Other formats

    TEI XML file   ePub eBook file  


    mail icontwitter iconBlogspot iconrss icon

Arts and Crafts of the Cook Islands




Except for purposes of modesty, there was little need for clothing in the tropics. Men wore a loin cloth (maro) and women a short skirt or kilt (pareu). A kind of poncho (tiputa) was made of bark cloth or sometimes of plaited pandanus leaves. For dances and festivals, kilts were made of ti leaves, coconut leaflets, or hibiscus bast. Chiefs sometimes wore a wide plaited band around the waist as part of their chiefly regalia. Headdresses of coir fiber and feathers were worn by chiefs and warriors on state occasions. Various types of body ornaments were worn in the different islands. Sandals protected the feet from the sharp coral of the shore reef and the inland upraised coral reefs of some of the islands.

In volcanic islands where the paper mulberry flourished, bark cloth was the general material for garments. Native materials have been superseded by |trade textiles, except for native dance costumes. Sandals, however, are still used on the reef.

page 64

Bark-cloth Garments

Loin Cloth

A loin cloth or perineal band (maro), which was worn by men, was a long narrow strip of bark cloth. It was the only garment worn while working. The wearer held one end on the breast or under the chin, while he passed the cloth back between his thighs to cover the genitals, passed it around the waist from the back above the right hip, looped it at the back through the right turn and thus reversed the direction from left to right in front. The end was tucked away between the folds and the skin and the held end was dropped to hang down in front. The front end was termed taumua (tau, to hang; mua, in front).

Cook (20, p. 171), the European discoverer of Mangaia (1777), says of the people he saw on the shore: "Most of them appeared naked except having a sort of girdle, which, being brought up between the thighs, covered that part of the body." Later, he says of two men who came out in a canoe:

They wore such girdles as we had perceived about those on shore, and we found that they were a substance made from the Morus papyrifera, in the same manner as the other islands of the ocean. It was glazed like the sort used by the natives of the Friendly Islands. …

In speaking of the natives of Atiu, Cook (20, p. 183) says, "Like those of Mangeea [Mangaia], they had girdles of glazed cloth, or fine matting, the ends of which, being brought betwixt their thighs, covered the adjoining parts." The glazed appearance was produced by mixing coconut cream or coconut oil with the vegetable dye.

Kilts and Skirts

More a kilt than a skirt, the pareu was worn by females. It was formed of a strip of bark cloth wide enough to reach from the waist to the depth required, above the knees for unmarried girls and below for married women. The cloth was wrapped tightly around the waist and the overlapping ends were tucked inside the folds to keep the garment in position without a belt.

Cook (20, p. 189), in speaking of a group of female dancers in Atiu, said, "… their dress consisted only of a piece of glazed cloth fastened about the waist, and scarcely reaching as low as the knees.…"


An upper body garment consisting of a length of bark cloth with a hole in the middle was called a tiputa. The head was passed through the hole, and the cloth hung down in front and behind, leaving the arms free and the sides uncovered. The word puta in the name tiputa means a hole. Cook (20, page 65p. 171), speaking of the clothing of the Mangaians, said: "But some of them had pieces of cloth of different colors, white, striped, or chequered, which they wore as a garment thrown about their shoulders." These garments were probably tiputa, for my Mangaian informants had no knowledge of any other form of body garment.

Bloxam (8, p. 86), who visited Mauke in 1825, says: "Several of the inhabitants wear their dress in the manner of the Spanish poncho and these were painted in very gay colors." On the following page, he states: "They wear their tapas in a variety of ways—one had it similar to the American poncho, which being spotted and starred over with red and black gave it a very peculiar appearance."

Even after European cloth came into general use, the people continued to make tiputa out of bark cloth for their native festivals. Many have found their way into museums, and it is somewhat difficult, owing to the uncertainty of their date of manufacture, to determine what was the original form and what innovations were added in post-European times.

Two specimens from Mangaia in the Bishop Museum have circular neck holes and long side fringes. Similar garments in the British Museum were made in the same way, but one British Museum garment, figured by Edge-Partington (24, II-ll-l), has a longitudinal slit for the neck and has no side fringes (fig. 266, b). It was evidently obtained through the missionary, Gill, for Edge-Partington quotes the following note from him, "This is the tartan of a clan. No other than a member can wear it under pain of death. W. W. Gill." The cloth is covered with a lozenge pattern in black (fig. 28, e), and both the pattern of the garment and the colored design appear old.

The two Mangaian tiputa in Bishop Museum are decorated with perforations in the cloth. They take the form of small triangles in horizontal rows with the bases on the same line (fig. 27, a) or with alternate triangles inverted (fig. 27, b). Longer triangles with narrow bases are arranged in groups of three which may alternate as to the position of the bases (fig. 27, d). Lozenge-shaped perforations were formed by pinching up the cloth and cutting out V-shaped pieces (fig. 27, c). Another motif of larger size has serrated edges and pointed ends (fig. 27, e). Perforations, usually triangular, were made in three concentric circles around the neck opening, and the different motifs were arranged in horizontal or vertical rows over the body of the garment. According to my native informant, the perforated technique was peculiar to Mangaia and hence is of value in determining locality. (See plate 6, B.) The cutting out of triangles and lozenges in cloth seems to be a natural sequence to the common Mangaian technique of cutting similar figures in their wood carving.

A Mangaian tiputa (fig. 266, d) in the British Museum, figured by Edge-Partington (24, II-11-2), is stated by him to be "Mourners dress of the better page 66class of black colored bark cloth pierced with diamond shaped holes in parallel rows." The information seems to come from Gill whose long residence in Mangaia, when information was firsthand, entitles his statements to respect. It may be that before the culture was affected by foreign influence, the people wore the painted tiputa without perforations as their ordinary dress and the dark colored ones with perforations as a sign of mourning. Later, as native tapus lost their dread, the mourners' costumes were given circular neck holes and fringed at the edges to be used at dances and festivals. The dark cloth was easy to stain and did not require the painstaking task of covering the cloth with an intricate painted design.

Figure 27.—Mangaian perforation motifs: a, triangles on same base line; b, alternate triangles inverted; c, lozenge perforations in rows; d, larger triangles in group of three resembling tattoo motif (fig. 71, a) said to symbolize the three original tribes of Mangaia; e, serrated motifs.

Figure 27.—Mangaian perforation motifs: a, triangles on same base line; b, alternate triangles inverted; c, lozenge perforations in rows; d, larger triangles in group of three resembling tattoo motif (fig. 71, a) said to symbolize the three original tribes of Mangaia; e, serrated motifs.

In making one of the later Mangaian garments (fig. 266, f), which is now in Bishop Museum (C8235), a piece of cloth 48 inches wide and 90.5 inches long was used. A circular piece about 6.5 inches in diameter was removed from the middle and the edge given a serrated form by cutting out small triangular pieces. The sides were fringed by horizontal cuts 6.5 inches deep and 0.5 inch apart. The two ends were treated with vertical cuts 11 inches long and half an inch apart. A rectangular piece, 6.5 inches wide by 11 inches deep, was removed from each corner, owing to the meeting of the fringe cuts from the sides and ends. When the cloth was worn, the side fringes hung down and reduced the actual width of the garment from 48 inches to 35 inches. The depth including the lower fringes remained half the total length of the cloth, 45.25 inches. The cloth was of the dark color obtained by immersion in a taro swamp, but the outer surface was rubbed with a reddish brown dye. Of the two perforated garments in the British Museum, the sides and bottom of one are serrated (fig. 266, d), whereas the other (fig. 266, e) has serrated sides and bottom fringes.

The other Bishop Museum specimen (8059) from Mangaia is made of yellow cloth with splotches of red and thus conforms to the type cloth termed 'upa'a. It has a serrated neck, side, and bottom fringes and is perforated with triangles, lozenges, and the serrated motif (fig. 27, e). Thus, even if the perforated technique was originally confined to the drab mourners' costumes, it has been used in later times on festival attire.

page 67

A tiputa from Atiu (fig. 266, c) in Bishop Museum (C2891) is made of breadfruit bast. The circular neck hole is serrated, but the side and lower edges are not fringed. No dyes were used and the natural color is grayish brown. The ridges made by the beater stand out in parallel lines.

For the distribution of the poncho, see pp. 432-434.


A cloak or wrapper of a rectangular sheet of tapa was often thrown over the shoulders to protect the body in the cool of the evening. In Tahiti, such cloaks were called tihei, the name consisting of the prefix ti and hei (to be around the neck). Savage (57) does not record this name for the Cook Islands, so it is probable that such wrappers, only worn occasionally, did not receive the dignity of a special name.


A belt or girdle received the name of tatua and a strip of bark cloth was sometimes used as part of the full dress. Plaited girdles of pandanus leaf however, were more appropriate in a chief's dress (pl. 6, C).


Bark cloth was used in the making of headdresses. (See p. 79.)

Bark Cloth

The best bark cloth was made from the bast of the paper mulberry, which was cultivated for that purpose. The general name for the plant is aute or anga, but Aitutaki people held that aute and anga were distinct varieties, and they added a third name purautea. It is possible that purautea (purau, Tahitian name for Hibiscus tiliaceus; tea, white) really refers to the wild hibiscus, for Savage (57) states that purautea was a soft quality of ancient garment made from the lemon hibiscus.

A good cloth was also made from the bast of shoots and younger branches of the breadfruit tree (kuru). A coarser, brown cloth was made from the aerial rootlets of the banyan (aoa). Gill (32, p. 337) states that the poor people of Mangaia manufactured their clothing from the bark of the "Eutada scandens" (evidently a misprint for Entada scandens which has since Gill's time been changed to Entada phaseoloides (Linnaeus) Merrill). The plant is a woody climber which is common in the forests.

The manufacture of bark cloth was the work of women, but in Mangaia a special thick white cloth was made by men.

page 68

A shell (kafi), consisting of one valve of a marine bivalve (Asaphis violacea), was used to separate the inner bast (kiko) from the outer bark (pakiri). The natural edge was used; shells were not ground as were the shell implements of Samoa.

Beaters (ike) were made from ironwood and sometimes from miro. Twelve beaters examined in Atiu ranged in length from 14.25 to 18 inches, with an average length of 15.5 inches. The wood was cut square in section with four equal surfaces ranging from 1.5 to 2 inches wide, with an average of 1.75 inches. The proximal third was gradually rounded to form a grip with the proximal end flared out to the original square section. The distal two thirds formed the beating part of the implement and was incised with grooves (ngao) to form ridges that would assist in beating out the bark. The grooves, which were cut parallel to each other and to the edges of the surface, varied in closeness according to their function in the various stages of beating the bark—widely spaced grooves for the initial stage of beating, medium spaced grooves for intermediate stage, and closely set grooves for the final stage. In Aitutaki (70, p. 78) the beaters are named according to the spacing of the grooves, and I counted the ridges per centimeter for the four varieties named.

  • ngao 'oa'oa: 1 ridge to 1 cm.
  • ngao papa: 3 ridges to 1 cm.
  • ngao I'ei'e: 4-5 ridges to 1 cm.
  • ngao ta'akaieie: 6.5 ridges to 1 cm.

In Atiu a similar system was followed, but the qualifying terms were different.

  • ngao tuanunui: coarse grooves for first stage (papa'anga)
  • ngao kikau: medium spacing
  • ngao makeke: closer spacing;to smooth ('akamnnia)
  • ngao 'unga'unga: fine, close grooves for finishing

In Mangaia, the beaters were named according to their function, which corresponded with the number of grooves.

  • ike papa: for initial beating; with coarser and more widely spaced grooves
  • ike toto: for intermediate stage; with medium-sized grooves set closer together
  • ike 'akatu: for finishing work; with fine grooves set closely together

Each surface of the beaters was carved in one pattern but they differed in the four surfaces so that one beater had coarse, medium, and fine surfaces for the different stages of beating (pl. 6, A). The fourth surface gave an extra medium or fine surface. The width of the blade was usually the same throughout, but sometimes the proximal end was slightly wider. The carved pattern was always in parallel lines.

page 69

The beaters from all the islands had the handles shaped with a proximal flare, except the fine beaters (ike 'akatu) of Mangaia in which the grooved surfaces extended from end to end of the implement. The Mangaians held that the beaters for the heavier stages of beating were shaped to give the hand a good grip, but, as the finishing process consisted of light tapping, it was not necessary to shape the handle.

For types of beaters, see pp. 429-431.

Each household had several beaters, and on the death of the mistress of a family the beaters were inherited by her daughters. In Atiu, I saw a beater in a burial cave, showing that sometimes these articles were placed with the owner after death.

Wooden anvils (tutunga; from tutu, to beat) were made of tamanu or miro logs that were squared. The length was usually about 13 feet and the average width and depth about 6 inches or a little more, but usually the end corresponding to the upper end of the tree trunk was less than the butt end. The under surface was hollowed out with a deep groove, and the ends were usually raised a little above the ground by the placing of stones or old mats beneath them. Beating the bark produced a resonant sound which the workers enjoyed. With a long anvil, a number of women could work side by side. They had a number of beats and chants to which they all kept time, thus getting fun and enjoyment from their labor. Certain signals were used that had a definite meaning, for instance, the entrance of a visitor at one end of the village was known at the other end long before he arrived there.

A Rarotongan legend states that the goddess Ta'akura had an anvil named Tangi-te-varovaro (making-a-prolonged-hollow-sound). One end had a secret recess in which was placed a kura (red feathers, something precious) and a rei (whale ivory ornament). During the absence of Ta'akura from her home, the anvil was taken by a woman named Tapairu-'aere-moana. On the return of Ta'akura the sound of beating on her anvil was borne to her on the south wind (tonga). She followed the sound until she reached the house of the thief, who was beating bark on the stolen anvil. Ta'akura demanded her anvil, but the other woman claimed it as her own. Ta'akura asked her if she had any special mark to denote her ownership. When the thief replied in the negative, Ta'akura opened the cavity at the end and produced the kura and the rei. The following verse records the incident:

Tonga e, matangi e! O South wind, O breeze!
I taku tutunga nei This my anvil was
'Omai e te tonga e. Returned to me by the south wind O.
Vero'ia mai taku tutunga Directed toward my anvil
E te matangi, e te tonga e. By the breeze, by the south wind O.

Wooden anvils were used throughout the volcanic islands of Polynesia except the Marquesas and Easter Island where flat stones took the place of wood.

page 70
Manufacture of Cloth

The general term for beating out the bark is tutu, but other terms are applied to the various stages of the process. The manufacture of cloth has been described for Aitutaki (70, pp. 79,80). Throughout the group, the process is the same, though there are slight variations which do not affect the main principles.

Paper mulberry saplings about as thick as the thumb are cut down in the plantations and the bark peeled off there or at home. The bark is cut longitudinally at the severed end of the thicker butt end and pulled away with the right hand while the left hand holds the peeled stick. The bark splits naturally down the length of the sapling and comes off in one long strip. As the bark tends to curl into tubular form, it is folded in short lengths inside out to flatten out the curvature. The inner bark or bast (kiko) is separated from the outer bark (pakiri) at the butt end with a ka'i shell and then carefully pulled away throughout the length of the strip.

The outer bark is discarded and the bast strips tied in bundles for future treatment. Usually the bundles are soaked in water for 24 hours before the first beating takes place. In Aitutaki they were soaked in sea water, but where there are streams they may be soaked in fresh water.

The first beating is done on the anvil with the coarse surface of the beaters. In Aitutaki this process is termed 'oa'anga, hence the coarse surface of the beater is termed ngao 'oa'oa. The first beating, applied to each individual strip, flattens it out and brings out the texture of the bark. The strips are then washed in fresh water to remove the sap, salt, and any green coloring that may have adhered from the outer bark. In Samoa, the green coloring is scraped off with a shell. The washed material is placed in a rough coconut-leaf basket (tapora) and allowed to drain. The next day the bast is wrapped in banana leaves or, in Aitutaki, in coarse taro leaves (puraka) for three days. This treatment is a form of retting. Some women said that they could tell by tasting the bast when the material was ready for further beating.

The second or final beating was commenced with the medium grooved surfaces. When a number of women were working at one anvil, each put several strips before her and beat out a section to the required thinness. The edges of the sections were overlapped and, by further beating, were felted together into one continuous sheet. The process of felting the material is the characteristic technique of central, eastern (except Easter Island), and northern Polynesia, whereas in the west, in Samoa and Tonga, the thin sheets from individual strips of bast were pasted together to the required thickness. In Aitutaki, the process of beating into a sheet is termed papa'anga. Beaters are usually held with both hands.

The finishing process ('akaotinga) is done by lightly tapping with the finely grooved surfaces of the beaters. The smoothed cloth is then spread out in the page 71sun to dry, the edges weighted down with basaltic stones to prevent it from blowing about.

In the felting process, the parallel marks of the ridges on the beaters show up as watermarks when the cloth is held up to the light. In the pasting process of western Polynesia, the coarse markings of the beaters do not show on the cloth. Hawaiian bark cloth is readily identified by the watermarks made by the various patterns on the beaters.

The completed cloth is white and may be used without further treatment. White cloth in Mangaia is termed autea and in Aitutaki, pa'oa. Savage (57) gives tapa as a general name for cloth, so tapa was probably used in Rarotonga. When colored with vegetable dyes, specific names were given to the different kinds of cloth.


The dyes used were prepared from the bark, roots, sap, and leaves of various plants to produce the colors, yellow, red, and black. Black was also obtained by soaking the cloth in the mud of taro swamps. Combinations of plants were also used, and coconut oil.

The addition of color to the white cloth was done by total immersion of the cloth in the dye solution, rubbing the dye over the outer surface of the cloth, or by freehand painting. The simplest form of immersion was by pressing the cloth down in the mud of taro swamps, leaving it there until the required depth of color was obtained, washing the mud off, and then drying it in the sun. A black color was obtained which in time fades to a brown. Red and yellow were obtained by soaking the cloth in the dye solution in wooden bowls made on the pattern of food bowls but set aside for dyeing. For large pieces of cloth the canoe-shaped bowl (paroe) was used. In this process the same color appears evenly on both surfaces.

To dye one surface in the same color, a small piece of cloth was dipped in the dye and rubbed evenly over the surface. The other surface remained white except for splotches where an excess of dye soaked through. The favorite cloth used for the tiputa in Mangaia was rubbed with a yellowish dye on the outer surface after the material had been immersed in a taro swamp. The inner surface remained brown, whereas the outer surface had a ribbed appearance because the dye took better on the parallel ridges made by the beater.

A yellow color was obtained from turmeric (renga), nono, and in Aitutaki from the kavapiu (?). The turmeric roots were washed and grated on a rough stone or rough coral. The roots of the nono were scraped with a ka'i shell to remove the outer bark; the inner bark scrapings produced the dye. The roots of the kavapiu were treated like turmeric. The scrapings of the root used were thoroughly mixed and worked with the hands in water in a wooden bowl. The insoluble material, caught in a strainer made of dry strips of hibiscus bast, was wrung out, and flicked off as in preparing the kava beverage. Some women page 72added a little sea water, and in Mangaia, coconut cream was added to grated turmeric.

Red was obtained from the berries of the mati, combined with leaves of the tou in Aitutaki and with the leaves of the morirei in Mangaia. The ripe berries with their ends pinched off, were placed in water in a wooden bowl and removed when the water became cloudy or gray. The fluid was termed vai mati. Crushed leaves of the tou or morirei turned the fluid red. The leaves were removed when the right shade of red was reached. In Aitutaki a little coconut cream was added.

A red color is also obtained from the bark of the candlenut (tuitui), which is scraped and wrung out in a wringer (taka) of plaited hibiscus bast. The fluid (vavai'iri) was stored in a coconut shell container until required. In Mangaia, the fluid obtained from beating the bast of the toa was mixed with vavai'iri to produce a reddish color; flowers of the hibiscus (kaute) were pounded and rubbed on the cloth to produce red splotches. The process was done at sunset and the cloth dried at night to avoid fading.

The commonest method of obtaining black was to immerse the cloth in the black mud of taro swamps and leave it there until it had assumed the right depth of color. The cloth was then washed and dried. In Mangaia, such cloth was termed pakoko. A black color was obtained in Aitutaki by soaking the cloth in candlenut juice (vai'iri), drying the cloth in the sun, and then soaking' it in ironwood juice. After sun drying, the cloth was cooked in an earth oven lined with a layer of banana stem bark and with a layer of candlenut leaves under the cloth. The whole was covered with leaves. Tasting determined whether the cloth was cooked (maoa), and it was removed and dried in the sun Next, the cloth was pressed down in the mud of a taro swamp (repo) and left until it acquired the desired shade of black. Finally, it was washed and dried.

In Mangaia, the pounded leaves of the 'ange shrub were mixed with coconut cream and wrung out. The cloth soaked in the mixture and sun dried had a gray color and a sweet scent.

The information about the above colors was obtained in Aitutaki and Mangaia, but it is probable that additional mixtures were used in others of the Cook Islands.

Paper-Mulberry Cloth

In Mangaia, the completed white cloth was termed autea or tikoru. When treated further with dyes or coconut cream, the cloth received the general name of parai (to smear or paint). The various parai were as follows:

1.pa'oa-tea: a yellow cloth soaked in a mixture of turmeric and coconut cream.
2.'upa'a: the pa'oa-tea colored with red splotches of pounded red hibiscus flowers; dried out at night to prevent fading by the sun.
3.tukntoa: a red cloth obtained by immersion in a mixture of candlenut and ironwood (toa).
4.tuku'anga: a gray cloth from immersion in a mixture of 'ange leaves and coconut cream.
page 73

The cloth soaked in taro swamp mud received the general name of pakoko and further treatment also made it a parai. of which one kind was known.

Parai mangu (mangu, black): the pakoko cloth was rubbed on the outer side with a mixture of grated turmeric and grated coconut, which gave a yellowish-black polish to the surface treated.

In Mangaia, a special thick cloth called tikoru mata'iapo was made as a covering for the gods which was never removed except during temple ritual or when the keeper of the gods replaced it with a new one. The white cloth was also used as clothing by high chiefs and priests, hence its name, which means cloth of the first-born. The manufacture of the cloth was a secret process known only to the priests of the Tonga'iti tribe. Its association with the gods saved the Tonga'iti tribe from extermination when a plot to overthrow the ruling tribe was discovered. They were exiled instead, and Gill (33, p. 132) gives the following reason:

The alleged motive for not permitting these Tongans to be slain was, "that the sacred clothing of the gods might not be defied with human blood", and so draw down vengeance upon the ruling race. Pati, priest of the exiled tribe was the sole depository or maker of the superior paper-mulberry cloth as thick as cardboard, used exclusively for clothing the gods, great chiefs, and priests. This "lordly clothing" was actually worshipped by the tribe that manufactured it, under the name of te tikoru mataiapo.

In Aitutaki (70, pp. 80-83), the white cloth (pa'oa) was, according to my informants, treated in seven ways.

1.'apa'a (hapaha): a yellow cloth by immersion in any of the yellow dyes, but turmeric was the best dye. This was the appropriate cloth to spread over a mother after childbirth.
2.pungavere: a red cloth from immersion in the mixture of mati berries, tou leaves, and coconut cream.
3.rarua: a white cloth (pa'oa) with a pattern of lines made with the red dye used in the pungavere cloth.
4.pa'oa verevere ki te repo: a black cloth produced by the method described in the second process under black dyes (p. 72).
5.piri: a black cloth treated for one day in the mud of a taro swamp and after washing and drying, rubbed with a mixture of iron-wood sap (toto) and coconut oil. The cloth after drying in the sun has a varnished appearance due to the oil and is less affected by wet than cloth not so treated. The cloth has a sticky feeling and hence is called piri (to stick)
6.pareu: a reddish cloth obtained by soaking in candle-nut fluid and with straight, freehand lines drawn in the black of taro swamp mud with a dry pandanus fruit key (kati 'ara) used as a brush. Used as a bed cover.
7.inaina: white cloth coated with coconut oil ('inu 'akari) which gave it a shiny appearance (karaparapa).

The other islands had their own names for the various types of cloth.

Breadfruit Cloth

Cloth made from the bast of shoots or young branches of the breadfruit is grayish-brown in color and is of a coarser texture than paper-mulberry cloth. page 74 Bishop Museum has a good sample of this cloth collected in Rarotonga (no. 8269). When loom woven textiles displaced bark cloth as clothing, the people stopped growing the paper mulberry. When those who knew how to make cloth wanted some for native festivals they had to use the bark of the breadfruit tree as raw material. In 1929, when the process of manufacture was demonstrated to me on Atiu, breadfruit bast was used. The grooves made by the beater show up well. In Aitutaki, the cloth was called inaina. It is curious that though I obtained the name pokuru in Mangaia for breadfruit cloth, Gill (32, p. 338) states that breadfruit bast was not used for cloth in Mangaia.

Banyan Cloth

Cloth made from the bark of the aerial rootlets of the banyan (aoa) was brown. In Mangaia it was called aoa after the tree. Gill (32, p. 338) says that in Mangaia it is worn by the poorer people.


Designs in yellow, red, or black were painted on plain white cloth or cloth that had been stained to one color. Small brushes were made from the dry keys of pandanus fruit. Probably other forms of brushes were also used.

Dyeing frames (rakau takiri pa'oa) were said by native informants to have been used in Aitutaki (70, p. 78). Dried coconut-leaflet midribs were tied in parallel lines with fine sennit to a rectangular frame made of strips of pandanus aerial rootlets. A pattern made with the midribs crossing at right angles with the frame was termed ngaito, and a set crossing obliquely was termed okaoka. The dye was dabbed onto the midrib lines and the frame then pressed down on the cloth. This makes a fourth process that may be termed stamping.

The western process of rubbing and the mistakes that have arisen in identifying specimens as Cook Islands are discussed on pages 431-432.

The painted patterns on a few old specimens of Cook Islands tapa are recorded here to aid in future identification.

One specimen collected in 1825 on the island of Mauke by Andrew Bloxam, chaplain on H.M.S. Blonde , is authentic, for he bound a piece of it in his diary which is now in the possession of Bishop Museum. As the Blonde did not touch at Tahiti, nor any other island in the Cook group, the locality of Mauke is absolutely correct. (See figure 28, a.)

Old pieces of Rarotongan tapa are found wrapped in a thick roll about the middle section of some of the staff gods that are peculiar to Rarotonga. The pattern on the tapa bundle around two of these gods is shown in figure 28, b, c. A similar pattern occurs on a piece of cloth attached to a large feather headdress in the British Museum said to belong to a Tahitian mourner's costume. The identification of the headdress as Tahitian is open to several objections. The few complete Tahitian mourner's costumes that have been saved in museums have a headdress composed of a coil of colored cord worn like a turban. Furthermore, the wooden framework which carries the feathers is attached page 75to a stiff cap of coiled sennit that follows the technique used in the Cook Islands. The importance of authentic old tapa patterns as a means of identifying doubtful artifacts is thus obvious. The coiled cap indicates that the head-dress must belong to the Cook Islands but the bark cloth pattern further identifies it as being Rarotongan. (See figure 28, d.)

Figure 28.—Painted tapa patterns. a, Mauke: white cloth painted with black bars, black triangles in pairs with apices meeting giving hour-glass appearance and internally blaked with thin parallel lines, and zigzag bars formed by crossing black lines; horizontal bars (1) and triangles (1) painted yellow; rectangles (2) and triangles (2) in red (Bishop Mus., Bloxam ms.). b, shows carved heads (1) and tapa bundle with zigzag pattern (2) at one end and double rows of black lozenges separated by two thin lins (3) at other end (British Mus., L.M.S. coll.). c, carved god (1) with bundle of white tapa (2) painted with wide bars with serrated edges (W. O. Oldman coll., (424) d, coiled coir headdress (1) with tapa cloth (2) painted with wide black bars with serrated edges similar to c (British Mus., L.M.S. coll.). e, tiputa garment from Mangaia with lozenge pattern in black as drawn by Edge-Partington (British Mus.).

Figure 28.—Painted tapa patterns. a, Mauke: white cloth painted with black bars, black triangles in pairs with apices meeting giving hour-glass appearance and internally blaked with thin parallel lines, and zigzag bars formed by crossing black lines; horizontal bars (1) and triangles (1) painted yellow; rectangles (2) and triangles (2) in red (Bishop Mus., Bloxam ms.). b, shows carved heads (1) and tapa bundle with zigzag pattern (2) at one end and double rows of black lozenges separated by two thin lins (3) at other end (British Mus., L.M.S. coll.). c, carved god (1) with bundle of white tapa (2) painted with wide bars with serrated edges (W. O. Oldman coll., (424) d, coiled coir headdress (1) with tapa cloth (2) painted with wide black bars with serrated edges similar to c (British Mus., L.M.S. coll.). e, tiputa garment from Mangaia with lozenge pattern in black as drawn by Edge-Partington (British Mus.).

An old Mangaian pattern has been recorded by Edge-Partington (24, II-11-1) from a poncho in the British Museum (fig. 28, e); and he quotes Gill as saying that different tribes had their own designs which only members page 76of the tribe could wear. Gill (33, p. 334) says: "Even in the matter of clothing there were special differences. I have seen a man stripped naked for presuming to wear the garments of another tribe. The meek defence was that his grandmother was a member of the said tribe."

Plaited and Leaf Garments

Loin Cloth

Though bark cloth was the general material on volcanic islands, an observation by Cook (20, p. 194) shows that in Atiu, at least, loin cloths were perhaps made of matting, presumably of pandanus leaf. He writes: "Their general dress was a piece of cloth, or mat, wrapped about the waist, and covering the parts which modesty conceals." It is possible, however, that the term "mats" may have referred to a form of kilt.

In the atolls, called the Hervey Islands by Cook (Manuae and Te Au-o-tu) the material was of necessity pandanus. Cook (20, p. 210) says of these islanders who came out in double canoes: "Their only covering was a narrow piece of mat, wrapt several times round the lower part of the body, and passed between the thighs; …"

Kilts and Skirts

Kilts of ti leaf (pareu rau ti) were made by attaching the leaves to one side of a three-ply braid formed of the butt ends of the leaves from which the stiff midribs were removed (fig. 29, a), and sometimes partly dried leaves were added for their golden color. The technique has been described (70 p. 86). They were worn by both sexes in dances and festivals, and they were; not worn everyday as they were in Samoa.

Figure 29.—Kilts, a, ti leaf kilt: three-ply braid commenced with butt ends: of three leaves (1), and fresh leaf added each time lowest ply crosses over to middle position; finishing end (2) is braided on for a few turns and knotted; strips of bast are attached to each end for tying around waist. b, hibiscus bast kilt: made of strip of bast doubled over two cords (1) and fine cord (2) doubled around left side strip a little below supporting cords and run across in single-pair twine. c, method of doubling bast strips over two cords as in b. d, method of looping bast strips over a single cord alternate method.

Figure 29.—Kilts, a, ti leaf kilt: three-ply braid commenced with butt ends: of three leaves (1), and fresh leaf added each time lowest ply crosses over to middle position; finishing end (2) is braided on for a few turns and knotted; strips of bast are attached to each end for tying around waist. b, hibiscus bast kilt: made of strip of bast doubled over two cords (1) and fine cord (2) doubled around left side strip a little below supporting cords and run across in single-pair twine. c, method of doubling bast strips over two cords as in b. d, method of looping bast strips over a single cord alternate method.

page 77

Temporary kilts of coconut leaflets (pareu nikau) were made with a technique similar to that for ti-leaf kilts. The leaflet midribs were removed.

Kilts of hibiscus bast (pareu kiri 'au) were made of strips of bast attached to a single or a double cord which served as a waist cord. One or two rows of spaced single-pair twine (70, pp. 87-90) helped to keep the strips together (fig. 29, b). These were also used in dances but, as the material was durable, they were kept for future use. Ornaments (rakei) of colored seeds and shells were sometimes used to decorate the kilts. In modern times, dance kilts are made from the stout fibers of a low shrub named purumi.


I have seen tiputa made of plaited pandanus for dances and festivals. That their use was not modern is vouched for by Cook (20, p. 194) who saw some in Aitu. "But some had pieces of mats, most curiously varied with black and white, made into a sort of jacket without sleeves. …"

A rough tiputa for wear in the cultivations was made from a banana leaf. A section of midrib was removed from the middle of the leaf to admit a person's head, and the midrib cut ends were tied with hibiscus bark. The midrib was cut through on the middle side of each tie and the section removed by cutting the leaf on each side of the section. The head was passed through the hole and the rough garment protected the front and back from sunburn.

Belts or Girdles

A belt or girdle made of plaited pandanus leaf was called tatua. Though the waist cloth and the kilt were self fastening, plaited belts were worn as ornaments in full dress by high chiefs.

A belt from Rarotonga in the British Museum, which was obtained from Gill, is about 28.5 feet long including the fringe at each end and 2.5 inches wide. The band is plaited a little over double the width of the belt, but the side edges are turned in, and the band doubled. This technique resembles that of the Maori war belts termed tatua whara or tatua kotara (68, pp. 346-348, pl, 34). Though the Maori belts are made of flax (Phormium tenax), it is interesting that they should retain the name of whara which would be the Maori form of the Rarotongan word 'ara (hara) meaning pandanus. The Rarotongan belt is in both check and twill, whereas the Maori belts are in twill only. Both belts have geometrical patterns worked in black; in the Rarotongan belt this is done with the colored wefts overlaid on the foundation wefts, whereas in the Maori belt some of the foundation wefts are dyed to get the effect. (See plate 6, C.)

page 78
Sacred girdles

In Mangaia, a special girdle was fastened to warriors by the king to give them success in war or in securing a victim for human sacrifice. In an attack on a party of fishermen, the king Vaeruarau halted the warriors on the central hill of Rangimotia where, Gill (33, p. 84) says, "Vaeruarau adjusted on each the girdle sacred to Rongo, invoking the aid of the raven-haired god of war." About the occasion when the last human victim was sought in 1815, Gill (28, p. 302) states, "On a certain evening the victim-seekers assembled on the level top of the central hill, to receive at the hands of the king 'the sacred girdle.' "There is no information as to the form or material of the girdle which was evidently peculiar to Mangaia.


A baldric (tope) was plaited from pandanus leaf into a long, fairly wide band. The length was said to be "e maro e tape" (an arm span and a part). They were favored in Aitutaki, where fancy patterns were formed with overlaid colored wefts (pl. 1, B). The middle of the tope was passed behind the neck, the ends crossed over the breast, and passed backward under the arms. A second tope could be worn with the middle in front of the neck, the two ends crossed over the back and brought forward. The lower ends were fixed by a belt (tatua) of bark cloth wound around the waist. The baldric formed part of the full dress at festivals and was not an article of common dress.

It is evident that the ornamental cord with wrappings of human hair used in Mangaia was also in the nature of a baldric (pl. 8, C).


Sandals (tamaka) are still made of hibiscus bast (kiri 'au) because they are useful for walking over the sharp coral points while fishing and loading and unloading cargo on the fruit ships which visit the islands. The bast is twisted into a single-ply strand which is arranged in longitudinal loops around two sticks, one held between the toes to define the distal end of the sandal and the other held by the left hand parallel with the first at the distance required for the length of the sandal. The bast strand is then passed transversely over and under the longitudinal loops in check to form the sole of the sandal. At the proximal end, a heel loop is formed which with two front strands is used in fastening the sandal. The details of technique have been described for Aitutaki (70, pp. 93-99). In Atiu, the same technique is followed but only the toe stick is used.

Rori, the Mangaian carver, who sheltered in the makatea, owed his many escapes from enemies to his sandals that carried him with ease over the sharp page 79points of makatea to his place of concealment. The Mangaians so associate sandals with the often told story of Rori that they state that sandals were invented by him. However, the distribution of an identical technique throughout the Cook Islands proves that sandals were made long before the period of Rori.

Clothes Containers

Clothes and other articles were probably kept in the house in closely plaited baskets of coconut leaf or pandanus leaf. Tamuera Ariki of Mauke said, however, that a wooden box termed avata was made for storing cloth and cloth beaters. It was made of tou wood and shaped like a large bowl. He quoted the story of Taura-vi 'i, a defeated warrior, who was hidden by his wife in an avata box to escape his enemies. No such box has yet been located in collections.