it grew mixed with other trees, it was apparently in local concentrations. Air photographs show that on those areas from which pine (Matai—Podocarpus spicatus and Rimu—Dacrydium cupressinum) was removed in commercial quantities there are neither signs of umu-ti nor of dimpling. A few broadleaf trees lie on the ground, possibly left standing after the clearing of the pine, whereas on other slopes the broadleaf lie profusely.

It is important, but difficult now, to estimate the quantity of timber removed, or otherwise gain an idea of the localization and concentration of the pine on the Otago Peninsula. It may be noted, however, that at an early date, possibly before 1860, a mile or more of wooden railway was laid across the mud flat of Hoopers Inlet (originally Cooper's) to facilitate the removal of the pine from the hillsides referred to, and transport it to Portobello for shipment. Also that the Weller Brothers, who had a whaling station at Otakau in the 1830's were unable to find timber - 343 for ship-building (a side-line industry) on the hillside behind the station, where umu-ti and dimpling are now in evidence. It is on the edge of the bush and in areas of stone outcrop that the Cordyline would grow most abundantly, and it would appear that it preferred a degenerate bush to areas where the cover included a higher proportion of podocarp and dacrydium.

The Ti as a Food

When a piece of the stalk of the ti plant is put in the ground, it grows a swollen underground tap-root which contains saccharine material. Species of the ti plant grow throughout Polynesia, and it is generally this swollen underground stem or tap-root that is spoken of as being cooked in the earth oven, and then chewed like sugar cane. In the itemized references concerning the part of the ti tree which was used, it will be noted that Best 17 is not clear, because he relies on the information given him by Tikao who first states “the tap-root was not steamed by his people as the stems were; it was covered with hot ashes and so baked,” and then later states that “the roots were sometimes steamed instead of being baked.” Another South Island Maori, quoted by Best, states that, when the Cordyline was being collected for cooking, the first root taken up was not utilized for food, but was hung up as an offering to the gods. The C australis was not the most popular species when others, such as C. terminalis and C. pulilio were available, and accounts of cooking the food do not always state which species is being spoken of. There can be no doubt, however, that the Otago species of ti was C. australis. Confusion has been caused by the use of the name ti para, which in the North Island refers to a special species of the Cordyline, but which is used by the South Island Maoris for the food product of C. australis. Hay writes as if he had been an eye witness of the stems of the young cabbage tree being stripped and laid on the stones, while John White describes large steaming pits, about eight feet in diameter, for cooking the roots. Best 18 quotes White also as informing us that the roots were wrapped in leaves and tied up so that when the cooked root was taken from the oven each family would be able to recognize its own bundle by means of peculiar knots. The possibility that the method of cooking the ti, described during the period of European contact, had changed, as a result of tribal movement, since that practised at the period when the ovens examined on the Otago Peninsula were in use, must be admitted.

The ti was cooked in most parts of New Zealand. On the Taranaki coast it was sometimes planted over large stones to prevent the tap-roots growing too deeply in the ground. In the Hawke's Bay area it was cooked in iron pots during times of scarcity. The butt of the unfolded shoots but not the ti root is still familiar to many North Island Maoris who boil it with fatty meat. It is commonly reported that early missionaries concocted some form of beer from the root of the Cabbage Tree.

In the south, the ti must have been a necessity to balance the diet, and the collection of the stems and roots and the cooking may have con- - 344 stituted a ceremony, and this is suggested by Tikao, who contributed to Elsdon Best's studies in the South Island, and by others. In view of this possibility, ethnological accounts of umu-ti and ti ceremony in Eastern Polynesia, some of which are quoted later in this article, are of interest, and it is to be noted that these descriptions so far as they concern the structure of the oven for cooking the ti, are more in accord with the archaeological findings in Otago than are the local descriptions contained in the early literature cited.

Differentiation of Umu-ti from other types of oven

In the immediate area of the nine umu-ti in Grid Square S.164/300770 no ovens of other type can be traced. The nearest ovens to the group are found between 1 and 2 km. to the north on the seaward side of Akapatiki Flat, and are shallow, conventional ovens, in sand, occurring in a layer determinable by the spread of charcoal, and in association with shell and fish remains, and some with moa remains. In another Otago Peninsula area umu-ti dug in clay and conventional ovens in sand, both relatively isolated from other evidences of occupation, are within 500 m. of each other. At higher altitudes, the umu-ti are not found associated with other types of ovens. However, shallow ovens of no definite structure are found elsewhere on the Otago Peninsula at altitudes between 200 and 600 feet above sea level. These have shown up in a number of places as a result of bulldozing trackways around the contours of the hillsides. In nearly every instance these ovens have shown up on spurs. None is more than 50 cm. deep, or more than 2m. wide. Such shallow ovens do not usually leave any signs of a raised rim, and those exposed by bulldozing would not have been detectable from the surface. The stones are small and fragmented. There is no trace of faunal remains, and a vegetarian use of the oven is therefore suggested. If the umu-ti is fundamentally a steam pit oven, then the deep pit is functionally essential to produce and retain the steam generated by the addition of water to the heated stones. It is possible that the shallow ovens were used for the process of roasting the ti root in hot ashes as described by Tikao, or for roasting the underground stems of the aruhe (Pteridium esculentum) which contains a good deal of starch and provided an important and nourishing farina in the South Island. The aruhe, according to tradition, 19 was not steamed in the umu, the hot oven, but after soaking in water was roasted over the fire. Ovens of this type have also been noted in Otago along the route of known Maori trackways.

OVEN S.164/244736

This oven, above Company Bay on the Otago Peninsula, was excavated to make clear the differentiation of the umu-ti from other types. It was revealed by bulldozer when a trackway was being widened. The plan is a circle with an outside diameter of 2 m. Within the oven was found black soil and small stones, most of them weighing less than 4 kg. A layer of charcoal extended across the bottom of the oven, - 345 beneath which the natural clay was found to have been subjected to heat. The total depth of the oven nowhere exceeded 45 cm.

This oven would appear to have served a vegetarian (possibly aruhe) cooking purpose, but in no way does it compare, on excavation, with the umu-ti, although from a surface examination there would be only the absence of a raised rim to distinguish it from the deep pit oven.

It would appear that the Otago Peninsula Maori lived on the central high ground. He would have crossed it when passing from harbour to ocean fishing grounds, and it appears he covered most parts of it in his search for two vegetable food sources, the ti and aruhe. The latter, the South Island Maori most likely dug at all times of the year, although the bracken rhizome would be richest in food value in November. On the evidence of the ovens, it would appear that the two plants did not thrive on the same hillsides.

Polynesian References

Miss Teuira Henry of Honolulu described in this journal in 1893 the Raiatean ceremony of the Umu-ti. 20 In the Society Islands, she says: “the ti ovens are frequently 30 feet in diameter, and the large stones heaped upon small logs of wood, take about twenty-four hours to get properly heated. Then they are flattened down, by means of long green poles, and the trunks of a few banana trees are stripped up and strewn over them to cause steam. The ti roots are then thrown in whole, accompanied by short pieces of ape root (Arum costatum) that are not quite so thick as the ti, but grow to the length of 6 feet or more. The oven is then covered over with large leaves and soil, and left for about three days, when the ti and ape are taken out well cooked and of a rich, light brown colour. The ape prevents the ti from getting too dry in the oven.” The ti root is described as being 2 feet long and from six to ten inches in diameter, and as requiring to be well baked before eating, when it has “something of the texture of sugar-cane and its thick juice is very sweet and nourishing.” A picture of the ceremony is included with Miss Henry's article which suggests an umu-ti of dimensions closer to the Otago examples than those she states.

An informative ethnographical description is given by Young 21 from which the following is quoted: “The ceremony of the umu-ti, oven or fire pit of the ti plant is said to have been first practised in the Eastern Pacific on the Island of Huahine at the marae of Fare-ti. It was held to be essential that the sun, soon after appearing above the eastern horizon, should shine on the fire pit which was situated close to the marae. The pit was generally circular from three to four fathoms (18 to 24 feet) in diameter and 4 to 6 feet in depth. Occasionally the form of the pit was quadrilateral but the area of the heated stones was always circular. A fire of logs inter-mixed with smaller timber was lighted in the pit at sunset on the day before the ceremony, and kept burning all night, a number of large, hard, volcanic stones being laid on top of the wood, which stones of course became heated as the wood was consumed. The fire was - 346 replenished with wood during the night until the floor of the pit was a bed of live coals among which lay the heated stones.”

Describing the vegetable foods of Samoa, Buck 22 says that the underground stem of the ti was cooked for the saccharine material it contained. “A very large oven (umu-ti) is prepared communally by the village (nu'u). The families dig up the stems and bring them in baskets to the common oven. The oven is dug out to about ten feet in diameter. Larger stones than usual are heated on a fire made of logs. The levelling of the stones takes some trouble. Six or more men dressed in ti leaf kilts and wreaths do the work with long poles (sosofa). Sometimes a heavy log tied with ropes is dragged from side to side. The levelled stones are covered with the leaves of the ti. The underground stems, still in baskets, are placed in the oven in the parts assigned to the different families. They are next covered over with more leaves and buried with earth. The oven is left covered for a week or even longer. When it is opened the families take their own baskets. The outer bark is removed. The cooked ti can be chewed direct.” Buck says that the Samoan method of cooking was normally by dry heat. 23 Further, that in the ordinary cooking oven such as would be used for the cooking of a pig with vegetables, earth was not used to cover, but a leaf cover or, later, sacking was used. In such an oven water was not used to raise steam. The Cook Islands method resembles the Samoan in not using water. The umu-ti, however, was different in that earth was used to cover up the whole oven in which the stems of the ti were cooked. Buck does not state whether or not water was used in the Samoan umu-ti.

A large oven on Pitcairn Island was examined by the writer. This oven is on the west side of the island between three and four hundred feet above the sea, and just below the supposed site of a marae. The oven is 2.3 m. wide and more than 1 m. deep. It was found to contain a deposit of stones of vesicular basalt and to have a thick layer of charcoal at its base. It is flat at the bottom, steep sided, and the walls are reddened by heat. The oven appears to have been deliberately filled after removal of the larger oven stones, and levelled with the surface so that no raised rim remains. A ti plant, a variety of Cordyline, still grows on Pitcairn and was formerly more abundant.

  • ANDERSON, J. C., 1907. Maori Life in Ao-tea. Wellington, Whitcombe and Tombs.
  • — — 1916. Jubilee History of South Canterbury. Auckland, Whitcombe and Tombs.
  • BEST, Elsdon, 1924. The Maori. Published by The Board of Maori Ethnological Research for the Author and on behalf of the Polynesian Society, Wellington.
  • — — 1925. Maori Agriculture, Dominion Museum Bulletin No. 7, Wellington: Dominion Museum.
  • — — 1931. “Maori Agriculture.” Journal of the Polynesian Society 40:1-22.
- 347
  • BREES, S. C., 1847. Pictorial Illustrations of New Zealand. London.
  • BUCK, P. H., 1930. Samoan Material Culture. Bernice P. Bishop Museum Bulletin No. 75. Honolulu.
  • COLENSO, W., 1880. “On the Vegetable Foods of the Ancient New Zealanders,” Transactions and Proceedings of the New Zealand Institute, XIII:3-38.
  • HAY, James, 1915. Reminiscences of Earliest Canterbury. Christchurch, Simpson and Williams.
  • HENRY, Teuira, 1893. “The Umu-ti, a Raiatean Ceremony.” Journal of the Polynesian Society, 2:105-108.
  • HOCHSTETTER, Ferdinan von, 1867. New Zealand, Its Natural History. Stuttgart: Cotta.
  • ROGERS, L. S., 1922. “Notes on the occurence of Umu in the Warepa Survey District, Otago.” Journal of the Polynesian Society, 31:155.
  • SHORTLAND, E., 1851. The Southern District of New Zealand. London.
  • TEVIOTDALE, David, 1939. “Excavation of Maori Implements at Tarewai Point—Otago Heads.” Journal of the Polynesian Society, 48:108-115.
  • YOUNG, J. L., 1925. “The Umu-ti ceremonial fire walking, as practiced in the Eastern Pacific.” Journal of the Polynesian Society, 34:214-222.
1   Anderson 1907:512.
2   Best 1925:139.
3   Brees 1847: Plan.
4   Teviotdale 1939:108.
5   Shortland 1851:234.
6   Hochstetter 1867:157.
7   Best 1925:142.
8   Best 1925:142.
9   Best 1924:394.
10   Best 1931:13.
11   Best 1931:14.
12   Best 1925:142.
13   Hay 1915:15.
14   Colenso 1880:28.
15   Rogers 1922:155.
16   Anderson 1916:40, 554.
17   Best 1931:17.
18   Best 1931:15.
19   Anderson 1907:5.
20   Henry 1893:105.
21   Young 1925:214.
22   Buck 1930:136.
23   Buck 1930:103.