Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Arts and Crafts of the Cook Islands

Religion

page 308

Religion

The material objects associated with the religion of the Cook Islands were probably as varied as those of the Society Islands. Unfortunately, no artist such as Webber had the opportunity of depicting the native temples when religious ritual was being conducted, hence material for study is meager in comparison with that for the Society Islands and Hawaii. The subject may be divided into the temple with its appurtenances and the material symbols of the gods.

The Temple

The temple (marae) consisted of an open court built up or defined by a curb of rocks, filled in to level the area, and carpeted with gravel, preferably of white coral. A raised platform of stone or wood termed an atarau was erected at one end of the court. Decorative carved slabs termed unu were set up on the platform and on the margins of the court. There were houses for storing the gods, carved slabs, and such other things as may have been used by the priests.

In the campaign conducted by the white missionaries and Tahitian native pastors for the conversion of the people to Christianity, the first requisites were the demolition of the temples and gods. In Aitutaki, the Tahitian pastor, Papeiha, called a meeting of the people and urged them to burn all the temples in the island and bring the remaining "idols" to him. Williams (81, pp. 62, 63) says:

…the assembled multitude yielded their cordial assent. As soon as the meeting broke up, a general conflagration of the maraes took place; and so complete was the destruction, that, on the following morning, not a single idol temple remained unmutilated.

The same Papeiha was later stationed at Rarotonga. When Tinomana Ariki desired to become a Christian, the initial methods of Aitutaki were carried out. Williams (81, p. 152) states:

The teachers informed him that he must destroy his maraes and burn his idols, to which he instantly replied, "Come with me and see them destroyed." On reaching the place he desired some person to take a fire-brand, and set fire to the temple, the atarau, or altar, and the unus, or sacred pieces of carved wood by which the marae was decorated.

A similar fate overtook the maraes in the other islands. All that was inflammable was burnt and even the stones were displaced. In later years, stones from the maraes were used for the walls of taro patches and for other modern utilitarian purposes. The Mokoiro marae in Atiu still shows part of the boundary curbs, consisting of rectangular slabs of coral limestone set longways in the earth, which with the filling raised the marae court above the surface of the surrounding ground. The Maputu marae in Mangaia (76, p. 45) still has its boundaries sharply defined by basaltic rocks imbedded in the ground, forming a low raised rectangular court roughly 116 feet long by 32 page 309feet wide. This marae was supposed to have been filled in with enemy heads. The Court of Royalty named Arai-te-tonga in Rarotonga is not a marae but is an open court with stone rests at one end and a low platform composed of heaped-up rocks at the other. Behind the platform stands the pillar of investiture, named Tau-ma-keva.

Though the names and sites of the maraes in each island are remembered by the older people, most of them are marked only by a few stones that remain imbedded in the ground.

The platforms (atarau) on the court evidently were made either of wood or stone. In Mangaia, the platform on the Akaoro marae, upon which human sacrifices were laid, was definitely made of pandanus wood. On the Orongo marae, the sacrifice was laid on a smooth block of sandstone (76, p. 176).

The decorative slabs or posts (unu) were part of the temple appurtenances and were evidently erected after the court was weeded and before the ritual took place. The Rarotongan history (65, vol. 3, p. 262), in describing a ceremony on the court of Arai-te-tonga, gives the following account in the native text for which I have used my own translation:

Kua vaere te ta'ua Weeded was the court of
i Arai-te-tonga ma Tau-ma-keva. Arai-te-tonga and Tau-ma-keva.
E oti ake te ta'ua kua ipuipui te marae, kua 'akatu te umu [unu], kua iriiri ki te pou ki te marae, kua tau te kai i te takurua. When the court was cleared, set in order was the temple, erected were the carved slabs, hung to the posts on the temple, laid out was the food for the ceremonial feast.

In Savage's translation of the lines "kua iriiri ki te pou ki te marae", he gives "The posts of the marae were hung with flowers", but it is possible that iriiri (to hang upon) refers to unu. It would seem that between ceremonies, the carved slabs were stored away in a house on the marae and set up again on posts after the court was weeded in preparation for the next ceremony. I know of no carved unu slabs from Rarotonga, but several carved slabs from Aitutaki have been preserved and it is probable that these are unu. A bottom projection on these slabs looks as if it might be stuck into the ground, but some of the slabs have coir decorations that would not show up on the ground. If hung on posts, however, the decorations would hang down and render this decorative technique effective. (See figs. 203, b, d and 204.) A similar explanation applies to the spatulate lower ends of the Mitiaro carvings (fig. 225).

Houses on the temple courtyard were a constant feature of the Tahitian and Hawaiian temples, and it is probable that they were erected on the Cook Islands temples. The gods were wrapped in tapa and stored in such houses when not in use in religious ritual. In Mangaia, a miniature house was erected on the marae as a residence for the tribal god after peace had been declared. Such houses ('are ei 'au, house for peace) were about six feet long, were well thatched with pandanus leaves, and had a small doorway screened with white bark cloth (76, p. 173). In Mangaia, representatives of the tribal gods were page 310kept in a national godhouse erected between two important maraes (76, p. 172).

Mangaian gods were wrapped in a special, thick, white bark cloth termed tikoru. The priests also used tikoru, and it may be assumed that the priests of the other islands had special garments that were used during religious ritual. Such garments would be tapu and would probably be stored in a marae house when not in use. Sacred things were too dangerous to keep in the ordinary dwelling houses. The priests of Mangaia overcame the difficulty by partitioning off a portion of their dwelling houses with a thick curtain of bark cloth (76, p. 173). Here they kept a second symbol of their gods and presumably their sacred garments made of tikoru.

Gods

Though most of the gods were committed to the flames on the conversion of the Cook Islanders to Christianity, many were spared to be exhibited at the missionary headquarters in Raiatea. Writing of Aitutaki, Williams (81, p. 53) says:

The gods and bundles of gods, which had escaped destruction, thirty-one in number, were carried in triumph to the boat; and we came off to the vessel with the trophies of our bloodless conquest, "rejoicing as one that findeth great spoil."

Williams (81, p. 91) describes the arrival of the missionaries at Raiatea as follows:

With grateful hearts we now turned our faces homewards; where, after eight or ten days' sail, we arrived in safety. And, as other warriors feel a pride in displaying the trophies of their victories, we hung the rejected idols of Aitutaki to the yard-arms and other parts of the vessel, entered the harbour in triumph, sailed down to the settlement, and dropped anchor amidst the shouts and congratulations of the people.

On the following Friday evening the idols were suspended about the chapel, the chandeliers of which were lighted as before.

From Raiatea the "trophies" were sent to London where the London Missionary Society exhibited them in their museum. Later the collection was transferred to the British Museum, where they can be studied by students of Polynesian culture. Many gods were given to Tyerman and Bennet on their visit of inquiry and Bennet seems to have distributed them among institutions and personal friends. Other gods found their way to Europe and America and into museums where they are reasonably safe. Ethnologists owe a debt of gratitude to the early missionaries for sparing what they did, but it is irritating that details such as their names and functions were not preserved. Williams (81, p. 92) states that he numbered 25 of the Aitutaki gods and transmitted them with their names and histories to the Tyerman and Bennet deputation then at Tahiti.

A study of the material gods shows that though all the islands of the group shared the same major gods such as Tangaroa, Tane, Rongo, and others, the page 311forms differ in each island. In Rarotonga, the technique was confined to wooden images of various forms. Aitutaki had both wooden images and carved slabs. In Atiu and Mitiaro, carved wooden stands were made on the round, with arches for feather decoration. Atiu also used sennit bundles covered with red feathers. Unfortunately, two gods attributed to Mauke are more than doubtful. In Mangaia, grooved slabs on a round base were carved with a characteristic K-motif and decorated with feathers and human hair. Small objects made of tapa cloth, feathers, and human hair represented minor deities. In an island group with the two most widely separated islands only 180 miles apart and with many gods identical in name, there is an astounding variation in the technique and carving motifs. The minds of people seem to have carried similar ideas regarding ritual, but human hands shaped different theological conceptions into material form. The material forms are treated under each island.

Rarotongan Gods

Rarotongan gods were of wood carved in conventionalized human forms. Those available for study comprise two groups: (1) images consisting of one main figure and (2) multiple figures with an intermediate plain staff around which tapa was wrapped originally.

Images

The images consist of two forms: the so-called fisherman's god and a figure with three small figures on its breast, said to be Te Rongo.

A number of images called the fisherman's god are found in museums (pl. 13, D, E). Their function rests solely on the authority of Williams, who figured one of them (81, p. 100) and referred to it thus (81, p. 99):

An idol, of which the figure on the opposite side is a correct representation, was placed on the fore part of every fishing canoe; and when the natives were going on a fishing expedition, prior to setting off, they invariably presented offerings to the god, and invoked him to grant them success.

Ellis (25, vol. 2) includes a similar figure in a frontispiece with a number of other gods, and labels it and a Tahitian image with the same number, 7. In the figure legend he refers to 7 as "Orametua." He has thus confused a Rarotongan image with the orometua or ti'i images of Tahiti. The example shown in figure 191 in the London Missionary Society collection bears the following inscription: "Figure of Tarianui (Great Ears), A God of fishing, carried on the prow of a fishing canoe." Williams and Ellis carelessly failed to keep to the dialect of the island for which they described artifacts and customs. It is obvious that Tarianui is the Tahitian form (Tari'a-nui) of the Rarotongan Taringa-nui (Big Ears), and it would be more reassuring had the correct spelling been used. When Williams (81, p. 74) visited Mitiaro with the Atiuan chief, Rongo-ma-Tane, the people asked whether they should burn page 312Tarianui, or Great Ears. Here again the Tahitian dialect was used by Williams, for the ng sound in taringa and other words was used throughout the Cook group. There is no evidence that the Rarotongan god had the same name as the Atiuan god, and it is probable that the name was given by museum people because the fisherman's god happened to have big ears (fig. 191).

Figure 191.—Fisherman's god (British Mus., L.M.S.): height 440 mm.; head width above ears 184 mm.; width at ears 240 mm.; width at shoulders 240 mm.; narrowest width of body 111 mm.; penis amputated and navel protuberance broken off. a, front; b, back; c, right side.

Figure 191.—Fisherman's god (British Mus., L.M.S.): height 440 mm.; head width above ears 184 mm.; width at ears 240 mm.; width at shoulders 240 mm.; narrowest width of body 111 mm.; penis amputated and navel protuberance broken off. a, front; b, back; c, right side.

The characteristics of the fisherman's god are the short, thick-set, sturdy body with a large well-formed head with high forehead and specialized eyes, mouth, and ears. The head is sufficiently large to allow the craftsman to carve in details in a bold and striking manner. The eyes are extremely large, the inner angles almost meeting and the outer angles well out on the sides of the face. The eye areas are left as large elliptical areas projecting above the general plane of the face. Each eye consists of four elements: eyeball, lower and upper lids, and brow. The eyeball is elliptical, being pointed at the inner and outer angles. The lids and brow are formed in bold curves which form raised flanges. The two brows meet in the middle line. The curved planes of the cheeks meet in the middle line to form a vertical edge which extends downward in a straight line from the eyes to the mouth with no attempt to define a nose. The mouth is formed by four curved lines cut down into the general plane of the face, wider in the middle and narrowing to the outer ends. They define what I assume to be a middle tongue and upper and lower lips. The lowest of the curved lines is placed so that the chin is of the same thickness as the lips and tongue, which lends another characteristic feature to the appearance of the images.

page 313

The ears are set slightly back from the outer angles of the eyes and form projecting ellipses from which the inner part has been cut away to conform to the outer shape. A hole, usually square, is pierced through the posterior part. The straight chin and lower jaw line is carried obliquely upward below the ear and is continued horizontally across the back of the head. The wood is cut away between this line and the shoulders, both at the sides and back, and thus results in a distinctive, clearly denned, short neck.

The shoulder line is carried across the back as a straight edge. The shoulders are square, and large perforations free the upper arms from the chest. The forearms are flexed and the large expanded hands rest on the sides of the abdomen. Five or sometimes four fingers are defined by straight incisions. Nipples are shown, and the abdomen is large and distended with a protruding navel usually circular in form. A phallus with a large glans was always present, but in some museum specimens it had been amputated either by the missionaries or native converts. It is absent in the British Museum specimen figured (fig. 191), and the drawings given by Williams and Ellis were probably drawn from this specimen. A similar image in the Munich Museum (no. 191) has the phallus, but one in the Royal Scottish Museum, Edinburgh (1923, 360) has been mutilated. The thighs and legs are flexed and the gluteal region, somewhat shallow in depth, has a marked backward projection. The feet are solid blocks, but some are notched on the front lower edge to represent toes.

Figure 192.—Painted fisherman's god (British Mus., 9866): height 315 mm.; head width 129 mm.; shoulder width 155 mm.; hip width 148 mm. a, front; b, back; c, right side, showing sites of painting; d, painted motifs: 1, on outer sides of thighs, with points toward knees; 2, on both sides of breast; 3, in front of ears; 4, on shoulders; 5, on front of forearms; 6, on sides and back of upper arms. Inner grooves of eyes and mouth filled in as plain bands; simple triangles along upper boundary of mouth; serrated collar around back of neck, navel, and glans penis.

Figure 192.—Painted fisherman's god (British Mus., 9866): height 315 mm.; head width 129 mm.; shoulder width 155 mm.; hip width 148 mm. a, front; b, back; c, right side, showing sites of painting; d, painted motifs: 1, on outer sides of thighs, with points toward knees; 2, on both sides of breast; 3, in front of ears; 4, on shoulders; 5, on front of forearms; 6, on sides and back of upper arms. Inner grooves of eyes and mouth filled in as plain bands; simple triangles along upper boundary of mouth; serrated collar around back of neck, navel, and glans penis.

Two images in the British Museum and the Peabody Museum, Cambridge, are painted. Some of the motifs are the same as those used in tattooing, but they are not always placed in the same position as similar designs on the page 314human body. The hands have four fingers, the protruding navel is round, and the feet blocks are notched on the lower border in front and on the outer sides (fig. 192).

The three-limbed motif (fig. 192, d, 3), somewhat resembling the rau-teve motif in tattooing, is placed on the front of the ear instead of the back. The vertebra motif (fig. 192, d,1, upper part) is on the thighs instead of the back of the neck. The other painted image has a simple motif of wide bands, most of them with serrated edges (fig. 193).

Figure 193.—Painted fisherman's god (Peabody Mus., Cambridge, 53,517): height 420 mm.; head height 224 mm.; shoulder width 180 mm. a, front: typical eyes and mouth, painting around eyes and navel, small nipples, round navel, anterior axillary line denned by edge, five fingers, serration on outer and front sides of feet, b, back: protruding ears perforated, denned neck, straight edge across shoulders, slender waist, shallow gluteal region, c, right side: protruding abdomen and navel, arm with distinct outer surface, squat legs. Painted patterns confined to bands, some with serrated edges, as shown.

Figure 193.—Painted fisherman's god (Peabody Mus., Cambridge, 53,517): height 420 mm.; head height 224 mm.; shoulder width 180 mm. a, front: typical eyes and mouth, painting around eyes and navel, small nipples, round navel, anterior axillary line denned by edge, five fingers, serration on outer and front sides of feet, b, back: protruding ears perforated, denned neck, straight edge across shoulders, slender waist, shallow gluteal region, c, right side: protruding abdomen and navel, arm with distinct outer surface, squat legs. Painted patterns confined to bands, some with serrated edges, as shown.

The second type of human figure is larger than the fisherman's god and bears three small human figures on its breast. The head is large with well-flanged eyes, notched mouth, and ears of the same form as those of the fisherman's god, but the head and face are narrower in proportion to height. A new element is the well-shaped nose with carved nostrils. A long, well-shaped neck separates the jaw line from the shoulders, and the upper shoulder line is carried across the back as a straight edge. Two projecting breasts are high up on the chest, and the three small figures in full face appear below them, the middle figure being in the mesial line. The abdomen is large and protruding with a circular raised navel pitted in its center. The upper arm is long and vertical, free from the chest, and is carved with full-face human figures on its outer surface. The hand is large, four-fingered, and flexed onto the side of the page 315abdomen with no distinctive forearm. The gluteal region projects outward at a marked angle with the body and is comparatively short. The long legs are flexed on short thighs, and the feet form solid blocks which are notched on front and sides. The phallus is present with a well-marked glans (fig. 194).

Figure 194.—Image with smaller figures (British Mus., L.M.S., 42). a, front: total height 690 mm.; width of head with projecting ears 127 mm.; length of head 115 mm.; shoulder width 162 mm.; abdominal width at navel 127 mm.; hip width 160 mm., showing three figures on breast. The rectangular feet are notched on upper and lower front edges. b, side: showing projection outward of 3 breast figures, and 2 figures carved on flat on outer side of arm; middle part of arm uncarved for length of 43 mm., covered with thin white tapa over which sennit braid 3 mm. wide is wrapped with close transverse turns; remains of feathers indicate that the braid turns formed lashing for feather decoration; hand similar to fisherman's god but 4 fingers; side of foot ornamented with alternating angular notches to form zigzag decorations; nose well shaped, c, back: continuation of jaw line across the back of head defines distinctive neck, longer than in fisherman's god; shallow gluteal region distinctive, d, middle breast figure: flanged eye with brow flange, and mouth defined by three notches; protruding abdomen with circular navel, but navel absent in other two breast figures; all 3 have penis (1); arms (2) are represented by raised ridge; flexed lower limb (3) notched, e, arm figure: large ears formed of rectangular panels with corners rounded off; eyes without brow flange; mouth defined by two curved lines; arms (1) represented by horizontal ridge, abdomen (2) by knob, and flexed legs (3) by angular projections; vertical mesial element (4) represents penis; upper arm figure 63 mm. high and lower 55 mm.

Figure 194.—Image with smaller figures (British Mus., L.M.S., 42). a, front: total height 690 mm.; width of head with projecting ears 127 mm.; length of head 115 mm.; shoulder width 162 mm.; abdominal width at navel 127 mm.; hip width 160 mm., showing three figures on breast. The rectangular feet are notched on upper and lower front edges. b, side: showing projection outward of 3 breast figures, and 2 figures carved on flat on outer side of arm; middle part of arm uncarved for length of 43 mm., covered with thin white tapa over which sennit braid 3 mm. wide is wrapped with close transverse turns; remains of feathers indicate that the braid turns formed lashing for feather decoration; hand similar to fisherman's god but 4 fingers; side of foot ornamented with alternating angular notches to form zigzag decorations; nose well shaped, c, back: continuation of jaw line across the back of head defines distinctive neck, longer than in fisherman's god; shallow gluteal region distinctive, d, middle breast figure: flanged eye with brow flange, and mouth defined by three notches; protruding abdomen with circular navel, but navel absent in other two breast figures; all 3 have penis (1); arms (2) are represented by raised ridge; flexed lower limb (3) notched, e, arm figure: large ears formed of rectangular panels with corners rounded off; eyes without brow flange; mouth defined by two curved lines; arms (1) represented by horizontal ridge, abdomen (2) by knob, and flexed legs (3) by angular projections; vertical mesial element (4) represents penis; upper arm figure 63 mm. high and lower 55 mm.

The image is made of ironwood and has a fine polish. Its greater height gives it a more slender appearance than the fisherman's god, yet they have certain elements in common. The general shape of the head and the technique of the eyes, mouth, and ears is identical. The continuation of the jaw line across the back and the straight shoulder line across the back are similar and the resultant neck differs merely in length. The arm and hand technique is page 316similar, but utilization of the outer surface of the arm for carving human figures has led to the flexion being near the wrist. The protruding abdomen, circular navel, shallow gluteal region, flexed thighs and legs with solid rectangular feet are present in both types, but the notches in the fisherman's god have undergone less elaboration. The characteristic penis with large glans is present in both.

Elaboration has taken place in the longer neck, the definite nose with well-defined nostrils, and the addition of secondary figures. The secondary figures are of two types; breast figures carved in the round and arm figures carved on the flat. In both forms the artist's field is restricted in area, and the details of technique established on larger figures were difficult to produce on the smaller figures. The result has been the natural elimination of some details that would have required more space than was available. Thus while the characteristic eye pattern was observed, the brow element was omitted in both types. The full mouth pattern consists of four curved notches, but in the breast figures one notch was omitted and in the arm figures two were omitted. The ears offered no problem, and in the arm figures they are exceptionally large to balance the rectangular panel that forms the field of the figure. In the breast figures the large protruding abdomen is maintained because the figure is carved in the round, but in the flat arm figure the abdomen is represented by a knob. The upper limbs lose details as to forearm and hand in both types and are represented by raised ridges, which are separated in the round figure, owing to distance, but run together in the flat figure. The lower limbs in the round figure are capable of following the flexed pattern, but the feet are left out. The angular notches associated with the feet of larger figures are included on the lower edge of the flexed leg. In the flat figure the flexed limb pattern becomes flattened out with the knee angle toward the middle line, and neither feet nor notches are present. Thus extreme conventionalization occurs in the smaller figures influenced primarily, no doubt, by the limitations of space but also affected by an artist's ideas of values. A secondary figure is an ornamental detail and does not need the elaboration of the main figure.

The London Missionary Society catalog merely describes the figure in general terms but adds "From Rarotonga." In pencil on an old label are the words, "Te Rongo and his three sons from Rarotonga." While the allocation to Rarotonga is undoubtedly correct, the identification as Te Rongo and his three sons may be due to identifying the Mangaian god Rongo with the Raro-tongan god Te Rongo. The Mangaian god Rongo was the father of the three ancestors of the Ngariki tribe but there is no record that the Rarotongan god Te Rongo had three sons.

Staff Gods

Staff gods are long staves of ironwood with an expanded upper end carved into a large head in profile, succeeded by a number of smaller figures. Similar page 317small figures are carved at the lower end, which ends in a phallus. The middle staff part was covered with rolls of bark cloth, but in most museum specimens the cloth has disappeared. Gods with a middle staff may be divided into three varieties: large, small, and double-figure staff gods.

Apart from the presence of the face technique typical of the Rarotongan fisherman's god, the identification has been placed beyond doubt by Williams (81, pp. 98, 99):

A day or two afterwards, they requested us to take our seat outside the door; and, on doing so, we observed a large concourse of people coming towards us, bearing heavy burdens. They walked in procession, and dropped at our feet fourteen immense idols, the smallest of which was about five yards in length. Each of these was composed of a piece of aito, or iron-wood, about four inches in diameter, carved with rude imitations of the human head at one end and with an obscene figure at the other, wrapped round with native cloth, until it became two or three yards in circumference. Near the wood were red feathers, and a string of small pieces of polished pearl shells, which were said to be the manava, or soul of the god. Some of the idols were torn to pieces before our eyes; others were reserved to decorate the rafters of the chapel we proposed to erect; and one was kept to be sent to England, which is now in the Missionary Museum. It is not, however, so respectable in appearance as when in its own country; for his Britannic Majesty's officers, fearing lest the god should be made a vehicle for defrauding the king, very unceremoniously took it to pieces; and, not being so well skilled in making gods as in protecting the revenue, they have not made it so handsome as when it was an object of veneration to the deluded Rarotongans.

Makea Ariki of Rarotonga accompanied Williams on his return to Raiatea, and a number of the Rarotongan gods were taken with them. Williams refers to them thus (81, p. 145):

… we prepared for our missionary meeting; at which from two to three thousand people assembled, many of whom had come from Huahine and Tahaa; with the noble chief, also, from Rarotonga, whose presence, together with the exhibition of the neglected idols of his people, added to the interest of the occasion. This was the third time we had enjoyed the privilege of exhibiting to the Raiateans the abandoned idols of other islands.

It is evident that the various types of gods in the London Missionary Society case in the British Museum were sent home from the accumulation of "trophies" at Raiatea.

Large staff gods. The frontispiece in William's work (81) shows the missionaries Williams and Pitman with their wives seated on chairs in front of their home in Rarotonga. Lying on the ground before them are nine large images with the middle parts wrapped in tapa. Six bearers are carrying in another image standing erect on a litter formed of two poles with crossbars, resembling a ladder. The lower part of the tapa bundle rests on the poles and crossbars of the litter and thus supports the god in the erect position. One of these images sent to England to the Missionary Museum was figured by Williams (81, p. 100), thus placing the identification of these images beyond doubt. It is shown in figure 195 with additional details. Details of the large head are shown in figure 196, d, and the pattern of the bark-cloth wrapping in figure 28, b.

page 318
Williams (81, p. 99) mentions that red feathers were placed next to the wood under the cloth with pieces of polished pearl shell to form the manava or "soul" of the god. This may have been a constant feature, but, as the specimen sent to the Missionary Museum and now in the British Museum is the only large staff god with a tapa wrapping, it is impossible to obtain confirmatory evidence. Owing probably to the length of the larger staff gods, the specimens found in museums consist of the upper section with the large head
Figure 195.—Rarotongan large staff god (British Mus., L.M.S.), figured by Williams (81, p. 100): total length about 13 feet, a, full figure: 1, head length 102 mm., greatest width at back 46 mm., tufts of feathers attached to left ear (see fig. 196, d); 2, upper set of five small figures, arranged as in figure 197, c; profile figures with phallus decorated with feathers (fig. 195, d); 3, bundle of tapa cloth wrapped around middle staff section and tied with bands of cloth, red feathers and pearl shell were next wood; 4, lower set of five small figures; 5, phallic terminal, b, over 30 small pieces of pearl shell figured by Williams, said to be "manava or soul of the god." c, feather ear ornament: four split black feathers 96 mm. long tied to each of four holders of a loop carrier with same technique as in war cap; two-ply cord of loose white fiber threaded through loops of six carriers and tied through hole in ear. d, feathers tied to phallus below chin of profile figures with fine sennit, 3 mm. wi'de, passing in transverse turns over feather quill and tied to neighboring turn with overhand knot; feathers, single, split, black, and 90 mm. long.

Figure 195.—Rarotongan large staff god (British Mus., L.M.S.), figured by Williams (81, p. 100): total length about 13 feet, a, full figure: 1, head length 102 mm., greatest width at back 46 mm., tufts of feathers attached to left ear (see fig. 196, d); 2, upper set of five small figures, arranged as in figure 197, c; profile figures with phallus decorated with feathers (fig. 195, d); 3, bundle of tapa cloth wrapped around middle staff section and tied with bands of cloth, red feathers and pearl shell were next wood; 4, lower set of five small figures; 5, phallic terminal, b, over 30 small pieces of pearl shell figured by Williams, said to be "manava or soul of the god." c, feather ear ornament: four split black feathers 96 mm. long tied to each of four holders of a loop carrier with same technique as in war cap; two-ply cord of loose white fiber threaded through loops of six carriers and tied through hole in ear. d, feathers tied to phallus below chin of profile figures with fine sennit, 3 mm. wi'de, passing in transverse turns over feather quill and tied to neighboring turn with overhand knot; feathers, single, split, black, and 90 mm. long.

page 319 and first series of smaller figures, the lower two sections having been cut off and discarded by the original collectors because of the difficulty in transporting such large objects. (See plate 13, C.) The correct length of the image sent to London is 13 feet. The upper carved section of this image is 23.5 inches long. The upper carved sections of nine other images studied by me ranged from 24 to 42.5 inches, hence it may be assumed that the total length of each of them was greater than the 13-foot complete specimen. Ellis (25, vol. 2, p. 221) says that "Mr. Bourne in 1825, saw fourteen about twenty feet long and six feet wide." The six feet wide must refer to the bark-cloth wrapping, but even this seems exaggerated.

Of the nine incomplete specimens the British Museum, Cambridge University Museum, Peabody Museum (Cambridge, Mass.), and Munich Museum have one each; and the Oldman collection has five. Edge-Partington (24, 1-21-6) figured the complete specimen, and the upper segment was figured by Ellis (25, vol. 2, frontispiece) and Oldman (54, vol. 47, pis. 1-3).

In preparing the timber for the large gods, the craftsman adzed the upper part into a thick plank and trimmed the section for the large head so that the two surfaces formed two inclined planes which met at an edge in front. The back retained the full thickness of the plank. The front and back lines of the head are vertical and parallel, and the front mesial edge extends from the top of the head to the chin. The flanged eye is carved some distance below the vertex and gives the effect of a high vertical forehead. The eye technique is similar to that in the fisherman's god, with a transverse pointed ellipse for the eyeball, a curved raised flange below for the lower lid, and two similar flanges above for the upper lid and the eyebrow. The eye extends across the lateral surface of the face, leaving just enough space behind it for the raised ear. The ear and mouth have the same techniques as the fisherman's god, and there is no nose. From the chin, the lower jaw line inclines upward and backward in a straight line to meet the vertical posterior line of the head and then cuts horizontally across the back. Below this line, a square-cut groove at the sides and back form a short neck. Below the neck comes the shoulder, arm, and hand which form the posterior boundary of the first small figure. The posterior border of the shoulder and arm is serrated. In ten specimens the vertical height of the head from chin to vertex ranged from 195 mm. to 310 mm.; the head width from the front back from 81 mm. to 160 mm.; and the thickness at the back from 45 mm. to 72 mm. (See figure 196.)

In describing the upper series of small figures it is convenient to place the artifact on its back, as in figure 197. In this position the upper edge of the wood was cut with a number of fairly deep and wide notches to form rectangular projections. Commencing from the large head, the projections were carved alternately in full face and in profile. The full face figures were carved with page 320large projecting ears, an extremely conventionalized face, and with raised elements representing the body and upper and lower limbs. Similar carving is executed on each side of the projection. The full-face figures are double, being placed back to back and with a notch separating the upper parts of the head. The alternate figures in profile are single and have well-defined eyes similar to those of the large head but without eyebrows. They also have large ears, an element representing a limb, and a male organ projecting forward from beneath the chin. They all face toward the phallic terminal. The lower, unnotched part of the wood is pierced with triangular holes which help to define a second row of even more conventionalized figures, and they pair with the figures on the notched projections above them. The projection nearest the large head is always carved in full face, and there is no second row figure below it because the space is occupied by the arm of the large head. If the arm is high, the small space below it may be ornamented with vertical zigzag lines.

Figure 196.—Head of large staff god: a-c (British Mus., 1919-10-14); d (British Mus., L-M.S.). a, right side: height from chin 310 mm.; length from front back, at ear level, 160 mm. Figure shows the following features: 1, eyeball; 2, upper lid; 3, lower lid; 4, eyebrow; 5, tongue; 6, upper lip; 7, lower lip; 8, chin; 9, ear; 10, neck-groove; 11, shoulder with arm and notched posterior border; 12, hand with fingers; 13, first figure of upper series; 14, projection of small figure facing backward, b, front: various parts numbered as in a;13, top of first figures of upper series, showing that there are two facing outward on either side; figure shows how narrow head is as. compared with side view, c, back: width 72 mm.; ears (9) shown with square holes, (15) pierced through back part of ear flange and appearing on back of head; 10, neck groove; 14, small raised figure; back of head flat except for median edge (16) which extends down from top for short distance; back surface formed by posterior thickness of plank from which inclined planes proceeded forward. d, ornament of split black feathers attached to left ear through pierced hole. (See figure 195, a.)

Figure 196.—Head of large staff god: a-c (British Mus., 1919-10-14); d (British Mus., L-M.S.). a, right side: height from chin 310 mm.; length from front back, at ear level, 160 mm. Figure shows the following features: 1, eyeball; 2, upper lid; 3, lower lid; 4, eyebrow; 5, tongue; 6, upper lip; 7, lower lip; 8, chin; 9, ear; 10, neck-groove; 11, shoulder with arm and notched posterior border; 12, hand with fingers; 13, first figure of upper series; 14, projection of small figure facing backward, b, front: various parts numbered as in a;13, top of first figures of upper series, showing that there are two facing outward on either side; figure shows how narrow head is as. compared with side view, c, back: width 72 mm.; ears (9) shown with square holes, (15) pierced through back part of ear flange and appearing on back of head; 10, neck groove; 14, small raised figure; back of head flat except for median edge (16) which extends down from top for short distance; back surface formed by posterior thickness of plank from which inclined planes proceeded forward. d, ornament of split black feathers attached to left ear through pierced hole. (See figure 195, a.)

At the end of the double row of figures, there is usually a single full-faced figure carved on either side of the beginning of the staff section. The number of figures varies in the different gods, but the number in the top row ranges from four to eight (fig. 197).

page 321
Figure 197.—Small figures in large staff god. a, wood prepared by notching upper surface to form panels for alternate full-face figures (1) and figures in profile (2); part below upper panels with triangular perforations to define lower set of larger figures (3) below upper profile figures and smaller figures (4) below upper full-face figures. b, carved figures (British Mus., 1919-10-14): first full figure (11) of upper row with conical head (1), large ears (2) bounding highly conventionalized faces formed by 2 upper oblique grooves and 2 lower horizontal; upper limbs (3) formed of single raised ridge; lower limb (4) flexed with knees bent inward; abdomen (5) formed of rounded knob. Profile figure (12) of upper row with long flat vertex (1), large ear (2), shaped eye (6) formed of distinct eyeball with orbit, mouth (7) formed by 2 notches and marked chin and jaw line; upper limb absent; lower limb (4) flexed under jaw; and phallus (8). 13, second full figure of upper row with details similar to first figure (11) except that lower limb has 2 angular notches (9) to represent toes. 14, larger figure of lower row beneath profile figure of upper row; 2 large ears (2) bounding conventional face of same form as upper full figure, but part below second horizontal groove is divided vertically; two long triangular perforations (10) bound elongated body (5) with upper limbs (3) above carrying turned-up hands with 3 fingers; lower limbs (4) bent upward with feet carrying 3 toes bent downward. 15, smaller figure of lower row beneath full figure of upper row; two large ears (2) bounding conventional face similar to face of other full figures, with upper limbs (3) of single ridge; lower limb (4) in which upper and lower parts are fused through carver not cutting outer notches as in lower limb of 13, but notches for toes present; body represented by small knob (5). Parts left uncarved have figures corresponding to the above, except that there is no lower figure under the first full figure (11), as space beneath is partly occupied by arm (16) coming from the large head. This arm projects outward beyond lower limbs of full-face figure above it, and consequently the lower limbs of this figure are not notched for toes, c, staff god (Cambridge Univ. Mus., Z.6099). The alternate full face (1) and profile (2) figures resemble those in b, except that the upper profile figures have a single straight ridge to represent upper limb, while lower figures (15) beneath upper full faces are so compressed through smaller space that upper limbs are omitted; arm (16) of large head is serrated on its lower border and ends in a hand with 3 fingers; upper full-face figures are in pairs, and backs of opposite figures (17) may be distinguished, d, upper (front) view of c: tops of paired full-face figures (1, 1) and single profile figures (2) are shown, and chin (3) of large head. e, lower (back) view of c: conventional full-face figure (1) below neck of large head, with notches down middle line and perforations from side showing.

Figure 197.—Small figures in large staff god. a, wood prepared by notching upper surface to form panels for alternate full-face figures (1) and figures in profile (2); part below upper panels with triangular perforations to define lower set of larger figures (3) below upper profile figures and smaller figures (4) below upper full-face figures. b, carved figures (British Mus., 1919-10-14): first full figure (11) of upper row with conical head (1), large ears (2) bounding highly conventionalized faces formed by 2 upper oblique grooves and 2 lower horizontal; upper limbs (3) formed of single raised ridge; lower limb (4) flexed with knees bent inward; abdomen (5) formed of rounded knob. Profile figure (12) of upper row with long flat vertex (1), large ear (2), shaped eye (6) formed of distinct eyeball with orbit, mouth (7) formed by 2 notches and marked chin and jaw line; upper limb absent; lower limb (4) flexed under jaw; and phallus (8). 13, second full figure of upper row with details similar to first figure (11) except that lower limb has 2 angular notches (9) to represent toes. 14, larger figure of lower row beneath profile figure of upper row; 2 large ears (2) bounding conventional face of same form as upper full figure, but part below second horizontal groove is divided vertically; two long triangular perforations (10) bound elongated body (5) with upper limbs (3) above carrying turned-up hands with 3 fingers; lower limbs (4) bent upward with feet carrying 3 toes bent downward. 15, smaller figure of lower row beneath full figure of upper row; two large ears (2) bounding conventional face similar to face of other full figures, with upper limbs (3) of single ridge; lower limb (4) in which upper and lower parts are fused through carver not cutting outer notches as in lower limb of 13, but notches for toes present; body represented by small knob (5). Parts left uncarved have figures corresponding to the above, except that there is no lower figure under the first full figure (11), as space beneath is partly occupied by arm (16) coming from the large head. This arm projects outward beyond lower limbs of full-face figure above it, and consequently the lower limbs of this figure are not notched for toes, c, staff god (Cambridge Univ. Mus., Z.6099). The alternate full face (1) and profile (2) figures resemble those in b, except that the upper profile figures have a single straight ridge to represent upper limb, while lower figures (15) beneath upper full faces are so compressed through smaller space that upper limbs are omitted; arm (16) of large head is serrated on its lower border and ends in a hand with 3 fingers; upper full-face figures are in pairs, and backs of opposite figures (17) may be distinguished, d, upper (front) view of c: tops of paired full-face figures (1, 1) and single profile figures (2) are shown, and chin (3) of large head. e, lower (back) view of c: conventional full-face figure (1) below neck of large head, with notches down middle line and perforations from side showing.

page 322

On the back surface, there is a full-face figure on the same level as the first small figure near the large head and the rest of the surface is notched transversely at regular intervals (fig. 197, e).

The middle section which is the longest is a plain rounded staff covered with rolls of tapa forming a large-sized wrapping. It is absent in the museum specimens that have been cut short, but fortunately the technique is present in the complete British Museum specimen shown in figure 195. Even in this specimen, the upper part had been severed and afterwards attached to the middle staff by a metal screw.

The lower section also has small figures arranged in two rows and it ends in a phallus with a well-defined glans. In the British Museum specimen there are five raised figures in a group of three separated by a short interval from an end group of two. The three consist of a middle full-face figure with a profile figure on either side looking away from the middle figure. The end two are profile figures back to back with the end one looking toward the terminal (fig. 195). This is the only specimen of the large staff gods in which the lower section has been preserved.

Small staff gods. Small staff gods are like the large gods, but less carefully carved, with the edges between chipped hollows unsmoothed. They have been preserved whole, except one in the London Missionary Society's collection which has had the terminal phallus removed. Of eight specimens studied, page 323six ranged in total length from 736 mm. to 1,638 mm. The other two were rather long; one in the Oldman collection (435) is 2,534 mm. and the other, in the Royal Scottish Museum, Edinburgh, is 3,022 mm. The Edinburgh specimen is well carved and forms a link between the small and the large staff gods. The eight specimens studied are in the British Museum (2), Royal Scottish Museum (1), and the Oldman collection (5). Another in the Basle Mission House Museum was figured by Stolpe (64, p. 42, fig. 50). Oldman (54, vol. 47, pls. 1-3) figured his own collection. (See figure 198.)

Figure 198.—Smaller staff god (Oldman coll., 424): total length 45 inches, a, upper head end (1), 17.6 inches, part of middle section (2), wrapped in three layers of tapa, white, brown, and outer with black painted zigzag design: binding of sennit wound spirally in three crossing courses, b, lower section (3), 11.75 inches; with part of middle section (2); total length of middle section, 21.6 inches.

Figure 198.—Smaller staff god (Oldman coll., 424): total length 45 inches, a, upper head end (1), 17.6 inches, part of middle section (2), wrapped in three layers of tapa, white, brown, and outer with black painted zigzag design: binding of sennit wound spirally in three crossing courses, b, lower section (3), 11.75 inches; with part of middle section (2); total length of middle section, 21.6 inches.

In this series, the height of the large head ranged from 105 mm. to 182 mm.; the head width from the front back from 45 mm. to 81 mm.; and the thickness at the back from 33 mm. to 45 mm. The eyes of the large head were formed by incising curves on the general plane of the face and consequently there are no raised eye flanges. The lines of eyes and mouth are not always regular and they may be reduced in number. The ears are not pierced.

The small figures are fewer in this series. Though upper and lower rows are usual, some single figures, both full-face and profile, occupy the full depth of the surface. In the large staff gods, the first small figure is always full-face, but in the series of eight small gods, two are in profile. There are no triangular perforations to define the lower figures, but triangular depressions may be present in some. Below the hand of the large head, however, there is usually a square perforation. In some, the small figures are replaced by a full plain panel with serrated edges; in others, the lower figures are replaced by a serrated plain panel, zigzag lines, or dots. The profile figures are devoid of a phallus. Most of the specimens have the small full-face figure on the commencement of the staff section. Further details are shown in figure 199.

page 324
Figure 199.—Smaller staff gods (a-d, British Mus., L.M.S., 306-19); e-f, British Mus., L.M.S., 18-20). a, full-length figure showing head (1), upper series of small figures (2), middle staff (3), lower series of figures (4), and terminal phallus (5); total length 1,640 mm. b, head and upper series of figures: head (1) with eye elements of eyeball, upper and lower lids, and brow; mouth defined by three notches instead of four as in large staff gods; typical ear and no nose; notched arm (2) and hand of the large head forms lower boundary of first figure; first figure (3) is full-face with large ears, poorly carved conventional face, and limbs represented by a horizontal ridge on either side; second figure (4) in profile with secondary figure (5) below it in full face with large ears, conventionalized face, and triangular depressions defining body and bent leg elements; last secondary figure (6) with large ears, conventional face, and two limb elements on either side, has no projecting figure above it. Large head (1), 133 mm. high from chin to vertex and 72 mm. from front back. Upper series of figures extends for 170 mm. beyond chin of large head. Cross diameters of staff part (7) are 38 mm. by 32 mm. c, lower series of figures in upper and lower sets and terminal phallus: upper set consists of two profile figures (1, 1) looking away from a middle full-faced figure (2). Projecting profile figures consist of head with single horizontal element to represent limbs, and each has secondary full-face figure (3, 3) below it; secondary figures (3, 3) have well-marked ears, with conventional faces, and horizontal elements to represent limbs; middle full face (2) has part above one ear broken away, and with the two ears and extremely conventionalized face it occupies projecting part of wood, and part below (4), which, in large staff gods, would be occupied by a secondary figure is here elaborated into elements representing a body and limbs for head above; thus below the face are two horizontal ridges representing upper limbs with a rounded knob for the body between them; below these again are two horizontal elements on either side, representing lower limbs, but one is cut through at knee joint. Two last figures (5, 6) are in projecting profile looking away from each other and with gap between; first (5) has single upper limb element beneath it followed by full-face figure (7) with large ears bounding conventional face and bent elements for lower limbs; second (6) has conventional profile face with upper limb flexed to hand with three fingers and a bent lower limb. First series (1, 1) occupies length of 140 mm. and second (5, 6) a length of 98 mm. Phallus (8) is 110 mm. long, d, under surface, lower end: 1, notched part under first set of figures; 2, notched part under second set of figures; 3, phallus, e, head and upper figure series: head (1) conforms to previous pattern with serrated arm (2) and three-fingered hand, but the figure (3) above arm is in profile instead of full face; next projection (4) is notched on each border without any carved figure; height of head (1) 118 mm., maximum length 68 mm., width 33 mm.; staff (5) is 39 mm. in diameter, f, lower end of e with phallus amputated: first figure (1) in projecting profile with lower part (2) notched instead of secondary figure; last pair (3, 4) in profile facing away; one (3) consists of head only with part below it notched; the other (4) has normal head with upper limb and three-fingered hand, and horizontal ridge for lower limb. Total length of staff god without phallus 1,025 mm.

Figure 199.—Smaller staff gods (a-d, British Mus., L.M.S., 306-19); e-f, British Mus., L.M.S., 18-20). a, full-length figure showing head (1), upper series of small figures (2), middle staff (3), lower series of figures (4), and terminal phallus (5); total length 1,640 mm. b, head and upper series of figures: head (1) with eye elements of eyeball, upper and lower lids, and brow; mouth defined by three notches instead of four as in large staff gods; typical ear and no nose; notched arm (2) and hand of the large head forms lower boundary of first figure; first figure (3) is full-face with large ears, poorly carved conventional face, and limbs represented by a horizontal ridge on either side; second figure (4) in profile with secondary figure (5) below it in full face with large ears, conventionalized face, and triangular depressions defining body and bent leg elements; last secondary figure (6) with large ears, conventional face, and two limb elements on either side, has no projecting figure above it. Large head (1), 133 mm. high from chin to vertex and 72 mm. from front back. Upper series of figures extends for 170 mm. beyond chin of large head. Cross diameters of staff part (7) are 38 mm. by 32 mm. c, lower series of figures in upper and lower sets and terminal phallus: upper set consists of two profile figures (1, 1) looking away from a middle full-faced figure (2). Projecting profile figures consist of head with single horizontal element to represent limbs, and each has secondary full-face figure (3, 3) below it; secondary figures (3, 3) have well-marked ears, with conventional faces, and horizontal elements to represent limbs; middle full face (2) has part above one ear broken away, and with the two ears and extremely conventionalized face it occupies projecting part of wood, and part below (4), which, in large staff gods, would be occupied by a secondary figure is here elaborated into elements representing a body and limbs for head above; thus below the face are two horizontal ridges representing upper limbs with a rounded knob for the body between them; below these again are two horizontal elements on either side, representing lower limbs, but one is cut through at knee joint. Two last figures (5, 6) are in projecting profile looking away from each other and with gap between; first (5) has single upper limb element beneath it followed by full-face figure (7) with large ears bounding conventional face and bent elements for lower limbs; second (6) has conventional profile face with upper limb flexed to hand with three fingers and a bent lower limb. First series (1, 1) occupies length of 140 mm. and second (5, 6) a length of 98 mm. Phallus (8) is 110 mm. long, d, under surface, lower end: 1, notched part under first set of figures; 2, notched part under second set of figures; 3, phallus, e, head and upper figure series: head (1) conforms to previous pattern with serrated arm (2) and three-fingered hand, but the figure (3) above arm is in profile instead of full face; next projection (4) is notched on each border without any carved figure; height of head (1) 118 mm., maximum length 68 mm., width 33 mm.; staff (5) is 39 mm. in diameter, f, lower end of e with phallus amputated: first figure (1) in projecting profile with lower part (2) notched instead of secondary figure; last pair (3, 4) in profile facing away; one (3) consists of head only with part below it notched; the other (4) has normal head with upper limb and three-fingered hand, and horizontal ridge for lower limb. Total length of staff god without phallus 1,025 mm.

page 325

The middle staff section is very long in proportion to the other two carved sections in order to provide space for the cloth wrapping. Of the series of eight, only two in the Oldman collection retained the cloth wrapping (figs. 198, 200).

Figure 200.—Small staff god (Oldman coll., 428): total length 29 inches; head length 4.15 inches; head width 1.75 inches; length upper head section 9.25 inches; length lower section 6.75 inches; notches front and back (1) in upper section and in lower section (2) take place of small human figures; roll of tapa (3), decorated with black, wrapped around middle staff section; rough carving.

Figure 200.—Small staff god (Oldman coll., 428): total length 29 inches; head length 4.15 inches; head width 1.75 inches; length upper head section 9.25 inches; length lower section 6.75 inches; notches front and back (1) in upper section and in lower section (2) take place of small human figures; roll of tapa (3), decorated with black, wrapped around middle staff section; rough carving.

The lower carved section has an upper and lower row of small figures, except in two specimens in which one full-face figure occupies the whole upper and lower space. In all except one, the last figure is a single profile figure facing toward the phallic terminal. The one exception to the end profile figure is a specimen which has no small figures at all (fig. 200). In the upper row the profile figures greatly outnumber the full-face figures. In seven gods, there are 22 profiles to 3 full faces. They are usually arranged in two groups with a short length of bare staff between them. Adjacent profile figures are carved back to back. The lower row is always full-face of the non-perforated variety, except in three panels in which the raised zigzag motif is used.

Double-figure staff god. The third type is composed of three sections but differs from the preceding forms in that the upper section is flat and wide instead of a narrow plank set edgewise. The single large head in profile is replaced by two fully formed figures with flexed arms and fully shaped flexed legs. The small figures are in single rows looking outward from either side edge of the expanded upper portion. These figures are full face with well-formed head, body, and limbs. In the wide space between the lateral edges, page 326there is a mesial row of three heads in profile flanked by triangular perforations on either side. The perforations do not assist in defining any figures, but they seem to be a technical carry over from the large staff gods.

The middle section is a rounded staff that was probably covered with a cloth wrapping. The lower section is ornamented by two pairs of back to back profile figures and it ends in a phallus. The specimen in the British Museum seems to be unique (fig. 201).

Figure 201.—Double figure staff god (British Mus., 1910-6-9-1). a, full figure, front view: 1, two top figures; 2, wide part with lateral rows of small figures; 3, staff, 4, lower small figures; 5, phallic terminal, b, upper part, front: two full-face figures (1, 1) with typical eyes and mouth, flexed arms and three-fingered hands resting on abdomen; body and flexed legs; lateral rows of small figures (2, 2) with tops of heads and ears showing, six in each set; middle expanded part (3) with a row of five perforated triangles on either side of middle line occupied by three profile heads (4) of which vertex and two ears show; expanded part narrows into staff (5). c, side view: upper figure (1) with typical ear and jaw line, no nose, flexed arms and legs; full-face small figures (2) with large ears, less elaborate eyes and mouth, flexed arms resting on abdomen, and two distinct lower limbs; one of figures (3) broken, others of same type, d, back: upper figures (1, 1) showing projecting ears, straight edge (2) between shoulders, perforations (3) separating upper arms from body; expanded portion (4) showing triangular perforations; lower ends of legs (6) showing, e, lower series of profile figures in two pairs facing away, with typical profile heads, flexed arm elements, and acutely flexed lower limbs; straight cut down on back of each figure; staff thinned down and two slender connections between last pair of figures; last figure (1) has three-fingered hand (2); phallic terminal (3).

Figure 201.—Double figure staff god (British Mus., 1910-6-9-1). a, full figure, front view: 1, two top figures; 2, wide part with lateral rows of small figures; 3, staff, 4, lower small figures; 5, phallic terminal, b, upper part, front: two full-face figures (1, 1) with typical eyes and mouth, flexed arms and three-fingered hands resting on abdomen; body and flexed legs; lateral rows of small figures (2, 2) with tops of heads and ears showing, six in each set; middle expanded part (3) with a row of five perforated triangles on either side of middle line occupied by three profile heads (4) of which vertex and two ears show; expanded part narrows into staff (5). c, side view: upper figure (1) with typical ear and jaw line, no nose, flexed arms and legs; full-face small figures (2) with large ears, less elaborate eyes and mouth, flexed arms resting on abdomen, and two distinct lower limbs; one of figures (3) broken, others of same type, d, back: upper figures (1, 1) showing projecting ears, straight edge (2) between shoulders, perforations (3) separating upper arms from body; expanded portion (4) showing triangular perforations; lower ends of legs (6) showing, e, lower series of profile figures in two pairs facing away, with typical profile heads, flexed arm elements, and acutely flexed lower limbs; straight cut down on back of each figure; staff thinned down and two slender connections between last pair of figures; last figure (1) has three-fingered hand (2); phallic terminal (3).

page 327
Summary of Staff Gods

The order and arrangement of the multiple small figures in the staff gods follow a similar pattern but a good deal of variation exists in their anatomical details (fig. 202).

Large ears are prominent in Rarotongan images and staff gods, a feature which persists in the smaller figures of the staff gods. In the profile figures, the ears are readily distinguished because the eyes and mouth are well defined. The head parts of the profile figures are similar in the three types of staff gods. They all have recognizable eyes, mouth, and ears though the incised curves around the eyeball may diminish to one above and one below and the mouth may be defined by one notch. In one of the smaller staff gods (fig. 199, f, 4), the mouth is omitted. In the large staff gods, the area given to the profile figure is so short that the body and limb elements cannot all be included below the head; but, owing to the inclination upward and backward of the jaw line, it has been possible to include an upper limb element to balance the figure (fig. 202, f, g, h). In one figure where the jaw line is horizontal, the upper limb element had to be omitted (fig. 202, i). In all the profile figures of the large staff gods, the phallus is under the chin of the profile head. This is ridiculous anatomy, but it must be remembered that a horizontal slit between the upper adjacent figures served a useful purpose in fastening feather decorations. The phallus was a common motif in carving, hence it was natural enough to carve the bar above the slit into an art form to serve as a cleat for fastening feathers in an appropriate position. Function, not anatomical position, decided the form of motif selected.

In the large staff gods, the profile heads are above highly conventionalized figures in full face (fig. 202, f-i) with the limbs partly defined by triangular perforations. In the small staff gods, the first full-face figure of the lower row may occur under the first profile figure of the upper series (fig. 199, b,5), but the triangles defining the body and limbs do not perforate the wood. The other full-face figures of the lower row are usually simple in form without the defining triangles (fig. 199, c,3, 3, 7). Again in the double-figure staff god and some of the figures in the smaller staff gods, the area usually devoted to full-face figures of the lower row is given to limb elements for upper profile heads (fig. 199, c,6 and f,4).

The full-face, small figures are so highly conventionalized that their anatomical parts would not be recognized were it not that the key is supplied by the profile figures on the same god. The key is the consistency in the shape and the position of the ear motif. In the upper figures in figure 202, b, c, d, there is no difficulty in recognizing as ears the slanting, projecting, and elliptical motif hollowed out in the middle, because of their resemblance to the definite ears in the profile figures. The notched part between the ears must therefore page 328
Figure 202.—Variation in small figures on staff gods; selected panels with upper and lower figures: a-e, full face above and below; f-i, profile above and full face below. a, panel in rough, showing preliminary stage of defining raised areas by removing intermediate wood: 1, ears; 2, face; 3, upper limbs; 4, body; 5, lower limbs, b, technique similar to that of a, treatment of raised areas: ear areas are treated by hollowing out middle (1) and leaving a raised circumferential rim; face area has two oblique grooves (2) on either side and two lower horizontal grooves (3, 4) in which lower groove is from either side, which leaves small points above and below; upper limb areas (5) are merely trimmed and body area (6) is raised knob left between grooves made around it; lower limb areas are trimmed by angular grooves (7) on either side, while lower edge is cut with two angular notches. The above is the general treatment of full-face figures, though variations may occur in the number or direction of the grooves made. Lower figure (8) follows pattern of upper one, but has only one transverse face groove instead of two as in a. c, in face element lower horizontal groove (1) does not cut right across area, and this gives appearance of a mouth; in lower figure (2), upper limb element is omitted, d, a fullface figure above arm (1) of large head is not notched on lower edge of legs (2). e, upper figure (1) without arm element; lower figure (2) without serrations on lower limbs, f, usual profile figure with full-face figure below; only one complete horizontal cut (1) on face element, g, lower face, in which upper pairs of oblique grooves meet and separate a small upper knob (1); no complete horizontal face cut, so lower knob (2) not separated, h, upper arm element (1) of upper figure not grooved as in / and g; lower face area separated into four parts through cuts from sides meeting, i, limb element of profile face omitted; lower face area (1) treated with side notches; this pair is second set from large head, arm (2) of which reaches neighboring ear (3) of lower figure, hence figure has only one arm (4); large arm (2), however, does not encroach on leg area and hence lower figure has a leg on each side (5, 6). j, profile panel on small staff god, showing lower figure defined by triangular depressions not perforated as in large staff gods, k, profile panel with lower figure not defined by triangles and upper (1) and lower (2) limbs represented by horizontal elements; no body element. l, full-face panel with notched face (1) occupying upper part and body and limbs filling lower half; one leg (2) has cut running right across and severing two elements at knee. m, profile panel with head occupying upper half and well-defined upper limb and poorly executed lower limb with notches occupying lower half; last panel bounding phallic terminal (1). n, profile panel with upper projecting head and lower half carved into eleven raised knobs, o, upper profile head and lower half with two vertical notched flanges, p, upper profile head with lower half notched at sides.

Figure 202.—Variation in small figures on staff gods; selected panels with upper and lower figures: a-e, full face above and below; f-i, profile above and full face below. a, panel in rough, showing preliminary stage of defining raised areas by removing intermediate wood: 1, ears; 2, face; 3, upper limbs; 4, body; 5, lower limbs, b, technique similar to that of a, treatment of raised areas: ear areas are treated by hollowing out middle (1) and leaving a raised circumferential rim; face area has two oblique grooves (2) on either side and two lower horizontal grooves (3, 4) in which lower groove is from either side, which leaves small points above and below; upper limb areas (5) are merely trimmed and body area (6) is raised knob left between grooves made around it; lower limb areas are trimmed by angular grooves (7) on either side, while lower edge is cut with two angular notches. The above is the general treatment of full-face figures, though variations may occur in the number or direction of the grooves made. Lower figure (8) follows pattern of upper one, but has only one transverse face groove instead of two as in a. c, in face element lower horizontal groove (1) does not cut right across area, and this gives appearance of a mouth; in lower figure (2), upper limb element is omitted, d, a fullface figure above arm (1) of large head is not notched on lower edge of legs (2). e, upper figure (1) without arm element; lower figure (2) without serrations on lower limbs, f, usual profile figure with full-face figure below; only one complete horizontal cut (1) on face element, g, lower face, in which upper pairs of oblique grooves meet and separate a small upper knob (1); no complete horizontal face cut, so lower knob (2) not separated, h, upper arm element (1) of upper figure not grooved as in / and g; lower face area separated into four parts through cuts from sides meeting, i, limb element of profile face omitted; lower face area (1) treated with side notches; this pair is second set from large head, arm (2) of which reaches neighboring ear (3) of lower figure, hence figure has only one arm (4); large arm (2), however, does not encroach on leg area and hence lower figure has a leg on each side (5, 6). j, profile panel on small staff god, showing lower figure defined by triangular depressions not perforated as in large staff gods, k, profile panel with lower figure not defined by triangles and upper (1) and lower (2) limbs represented by horizontal elements; no body element. l, full-face panel with notched face (1) occupying upper part and body and limbs filling lower half; one leg (2) has cut running right across and severing two elements at knee. m, profile panel with head occupying upper half and well-defined upper limb and poorly executed lower limb with notches occupying lower half; last panel bounding phallic terminal (1). n, profile panel with upper projecting head and lower half carved into eleven raised knobs, o, upper profile head and lower half with two vertical notched flanges, p, upper profile head with lower half notched at sides.

page 329 represent the face, and the identification of the body and limb elements below the face is made easier. If, however, the lower secondary figures were detached without knowledge of the upper figures, the correct interpretation of the elements would not be so easy. For example, in the lower figure in figure 202, c,2, what we know to be ears could quite excusably be regarded as eyes and the middle face motif might well be regarded as a nose with expanded nostrils with a small mouth below it. I tested the drawing on members of the Museum staff and, without exception, they diagnosed the elements as eyes, nose, and mouth. Thus the inaccurate interpretation fits the actual appearance of the figure better than the true one.

The secondary, lower full-face figures below the profile figures in the large staff gods are formed by a more specialized technique than those below the full-face figures of the upper row. In these, the ears and notched face are similar but the body and limbs are defined by perforated triangles and both hands and feet are further distinguished by two incised lines which divide them into three fingers and three toes. Owing to the limitations of space, there is no neck and the arms branch off from the outer sides of the ears (fig. 202, f-i).

The highly conventionalized full-face figures by themselves would be regarded as examples of "degradation in art", but as they occur in association with well-executed figures on the same object made by the same artist, such a theory is untenable. It must be conceded that the artist was influenced not only by the limitations of space, but that he had a set of values in which secondary parts of his creation purposely did not receive the care devoted to the principal motifs.

page 330
Aitutaki Gods

The study of the material symbols of the gods of Aitutaki must also be based on museum material. Again the richest material is among the "trophies" saved by the London Missionary Society. How they were acquired may be briefly extracted from Williams' reports (81).

After the Tahitian teachers, Papeiha and Vahapata, were stationed in Aitutaki by the London Missionary Society in 1821, they made little progress for 15 months. Then a girl of the Tamatoa ruling family died in spite of all the efforts of the priests to effect a cure through the medium of their own gods. The grandfather of Tamatoa Ariki was so enraged at the lack of answer to ritual that he sent his son to set fire to the temple of his own god. Two neighboring temples also caught fire. An example having been set by the highest native authority, the people followed suit and several districts brought in their gods to the native teachers on the Sabbath following the death of the chiefly maid. During the following week, other districts did likewise. The two teachers were quick to make the most of an opportunity that their 15 months of teaching had not been able to procure. As already stated (p. 308), Papeiha ordered the burning of the temples and the bringing in of all the remaining gods to be forwarded to Raiatea. After the destruction of the temples, Williams (81, p. 63) wrote:

The whole population then came in procession, district after district, the chief and priest leading the way, and the people following them, bearing their rejected idols, which they laid at the teachers' feet, and then received from them in return a few copies of the gospels and elementary books.

Williams (81, p. 92) evidently had the right idea of the value of the gods as museum specimens for he says:

I obtained from the chief of Aitutaki a short account of the relics of idolatry. Twenty-five of these I numbered, and transmitted, with their name and history, to the deputation then at Tahiti; six others were sent to England, and many of them are now in the Missionary Museum.

Williams (81, p. 93) then described a selection to give the reader a general idea of the whole. As any information collected at that time may be valuable, the list of selected objects and their numbers are repeated here in full.

Number 2. An idol named Te Rongo, one of the great deities, called a kai-tangata, or man-eater. The priests of this god were supposed to be inspired by a shark.

Number 8. Tangaroa; the great national god of Aitutaki, and of almost all the adjacent islands. He holds the net with which he catches the spirits of men as they fly from their bodies, and a spear with which he kills them.

Number 15. A rod, with snares at the end, made of the fibers of the cocoa-nut husk, with which the priest caught the spirit of the god. It was used in cases of pregnancy, when the female was ambitious that her child should be a son, and become a famous-warrior. It was also employed in war-time to catch the god by his leg, to secure his. influence on the side of the, party performing the ceremony.

Number 18. Ruanuu; a chief from Raiatea, who, ages ago, sailed in a canoe from that island, and settled at Aitutaki. From him a genealogy is traced. He died at Aitutaki, and was deified, as Te atua taitai tere, or the conductor of fleets. The Raiateans have page 331several interesting traditions connected with Ruatiuu. To this idol was appended an old tattered silk handkerchief, and the foot of a wine-glass; both of which were obtained from Captain Cook's vessel, and dedicated to Ruanuu, "the god or guide of fleets", for conducting that celebrated navigator to their shores.

Number 25. Taau, with his fan, etc.; the god of thunder. When the thunder peals, the natives said that this god was flying, and produced this sound by the flapping of his wings.

Williams uses the Tahitian dialect with the dropped k, so Ruanuu and Taau would be Ruanuku and Takau in the Aitutaki dialect. Ruanuku is a widespread god, and it is evident that Williams mixed up the name of a god with an early voyager either Ru or Ruatapu, both ancestors of the Aitutaki people. The errors are continued in the statement that the silk hankerchief and the wine glass foot were obtained from Captain Cook's vessel and dedicated to Ruanuu for conducting Cook to their shores. Cook touched at Mangaia and Atiu and twice sailed past Hervey Island (Manuae) but he never saw Aitutaki, which was discovered later by Captain Bligh in 1789. In spite of such errors, it is a pity that details of the other 20 objects were not recorded in print.

The six gods that reached the Missionary Museum in London were presumably transferred to the British Museum with other material. I could locate only two gods definitely assigned to Aitutaki, but there were four carved slabs catalogued as idols and attributed to the Hervey Islands. A study of the carving motifs shows that the four slabs are from Aitutaki. Three of these were catalogued from the Missionary collection, but in none of them was the name of the god recorded.

The Aitutaki material located consists of eight carved slabs, two carved slabs with a human motif, and eight objects carved in human form. It is possible that some of the eight human objects are not gods but part of a canoe decoration. However, failing an accurate history, I have included them here for comparative purposes.

Carved Slabs

The carved slabs resemble some of the unu carved ornaments figured by Ellis (25, vol. 2, p. 217) on a picture of a Tahitian temple. Carved unu ornaments were used on the Rarotongan temples, and it is probable that they were used also on Aitutakian temples. Though each of the three slabs in the British Museum was described in the London Missionary Society catalog as "a carved flat club formed idol", it is possible that they were unu and were termed idols by the missionaries because they were used in the temples. They are flat pieces of wood, usually carved on both sides, and narrowed at the lower end into spatulate form.

In the simpler forms, the carving motifs consisted of serrated edges, raised serrated bars, and sunken or pierced triangles. In the better carved forms, page 332sunken crescents, both pierced and raised lozenges, and chevrons combined with triangles or lozenges to form chevron-triangle and chevron-lozenge motifs in addition to the other motifs. Three slabs in the British Museum are shown in figure 203. The serrated edge with alternating sunken triangles
Figure 203.—Carved slabs, Aitutaki. a, probably breadfruit wood; total length, 457 mm.; greatest width, 62 mm.; greatest thickness, 18 mm.: top section (1) 112 mm. long with irregular side notches and mesial hole: middle section (2) with serrated sides and triangular depressions opposite alternate serrations; mesial raised longitudinal panel divided into serrated transverse bars: lower plain section (3) 34 mm. wide and 14 mm. thick at bottom end. (British Mus., L.M.S. coll., 112). b, brown wood, 490 mm. long; top width 93 mm.; 24 mm. thick in middle line and 12 mm. at sides: 1, upper carving of three rows of sunken triangles opposite alternate serrations in serrated edge below; 2, two rows of sunken lozenges with holes pierced through two upper, outer lozenges; 3, rest of carving has sunken triangles along side edges with inner sunken triangles opposite alternate serrated points of edge; inner part composed of serrated bands and sunken triangles; 4, lower section plain with thin white tapa wrapped around lower end by two-ply coir cord. (British Mus., L.M.S., 26). c, surface shiny from handling; length 607 mm.; top width 123 mm.; thickness 30 mm.: 1, upper end notched, four projections with chevron-triangle motifs with opposing chevrons below and knobs of uncut wood between, and sunken crescents below. 2, two notched transverse bars. 3, repetition of top pattern (1). 4, rest of carving consists of transverse serrated bars with a median longitudinal panel carved with chevron-triangle motifs in pairs placed sideways. 5, three transverse edges. 6, uncarved lower end, 61 mm. wide at upper end and 32 mm. below, transverse hole, and single side notches, (British Mus., no number), d, back of b, with coir cord ornamented with transverse tufts of coir fiber, 30 mm. long, caught between plies and with lower end doubled around lower edge of slab and covered by thin tapa lashed with bast cord, e, back of c:1, chevron-triangle motifs serrated at base; 2, transverse serrated bars with middle panel of sunken lozenges; 3, crescent motifs; 4, transverse edges. [Catalog states that remains of feathery plumes were formerly attached at lower end.]

Figure 203.—Carved slabs, Aitutaki. a, probably breadfruit wood; total length, 457 mm.; greatest width, 62 mm.; greatest thickness, 18 mm.: top section (1) 112 mm. long with irregular side notches and mesial hole: middle section (2) with serrated sides and triangular depressions opposite alternate serrations; mesial raised longitudinal panel divided into serrated transverse bars: lower plain section (3) 34 mm. wide and 14 mm. thick at bottom end. (British Mus., L.M.S. coll., 112). b, brown wood, 490 mm. long; top width 93 mm.; 24 mm. thick in middle line and 12 mm. at sides: 1, upper carving of three rows of sunken triangles opposite alternate serrations in serrated edge below; 2, two rows of sunken lozenges with holes pierced through two upper, outer lozenges; 3, rest of carving has sunken triangles along side edges with inner sunken triangles opposite alternate serrated points of edge; inner part composed of serrated bands and sunken triangles; 4, lower section plain with thin white tapa wrapped around lower end by two-ply coir cord. (British Mus., L.M.S., 26). c, surface shiny from handling; length 607 mm.; top width 123 mm.; thickness 30 mm.: 1, upper end notched, four projections with chevron-triangle motifs with opposing chevrons below and knobs of uncut wood between, and sunken crescents below. 2, two notched transverse bars. 3, repetition of top pattern (1). 4, rest of carving consists of transverse serrated bars with a median longitudinal panel carved with chevron-triangle motifs in pairs placed sideways. 5, three transverse edges. 6, uncarved lower end, 61 mm. wide at upper end and 32 mm. below, transverse hole, and single side notches, (British Mus., no number), d, back of b, with coir cord ornamented with transverse tufts of coir fiber, 30 mm. long, caught between plies and with lower end doubled around lower edge of slab and covered by thin tapa lashed with bast cord, e, back of c:1, chevron-triangle motifs serrated at base; 2, transverse serrated bars with middle panel of sunken lozenges; 3, crescent motifs; 4, transverse edges. [Catalog states that remains of feathery plumes were formerly attached at lower end.]

page 333 is shown in figure 203, a, b; the raised serrated bar in 203, a, b, c; the sunken lozenge in 203, b; the chevron-triangle motif in 203, c; and the sunken crescent in 203, c.
Figure 204.—Carved slabs (Lille Mus., after Stolpe): a, made of brown wood, flat and thin, serrated line, sunken triangles near each side edge with apex outward and each triangle opposite alternate teeth in serrated edge. Upper larger trianfles tend to form crescents, base and sides curved. In middle vertical line is row of three sunken lozenges followed by three irregularly placed triangles and finishing off with two mesial lozenges. Carved portion narrows from above down. Lower staff part plain and wider at lower end; around its upper part a thin two-ply twist of coir fiber is wound for several turns, and a long free end through which transverse tufts of fiber have been inserted hangs down. Free cord is also decorated with red feathers not shown in figure, b, made of brown wood: upper end (1) serrated with three deep notches and cut to a lower plane than general surface which commences at serrated line (2); side edges serrated but somewhat irregular at upper part, and lower part has pierced triangles in relation to alternate teeth; vertically placed pairs of compound lozenge-chevron motif (3) cut from two vertical raised bars, showing slight knob (4) between each pair; two transverse raised bars (5) followed by two pairs of lozenge-chevron and two more transverse notched bars; carved portion narrows from above down and further narrows into plain lower staff (6) which is bound with bast (7) and coir fiber cord (8) which supports a twisted cord of coir fiber (9) 1,080 mm. long ornamented with transverse tufts of coir fiber and red and yellow feathers; human front tooth was placed under turns of bast.

Figure 204.—Carved slabs (Lille Mus., after Stolpe): a, made of brown wood, flat and thin, serrated line, sunken triangles near each side edge with apex outward and each triangle opposite alternate teeth in serrated edge. Upper larger trianfles tend to form crescents, base and sides curved. In middle vertical line is row of three sunken lozenges followed by three irregularly placed triangles and finishing off with two mesial lozenges. Carved portion narrows from above down. Lower staff part plain and wider at lower end; around its upper part a thin two-ply twist of coir fiber is wound for several turns, and a long free end through which transverse tufts of fiber have been inserted hangs down. Free cord is also decorated with red feathers not shown in figure, b, made of brown wood: upper end (1) serrated with three deep notches and cut to a lower plane than general surface which commences at serrated line (2); side edges serrated but somewhat irregular at upper part, and lower part has pierced triangles in relation to alternate teeth; vertically placed pairs of compound lozenge-chevron motif (3) cut from two vertical raised bars, showing slight knob (4) between each pair; two transverse raised bars (5) followed by two pairs of lozenge-chevron and two more transverse notched bars; carved portion narrows from above down and further narrows into plain lower staff (6) which is bound with bast (7) and coir fiber cord (8) which supports a twisted cord of coir fiber (9) 1,080 mm. long ornamented with transverse tufts of coir fiber and red and yellow feathers; human front tooth was placed under turns of bast.

page 334

Two carved slabs; in the Lille Museum have been described by Stolpe (64; pp. 41-43, figs. 47, 48) and are here reproduced in figure 204.

In the Lille slabs, the serrated edges with alternating sunken triangles and the raised zigzag bar in one slab (fig. 204, b) are characteristic of Aitutaki. Another motif appears in the lozenge-chevron (fig. 204, b,3) which is a further elaboration of the triangle-chevron motif (fig. 203, c). The coir cord ornament with transverse tufts of coir fiber occurs in both slabs and is identical in technique with that in figure 203, d. Stolpe (64, p. 41) states that the origin of the Lille figures is unknown but that they belonged to a large collection
Figure 205.—Aitutaki carved slabs (Oldman coll.: a, no. 431; b, no.429). a, height, 23 inches; width, 4 inches; thickness, 0.75 inch. Side edges serrated, but sunken triangles do not always coincide with alternate teeth; raised middle panel with short horizontal rows of serrated bars characteristic of Aitutaki; two lowest bars not serrated; carved on one side only; lower end flared at sides, b, dark heavy wood; height, 20.75 inches; width, 5.75 inches; thickness, 1.25 inches. Top widely serrated with pierced lozenges below; sides serrated but sunken triangles opposite each tooth; upper part has four rows of fine chevron-lozenge figures each row reversed to preceding row; lower middle panel has serrated edge (1) with sunken triangles (2) opposite alternate teeth and nine transverse rows of notched bars; two holes in middle at sides for attachment of feathers; lower end widens from above down.

Figure 205.—Aitutaki carved slabs (Oldman coll.: a, no. 431; b, no.429). a, height, 23 inches; width, 4 inches; thickness, 0.75 inch. Side edges serrated, but sunken triangles do not always coincide with alternate teeth; raised middle panel with short horizontal rows of serrated bars characteristic of Aitutaki; two lowest bars not serrated; carved on one side only; lower end flared at sides, b, dark heavy wood; height, 20.75 inches; width, 5.75 inches; thickness, 1.25 inches. Top widely serrated with pierced lozenges below; sides serrated but sunken triangles opposite each tooth; upper part has four rows of fine chevron-lozenge figures each row reversed to preceding row; lower middle panel has serrated edge (1) with sunken triangles (2) opposite alternate teeth and nine transverse rows of notched bars; two holes in middle at sides for attachment of feathers; lower end widens from above down.

page 335
Figure 206.—Carved slab with arches (British Mus., L.M.S. 28). a, front: upper wide section (1) 180 mm. wide and 120 mm. deep, including three upper projections (4) obscured by feathers; upper edge below projections serrated; five sets of rectangular perforations in three vertical pairs and outer singles with vertical and horizontal grooves between; spaces between vertical pairs and outer singles occupied by four arches formed of two pillars with outward projections carved into lozenge-chevron form; two middle arches broken but lower carved cleat (5) of one remains and two outer arches (6, 6) are unbroken; outer single holes bounded laterally by arches (7); middle section (2) 68 mm. wide at junction and 58 mm. wide at lower end; width with lowest arches 79 mm., depth in middle line 233 mm.; middle vertical raised panel (8) 17 mm. wide, uncarved, with forward projecting arches, of which two upper are broken and three lower intact with feathers attached; on each side of panel, 14 horizontal notched bars (9) on same level as panel; sides bounded by five arches (10) projecting laterally. Bottom staff section (3) plain; length 108 mm., top width 31 mm., bottom width 36 mm. b, back, feather decoration not shown: upper section (1) with three upper projections (4) with constricted neck (5) cut with two or three transverse grooves, middle expanded portion (6) with large circular hole, and an upper crescentic cap (7) pierced by a mesial crescentic hole with a small hole on either side; upper edge below projections unnotched; three vertical grooves take place of front arches between rectangular holes, and horizontal grooves occur between outer pierced holes; relation of outer single holes to lateral arches readily seen; in middle section (2) mesial vertical panel (8) and parts on either side have notched bars; sets of five lateral arches (10) with pillar projections are evident, one broken, c, side view of arch with pillars (1), arch (2), and outward projections (3). d, front view of arch showing pillar projection (3), formed with lozenge (4) and combined chevron (5).

Figure 206.—Carved slab with arches (British Mus., L.M.S. 28). a, front: upper wide section (1) 180 mm. wide and 120 mm. deep, including three upper projections (4) obscured by feathers; upper edge below projections serrated; five sets of rectangular perforations in three vertical pairs and outer singles with vertical and horizontal grooves between; spaces between vertical pairs and outer singles occupied by four arches formed of two pillars with outward projections carved into lozenge-chevron form; two middle arches broken but lower carved cleat (5) of one remains and two outer arches (6, 6) are unbroken; outer single holes bounded laterally by arches (7); middle section (2) 68 mm. wide at junction and 58 mm. wide at lower end; width with lowest arches 79 mm., depth in middle line 233 mm.; middle vertical raised panel (8) 17 mm. wide, uncarved, with forward projecting arches, of which two upper are broken and three lower intact with feathers attached; on each side of panel, 14 horizontal notched bars (9) on same level as panel; sides bounded by five arches (10) projecting laterally. Bottom staff section (3) plain; length 108 mm., top width 31 mm., bottom width 36 mm. b, back, feather decoration not shown: upper section (1) with three upper projections (4) with constricted neck (5) cut with two or three transverse grooves, middle expanded portion (6) with large circular hole, and an upper crescentic cap (7) pierced by a mesial crescentic hole with a small hole on either side; upper edge below projections unnotched; three vertical grooves take place of front arches between rectangular holes, and horizontal grooves occur between outer pierced holes; relation of outer single holes to lateral arches readily seen; in middle section (2) mesial vertical panel (8) and parts on either side have notched bars; sets of five lateral arches (10) with pillar projections are evident, one broken, c, side view of arch with pillars (1), arch (2), and outward projections (3). d, front view of arch showing pillar projection (3), formed with lozenge (4) and combined chevron (5).

page 336of sacred objects obviously from the islands south of Tahiti. He correctly assigned them to the Rarotongan group, and there can be little doubt that they are from Aitutaki.

Two carved slabs of a simple type in the Oldman collection are shown in figure 205. An old label on figure 205, a says, "Brought home probably in Captain Cook's ship. Belonged to Kirkpatrick, physician to Geo. III."

The most elaborately carved slab of the series brings in new features in the presence of pierced dome-shaped protuberances at the upper end and arches with cleats for the attachment of feather decorations (pl. 14, A). It was figured by Edge-Partington (24, 1-23-1). (See figure 206.)

The feathers used are red parakeet, split black man-of-war hawk, and small white and long white tropic bird feathers. The red feathers are tied in small bunches to a single cord holder and the split black feathers are bound into bunches with the quills seized to form a loop. These two forms are attached to the upper dome-shaped protuberances. All the feathers without holders or loops are attached by a fine thread to the arches at the sides and in the front mesial row. They must also have been attached to the arches on the upper part of the slab, but they have disappeared because some of the arches are broken. The technique of the feather work is shown in figure 207.

Figure 207.—Feather decoration of carved slab, a, small red parakeet feathers (1) attached to sennit cord holder (2), about 40 mm. long and 1 mm. thick; feathers in separate bunches bound by single fiber and lashed (3) to holder without doubling back end (4) of holder, b, black split feathers (1) with long quill ends laid together and seized with single coir fiber and seized end bent to form loop (2) and lashed (3)1 to standing part, c, black feather bunches (1, 1) and red feathers on cord holders (2) tied to sides of dome projections (3) through large hole by thin thread (4) which passes across hole to other side, d, top view of tropic bird (1), split black (2), and small red feathers (3) attached by transverse turns (4) around arch and figure-of-eight turns (5) around lower cleat of arch, e, side view of d, showing transverse turns (4) around arch and figure-of-eight turns (5) around lower cleat, f, shows arrangement of upper set (1) of red feathers and lower set (2) of same on either side of lashing, g, shows feathers (1) attached to arch (2) and lower cleat (3) by thread (4) which passes down to next arch (5) for a similar attachment of more feathers.

Figure 207.—Feather decoration of carved slab, a, small red parakeet feathers (1) attached to sennit cord holder (2), about 40 mm. long and 1 mm. thick; feathers in separate bunches bound by single fiber and lashed (3) to holder without doubling back end (4) of holder, b, black split feathers (1) with long quill ends laid together and seized with single coir fiber and seized end bent to form loop (2) and lashed (3)1 to standing part, c, black feather bunches (1, 1) and red feathers on cord holders (2) tied to sides of dome projections (3) through large hole by thin thread (4) which passes across hole to other side, d, top view of tropic bird (1), split black (2), and small red feathers (3) attached by transverse turns (4) around arch and figure-of-eight turns (5) around lower cleat of arch, e, side view of d, showing transverse turns (4) around arch and figure-of-eight turns (5) around lower cleat, f, shows arrangement of upper set (1) of red feathers and lower set (2) of same on either side of lashing, g, shows feathers (1) attached to arch (2) and lower cleat (3) by thread (4) which passes down to next arch (5) for a similar attachment of more feathers.

page 337
Carved Slabs with Human Motifs

Two carved slabs with the typical motifs of Aitutaki are elaborated by the addition of human motifs at the upper end. One has a human head and the other three full human figures (fig. 208). The Oldman specimen (fig. 208, a) is believed to have been taken home by John Williams. The specimen figured by Ellis (fig. 208, b) has not been located in any museum or collection. Concerning it, Ellis (25, vol. 2, pp. 220, 221) writes: "No. 2 is an image of Tebuakina, three sons of Rongo, a principal deity in the Harvey Islands." The "statement does not make sense, for the single name of "Tebuakina" (there is no b in the dialect) cannot denote the three sons of Rongo whose individual names were Rangi, Mokoiro, and Akatauira. Furthermore, the slab is undoubtedly from Aitutaki, whereas the myth attributing three sons to Rongo is from Mangaia. The correct name of the god must remain in doubt. The shape of the longer figures conforms to that of some of the Aitutaki gods.

Wooden Images

The wooden images of Aitutaki differ markedly in technique from those of Rarotonga. They are much more roughly made, with the eyes and mouth simply incised elliptical depressions or, at most, two curved cuts enclosing an elliptical surface. The nose is always indicated by an edge with the lower end defined by a straight cut or by a narrow raised ridge. Ears are absent in the small figures, whereas they were a constant feature in the smallest figures of Rarotongan craftsmanship. The arms are very slender and usually end in a small hand with three fingers. The legs are also slender, so slender that in some carvings the body is supported by a vertical, mesial column provided by leaving some of the wood between the legs and at the back. The thighs are often short and slant upward to form an acute angle with the slender legs. The shoulders are square across the back, the chest slopes downward and forward in a flat plane, the breasts are not indicated, the abdomen protrudes with a protruding navel usually low down and apt to be mistaken for the male sex organ. The brows are marked by curved cuts separating the higher plane of the forehead from the lower plane of the face.

The simplest image is a head set on a long column. Its identification is correct, for it was figured by Williams (81, p. 55), who describes his acquisition of it in a village in Aitutaki (81, p. 54):

While walking through the settlement, we saw two grim looking gods in. a more dishonorable situation than they had been wont to occupy, for they were sustaining upon their heads the whole weight of the roof of a cooking house. Wishing to make them more useful, we offered to purchase them from their former worshipper. He instantly propped up: the house, took out the idols, and threw them down; and, while they were prostrate on the ground, he gave them a kick, saying, "There—your reign is at an end."

page 338
Figure 208.—Carved slabs with human motifs: a, height 26 inches; width 3.25 inches; greatest thickness 1.75 inches; slab with sunken notches (1) along side edges with crescents (2) on inner side of alternate teeth and triangles (3) on inner side of crescents; raised serrated bars (4) spaced down middle line; lower end with three transverse slits (5) and then plain; few turns of cord (6) evidently for fastening a coir ornament which has been lost; head (7) with straight brows, narrow nose, and eyes and mouth formed by elliptical pits. (Oldman coll., 430.) b, two large human figures at top of slab have curved lines over forehead, slanting flat busts, and evidently slender arms; a small, human figure in middle line below them; slab has serrated edges with sunken triangles opposite alternate teeth (1, 2) and serrated side edges with similar relationship of sunken triangles (3); a middle raised panel has been cut into transverse serrated bars (4) and above, there is a mesial pierced lozenge (5); lower part (6) evidently corresponds to a staff covered with some form of decoration. (After Ellis, 25, vol. 2, frontispiece 2.)

Figure 208.—Carved slabs with human motifs: a, height 26 inches; width 3.25 inches; greatest thickness 1.75 inches; slab with sunken notches (1) along side edges with crescents (2) on inner side of alternate teeth and triangles (3) on inner side of crescents; raised serrated bars (4) spaced down middle line; lower end with three transverse slits (5) and then plain; few turns of cord (6) evidently for fastening a coir ornament which has been lost; head (7) with straight brows, narrow nose, and eyes and mouth formed by elliptical pits. (Oldman coll., 430.) b, two large human figures at top of slab have curved lines over forehead, slanting flat busts, and evidently slender arms; a small, human figure in middle line below them; slab has serrated edges with sunken triangles opposite alternate teeth (1, 2) and serrated side edges with similar relationship of sunken triangles (3); a middle raised panel has been cut into transverse serrated bars (4) and above, there is a mesial pierced lozenge (5); lower part (6) evidently corresponds to a staff covered with some form of decoration. (After Ellis, 25, vol. 2, frontispiece 2.)

page 339

Unfortunately neither name nor history has been recorded. The god was also figured by Edge-Partington (24, I-21-4) and it is reproduced here in figure 209.

Figure 209.—Aitutaki god (British Mus., L.M.S., 95). a, front view: light wood, poor carving, total height 1,282 mm.; head length 148 mm.; head width 168 mm.; lower end 180 mm. by 145 mm.; forehead and nose on same plane with wood cut away to form lower plane for eyes and face, eyes and mouth formed by simple excisions, no ears, chin pointed, simple bust, without arms, column cut in transverse levels as shown. b, side view: showing transverse cuts coming around to sides but not reaching back, square cut across upper shoulder level continued around back.

Figure 209.—Aitutaki god (British Mus., L.M.S., 95). a, front view: light wood, poor carving, total height 1,282 mm.; head length 148 mm.; head width 168 mm.; lower end 180 mm. by 145 mm.; forehead and nose on same plane with wood cut away to form lower plane for eyes and face, eyes and mouth formed by simple excisions, no ears, chin pointed, simple bust, without arms, column cut in transverse levels as shown. b, side view: showing transverse cuts coming around to sides but not reaching back, square cut across upper shoulder level continued around back.

The second image, also in the British Museum, is labeled "Hervey Islands. Figure of a goddess carved in the style of Aitutaki." This image forms the key to the identification of some of the artifacts discussed in the following pages. (See figure 210 and plate 13, A, B.)

A painted image in the Munich Museum, labeled "Gotze aus Neu Seeland", was later attributed to the Marquesas. Though some features, such as the form of eyes and mouth, differ from the preceding key image, there is little doubt that the true locality is Aitutaki (fig. 211). The serrated line with sunken triangles opposite alternate teeth is characteristic of the Aitutaki carved slabs. Stolpe (64, p. 41, fig. 46) figured this feature and suggested that it—as well as the shape of the ears, mouth, and eyes—pointed to the Cook page 340
Figure 210.—Aitutaki goddess (British Mus., L.M.S., no number), a, front: total height including pedestal 525 mm.; head height 162 mm.; pedestal height 60 mm.;shoulder width 155 mm.; hip width 171 mm.; forehead with curved transverse line, with similar line below for eyebrows; nose formed by vertical mesial edge due to meeting of inclined planes of cheeks and bounded below by transverse cut; eyes and mouth formed by simple elliptical excisions; hands without fingers clasped on body; rectangular protruding navel; two legs join pedestal; separate vertical mesial support (1) present; curved excisions and zigzag line carved on front of pedestal, b, side view: showing poorly shaped ear with pierced hole, square shoulder, upper arm attached to body throughout, forearm flexed, protruding abdomen and navel, acutely flexed thighs, foot and toes roughly shown on side of pedestal, c, back: shows perforation of ears, square cut across shoulders, mesial gluteal groove, and mesial support (1).

Figure 210.—Aitutaki goddess (British Mus., L.M.S., no number), a, front: total height including pedestal 525 mm.; head height 162 mm.; pedestal height 60 mm.;shoulder width 155 mm.; hip width 171 mm.; forehead with curved transverse line, with similar line below for eyebrows; nose formed by vertical mesial edge due to meeting of inclined planes of cheeks and bounded below by transverse cut; eyes and mouth formed by simple elliptical excisions; hands without fingers clasped on body; rectangular protruding navel; two legs join pedestal; separate vertical mesial support (1) present; curved excisions and zigzag line carved on front of pedestal, b, side view: showing poorly shaped ear with pierced hole, square shoulder, upper arm attached to body throughout, forearm flexed, protruding abdomen and navel, acutely flexed thighs, foot and toes roughly shown on side of pedestal, c, back: shows perforation of ears, square cut across shoulders, mesial gluteal groove, and mesial support (1).

Islands especially to Rarotonga. The prominent pierced ears are certainly reminiscent of Rarotonga, but the eyes are not flanged and the mouth lacks the extra notches that show up so plainly in profile. The slender arms with the upper arm adhering to the body and the small hand with three fingers are totally different from Rarotongan technique. The large lozenge and triangle motifs painted in rows on the back of the image are similar to the carving on an Aitutaki adz (fig. 106, c). Stolpe (64, pp. 41, 42) also drew attention to the resemblance between the serrated line and triangles to a similar technique in the Lille carved slabs which I have attributed to Aitutaki. He regarded this motif as representing human figures fused together, but I regard it as a simple geometrical combination in which any resemblance to degradation from the human form is accidental.
page 341
Figure 211.—Painted image (Munich Mus., 190). a, total height about 404 mm.; well-defined eyebrows meet upper end of nose, eyes formed by curved elliptical incisions, nose is low ridge wider at lower end, mouth formed by two small curved incisions, ears prominent, body rectangular with square, pointed shoulders, rectangular raised nipples, thin forearm lying horizontally across abdomen and terminating in three, pointed fingers, protruding round navel, female sex organ, straight legs with solid feet. b, back: prominent ears show pierced hole through back, rectangular back flat, right-angled shoulders with straight horizontal line across back, gluteal folds marked, backward projection of feet. c, side view: oval ear with groove of helix, sharp angular chin, low nose, upper arm formed by narrow raised ridge attached to body throughout, carved zigzag line (1) runs vertically with sunken triangles corresponding to alternate serrations, an important diagnostic feature.

Figure 211.—Painted image (Munich Mus., 190). a, total height about 404 mm.; well-defined eyebrows meet upper end of nose, eyes formed by curved elliptical incisions, nose is low ridge wider at lower end, mouth formed by two small curved incisions, ears prominent, body rectangular with square, pointed shoulders, rectangular raised nipples, thin forearm lying horizontally across abdomen and terminating in three, pointed fingers, protruding round navel, female sex organ, straight legs with solid feet. b, back: prominent ears show pierced hole through back, rectangular back flat, right-angled shoulders with straight horizontal line across back, gluteal folds marked, backward projection of feet. c, side view: oval ear with groove of helix, sharp angular chin, low nose, upper arm formed by narrow raised ridge attached to body throughout, carved zigzag line (1) runs vertically with sunken triangles corresponding to alternate serrations, an important diagnostic feature.

A figure in the Hunterian Museum, Glasgow, has some features in common with the preceding image. The eyes and mouth are more elaborate, the arms are slender with three-fingered hands resting on the abdomen, the abdomen protrudes with a low protruding navel, and the legs are but slightly flexed. It differs in that the upper arms are free of the body and the legs are more slender. The slender legs are reinforced by a mesial supporting column similar to that in figure 210. It also has two raised notched bars across the breast, a technique common in Aitutaki. The image was figured by Mackenzie (49, plate opposite page 288, no. 1), and the legend states that it was collected on Captain Cook's voyage and attributes it to New Zealand. It could not page 342possibly have been collected on one of Cook's voyages as Cook never visited Aitutaki, and it never saw New Zealand. For reasons stated above, it is identified as Aitutakian. (See figure 212.) The column which supports the figure seems to have been cut off at the lower end and undoubtedly formed a part of a larger object, perhaps a canoe.

Figure 212.—Aitutaki image (Hunterian Mus., Glasgow, E.395); total height with support, 16 inches. a, front: head pointed with sharp chin; eyes elliptical, formed by two curved incisions and also upper and lower lids; raised nose, narrow; mouth with upper and lower lips denned; small unpierced ears; square shoulders; slender arms with upper arms (1) free of body and small three-fingered hands; upper serrated edge and lower serrated bar on chest; protruding abdomen with low navel (2); high hips (3) with slender slightly flexed legs; forward projecting feet partly broken; figure mounted on column (4) which narrows above to form mesial support (5) to body; main column incised with crescent motifs, b, side: with same numbering as a; note backward piojection of gluteal region (6) and vertical row of rectangular knobs on main column.

Figure 212.—Aitutaki image (Hunterian Mus., Glasgow, E.395); total height with support, 16 inches. a, front: head pointed with sharp chin; eyes elliptical, formed by two curved incisions and also upper and lower lids; raised nose, narrow; mouth with upper and lower lips denned; small unpierced ears; square shoulders; slender arms with upper arms (1) free of body and small three-fingered hands; upper serrated edge and lower serrated bar on chest; protruding abdomen with low navel (2); high hips (3) with slender slightly flexed legs; forward projecting feet partly broken; figure mounted on column (4) which narrows above to form mesial support (5) to body; main column incised with crescent motifs, b, side: with same numbering as a; note backward piojection of gluteal region (6) and vertical row of rectangular knobs on main column.

Of two carved objects in the Oldman collection, one has four and one three human figures. The object with four figures was figured by Ellis (25, vol. 2, frontispiece, no. 5), who thus describes it (25, vol. 2, p. 220): "No. 5 is Terongo, one of the principal gods, and his three sons." Rongo, the principal god of Mangaia, did have three sons, but I doubt whether he had the prefix Te in Aitutaki, or whether Aitutakian theology gave him three sons. The slab god with three figures (fig. 208, b) was stated by Ellis to represent three sons page 343of Rongo, not Terongo. One wonders whether Ellis named the Aitutakian figures himself from a knowledge of Mangaian myth. The figures have characteristic Aitutakian heads, bodies, and limbs and each figure has the mesial support shown in figures 210, 212. Both artifacts were figured by Oldman (54, vol. 47, pl. 4). (See figure 213.)

Figure 213.—Multiple human figures (Oldman coll.; a, b,436; c, d,437). a, front: top figure (1) has characteristic brow cut, a small knob on vertex, and a more elaborate mouth; second (2) has more elaborate eyes, but two lower figures (3, 4) have simple form of eyes, nose, and mouth of type (fig. 210); all have slender arms with three-fingered hands, low protruding abdomen, and slender legs; mesial support (5) in all; flower end divides into two legs (6). Dimensions: height, 34.5 inches; top figure, 10.5 inches, b, side view: shows sharp chin, square shoulders with straight cut across back, slender upper arms not separated from body, low protruding abdomen, and fairly acute flexion of thighs. Dimensions: height, 27.5 inches; top figure, 8.25 inches. c, front: all figures have characteristic simple eye, nose, and mouth formation and sharp chins; lowest (3) has characteristic deep cut brow formation; upper two (1, 2) have small three-fingered hands, but in lowest, hands are not defined; upper two have mesial support (5) to body, but lowest ends without legs and its body is supported directly by bifurcated lower end (6). d, side view: shows square shoulders, slender upper arms, low abdomen, and slender legs.

Figure 213.—Multiple human figures (Oldman coll.; a, b,436; c, d,437). a, front: top figure (1) has characteristic brow cut, a small knob on vertex, and a more elaborate mouth; second (2) has more elaborate eyes, but two lower figures (3, 4) have simple form of eyes, nose, and mouth of type (fig. 210); all have slender arms with three-fingered hands, low protruding abdomen, and slender legs; mesial support (5) in all; flower end divides into two legs (6). Dimensions: height, 34.5 inches; top figure, 10.5 inches, b, side view: shows sharp chin, square shoulders with straight cut across back, slender upper arms not separated from body, low protruding abdomen, and fairly acute flexion of thighs. Dimensions: height, 27.5 inches; top figure, 8.25 inches. c, front: all figures have characteristic simple eye, nose, and mouth formation and sharp chins; lowest (3) has characteristic deep cut brow formation; upper two (1, 2) have small three-fingered hands, but in lowest, hands are not defined; upper two have mesial support (5) to body, but lowest ends without legs and its body is supported directly by bifurcated lower end (6). d, side view: shows square shoulders, slender upper arms, low abdomen, and slender legs.

An interesting artifact with multiple figures in the Marine Museum, Brest, has two full figures back to back at the top and another figure lower down in page 344front. They form the upper end of a fairly large four-sided piece of wood that increases in size toward the bottom. The Brest Museum labeled it as "avant de pirogue sculpte" from Tahiti. It is probably a carved figure for a canoe, and judging from the canoes of the other islands in the group, it was probably erected at the stern. The locality is undoubtedly Aitutaki, for the figures have the typical face, body, and limbs of figures from that island. The two upper figures have a mesial support (fig. 214).

Figure 214.—Aitutaki carved figure (Musee de Hopital de la Marine, Brest, 486). a, front: upper figure (1) with characteristic eyes, brows, nose, mouth, and sharp chin; ornamental band (2) across forehead; pierced ears; slender forearm with three-fingered hand; high hips, slender legs and mesial support (3); lower head (4) similar to upper but with mesial ridge on forehead; square shoulders with square slanting chest and three-fingered hands; lower part (5) with transverse notched ridges from sides. b, back: upper figure (1) with typical head, face, body, hands and slender legs; lower part (2) forms plain panel with side edges and lower border; shows side projections. c, side view: heads with sharp chins, flat chests, slender upper arms, protruding abdomen with low navel (1) on front figure; acutely flexed thighs and mesial support shared by both upper figures; lower front figure shows mesial forehead ridge (2) and flexed leg (3) with projections on leg; transverse notched ridges on lower part.

Figure 214.—Aitutaki carved figure (Musee de Hopital de la Marine, Brest, 486). a, front: upper figure (1) with characteristic eyes, brows, nose, mouth, and sharp chin; ornamental band (2) across forehead; pierced ears; slender forearm with three-fingered hand; high hips, slender legs and mesial support (3); lower head (4) similar to upper but with mesial ridge on forehead; square shoulders with square slanting chest and three-fingered hands; lower part (5) with transverse notched ridges from sides. b, back: upper figure (1) with typical head, face, body, hands and slender legs; lower part (2) forms plain panel with side edges and lower border; shows side projections. c, side view: heads with sharp chins, flat chests, slender upper arms, protruding abdomen with low navel (1) on front figure; acutely flexed thighs and mesial support shared by both upper figures; lower front figure shows mesial forehead ridge (2) and flexed leg (3) with projections on leg; transverse notched ridges on lower part.

page 345
Atiu Gods

In the London Missionary Society collection in the British Museum, there are two gods attributed to Atiu and two other objects that have an affinity with the identified gods. Of the two identified gods, one is made of wood and the other is a sennit bundle covered with red feathers. No wooden images are known, hence the art of Atiu differs materially from that of Rarotonga and Aitutaki. The wooden god differs from the flat slabs of Aitutaki, but it has arches with the lozenge-chevron carved motif similar to those of Aitutaki. The Atiu gods thus divide into two classes as regards technique; carved wood and sennit bundles.

Figure 215.—Wooden god from Atiu (British Mus., L.M.S., 94/35). a, wooden figure: height 350 mm.; width of head 105 mm. by 108 mm.; width lower end 113 mm.; head (1) dome-shaped with six vertical elliptical slots (2) crossed by middle bar (3), dividing head into six elliptical panels (4) which are carved on outer side; constricted neck (5) followed by vertical arches (6) on same vertical lines as head panels and radiating equally from common center to form upper series; upper ends of arches have projecting cleats carved in lozenge-chevron motif (7); lower end of arch projects downward like two legs (8) with hole (9) pierced below them; interior of arches completely hollowed out; upper series followed by thick interspaces (10) before lower series of arches (11) which resemble upper but are reversed with legs (8) above and lozenge-chevron motif (7) below; interior of second set also hollowed out. Lower part, destroyed by rats, probably was a spatulate staff similar to those in the Mitiaro gods. b, braid decoration (12) in position, tied to crossbars in head slots and hanging down in inter-arch spaces; with lozenge-chevron motifs of arches showing (7). c, upper end of head, showing slots (2) with crossbars (3) and panels (4); head is hollow. d, side of head showing slots and panels with carved motifs on middle wider parts of panels; the six panels are alternately wide and narrow. e, carving motifs on panels: 1, chevrons on all three wide panels; all narrow panels with two horizontal raised bars on lower end but motif in middle varies; 2, notched triangle-chevron; 3, transverse bar and chevron; 4, bar-double triangle-chevron.

Figure 215.—Wooden god from Atiu (British Mus., L.M.S., 94/35). a, wooden figure: height 350 mm.; width of head 105 mm. by 108 mm.; width lower end 113 mm.; head (1) dome-shaped with six vertical elliptical slots (2) crossed by middle bar (3), dividing head into six elliptical panels (4) which are carved on outer side; constricted neck (5) followed by vertical arches (6) on same vertical lines as head panels and radiating equally from common center to form upper series; upper ends of arches have projecting cleats carved in lozenge-chevron motif (7); lower end of arch projects downward like two legs (8) with hole (9) pierced below them; interior of arches completely hollowed out; upper series followed by thick interspaces (10) before lower series of arches (11) which resemble upper but are reversed with legs (8) above and lozenge-chevron motif (7) below; interior of second set also hollowed out. Lower part, destroyed by rats, probably was a spatulate staff similar to those in the Mitiaro gods. b, braid decoration (12) in position, tied to crossbars in head slots and hanging down in inter-arch spaces; with lozenge-chevron motifs of arches showing (7). c, upper end of head, showing slots (2) with crossbars (3) and panels (4); head is hollow. d, side of head showing slots and panels with carved motifs on middle wider parts of panels; the six panels are alternately wide and narrow. e, carving motifs on panels: 1, chevrons on all three wide panels; all narrow panels with two horizontal raised bars on lower end but motif in middle varies; 2, notched triangle-chevron; 3, transverse bar and chevron; 4, bar-double triangle-chevron.

page 346
Wooden Gods

The identified wooden god consists of a column with a dome-shaped head and a lower portion with vertical sets of arches radiating from a center. Feathers and thick sennit covered with red feathers are used as ornamentation (pl. 14, C). The London Missionary Society catalog gives the following reference concerning this god: "Tarignarue, the superior god of Atiu, a very deeply carved idol with red feathers, also a large piece made of cinet and feathers. These are but a portion of the god; the rats having made a nest of him, destroyed the remainder." Somebody, evidently seeing the error in the spelling of the name, wrote in pencil: "Taringa-nue—Great ears. The chief deity of Atiu, Mitiaro, and Mauke."

When Rongo-ma-tane, high chief of Atiu, visited Mitiaro with John Williams (81, p. 98), he ordered the people to destroy their gods. "But", said one, "Must we burn Taria-nui (Great Ears) ?" "Yes," replied the king, "Commit him and all evil spirits to the flames." If the translation of the name is Big Ears, the correct spelling is Taringa-nui. But Williams gives this name to the fisherman's god in Rarotonga (p. 312). It is curious that the same name
Figure 216.—Feather decoration of Atiu god: a, coir cord (1) 2 mm. thick with end (2) doubled over and feathers (3) attached to it with thread of bast (4). b, strip of coir fiber (1) 25 mm. long with straight end (2) to which feathers were tied, served as intermediate holder to attach to a cord (c). c, fiber holder (1) attached to doubled-over cord (2) by thread lashing (3). d, two separate bunches (1, 1) of black feathers lashed around quills (2) and lashed (3) to doubled-over coir cord (4). e, doubled-over fiber holder (1) with red feathers tied to cord with knotted end (2) and tropic bird tail feathers (3) tied lower down. f, coir cord carrier (1) with four holders 58 mm. long, with ends doubled over (2) to which feathers were attached by single coir fiber still in position; bast cord (3) looped between holders and end tied to standing part by two spaced overhand knots; cord was lowered through top of slits in dome head and used to pull carrier with feathers down into slit and tied to lower ends of head panels. g, thick coir braid (1) 22 mm. wide, 9-10 mm. thick, and 360 mm. long, covered on front with small red feathers either singly or in small bunches with spiral bast thread (2) passing over quills; cord or fiber holders (a, b) with feathers were also laid on in pairs (3, 3) and fixed in same way as red feathers; tropic bird feathers were added at lower end of braid. Four lengths of braid, in position, were hung with a fine thread through their commencement end to the dome slit crossbar; one was loose, and the sixth was lost. h, tuft of human hair seized with single coir fiber.

Figure 216.—Feather decoration of Atiu god: a, coir cord (1) 2 mm. thick with end (2) doubled over and feathers (3) attached to it with thread of bast (4). b, strip of coir fiber (1) 25 mm. long with straight end (2) to which feathers were tied, served as intermediate holder to attach to a cord (c). c, fiber holder (1) attached to doubled-over cord (2) by thread lashing (3). d, two separate bunches (1, 1) of black feathers lashed around quills (2) and lashed (3) to doubled-over coir cord (4). e, doubled-over fiber holder (1) with red feathers tied to cord with knotted end (2) and tropic bird tail feathers (3) tied lower down. f, coir cord carrier (1) with four holders 58 mm. long, with ends doubled over (2) to which feathers were attached by single coir fiber still in position; bast cord (3) looped between holders and end tied to standing part by two spaced overhand knots; cord was lowered through top of slits in dome head and used to pull carrier with feathers down into slit and tied to lower ends of head panels. g, thick coir braid (1) 22 mm. wide, 9-10 mm. thick, and 360 mm. long, covered on front with small red feathers either singly or in small bunches with spiral bast thread (2) passing over quills; cord or fiber holders (a, b) with feathers were also laid on in pairs (3, 3) and fixed in same way as red feathers; tropic bird feathers were added at lower end of braid. Four lengths of braid, in position, were hung with a fine thread through their commencement end to the dome slit crossbar; one was loose, and the sixth was lost. h, tuft of human hair seized with single coir fiber.

page 347 should be given to a fisherman's god in Rarotonga and to the chief deity of Atiu, Mitiaro, and Mauke. The name Taringa-nui is not in a list of gods given to me in Atiu, but the name Tangiia-nui is in the list, and Tangiia-nui was an important god in Atiu, Mauke, and Mitiaro. It would appear that the missionaries wrote the incorrect spelling of Taringa-nui for Tangiia-nui and then translated the wrong name correctly as Big Ears. Whatever the name of the god was, that part of his material remains which escaped the ravages of the rats is shown in figure 215.
The feathers used were red parakeet, black man-of-war hawk, and white tropic birds as in Aitutaki, but the small white feathers are absent. Feathers were lashed to cord and coir holders for attachment to the outer side of the dome panels by lashing around the constricted lower ends of the panels. Other holders and carriers were attached to a strip of bast fiber which was pulled down through the upper parts of the dome slits until the feathers projected above the dome. The bast strips were then tied to the lower ends of the panels. The thick braid lengths, covered with red feathers that were hung to the cross bars of the dome slits, have lost most of their feathers but the fine threads that originally fastened the feathers are still in position. Two tufts of human hair were also used. The six arches arranged in two horizontal series are
Figure 217.—Staff with uncarved arches (British Mus., L.M.S., 295). a, full figure: 1, upper terminal; 2, upper set of four arches with no central rod; 3, central rod between sets of arches; 4, lower set of arches, showing upper knobs (5) and lower knobs (6); 7, lower rounded terminal with transverse edge (8) and bottom end (9) iractured; length from upper end to lower transverse edge (8) 455 mm. b, set of arches: 1, 2, side arches; 3, back arch; 4, front arch removed to show back arch; 5, upper arch projection; 6, lower projections; length of arch between ends of upper and lower projections 108 mm.; width across the middle of opposite arches 72 mm.

Figure 217.—Staff with uncarved arches (British Mus., L.M.S., 295). a, full figure: 1, upper terminal; 2, upper set of four arches with no central rod; 3, central rod between sets of arches; 4, lower set of arches, showing upper knobs (5) and lower knobs (6); 7, lower rounded terminal with transverse edge (8) and bottom end (9) iractured; length from upper end to lower transverse edge (8) 455 mm. b, set of arches: 1, 2, side arches; 3, back arch; 4, front arch removed to show back arch; 5, upper arch projection; 6, lower projections; length of arch between ends of upper and lower projections 108 mm.; width across the middle of opposite arches 72 mm.

page 348 devoid of applied decoration but it is probable that they were originally covered with feathers. For the feather and hair technique, see figure 216.

A plain staff with two sets of uncarved arches and a terminal upper point in the London Missionary Society collection has a number but no description or locality. The technique consisted of cutting out four ridges on a round staff, shaping them into two transverse sets of four arches with upper and lower projections, and then excavating the interior part between the arches of each set. The work resembles that of the identified wooden god (fig. 215) and it is thus tentatively ascribed to Atiu. The arch projections are simply blunt knobs without any carving motif. The arches were undoubtedly meant for the attachment of feather decorations. The artifact was figured by Edge-Partington (24, I-23-3) without any identification. (See figure 217.)

Sennit and Feather Gods

Two gods in the London Missionary Society collection are longitudinal folds of sennit with the outer surface of the outer layer covered with small red parakeet feathers. One—roughly shaped to form a head, body, and limbs—is labeled, "God formed of Cinet and feathers, L.M.S. 36, Atiu?" The note of interrogation after Atiu indicates a doubt as to the locality. Edge-Partington (24, I-22-1) figures this object and adds the following legend, "No. 1. Ronga [Rongo] ? A rude human figure formed of cords of plaited sinnet covered with feathers from Orutanga [Arutanga] a district of Aitu-take [Aitutaki]." However, I believe that the designation of Atiu is correct, as the technique of using thick braid with feathers on the outer surface is similar to that of the braid pendants in the authenticated wooden god (fig. 215).

A new element is introduced in the use of coconut leaf stipule (kaka) to fold around the sennit foundation, and the outer lengths are covered with thin white tapa to provide a smoother bed for the feather decoration. The feathers are laid in transverse rows and fixed by spiral turns of a thin oronga thread passing over the quills. The red feathers, and occasionally white and yellow feathers, were tied in bunches of five or six with a thin bast filament before attachment to the braid lengths. Plumes of split black feathers and tropic bird tail feathers on holders or ring carriers were placed thickly around the feet, lower abdomen, arms, and neck. The wing and tail feathers of the parakeet were also used.

The various combinations of feathers were kept in position by spiral turns around the sennit lengths. Two tufts of human hair were also used. The god and the feather technique are shown in figure 218.

The second sennit bundle is two and a half times as long as the first one but it is merely a bundle, thicker in the middle than at the ends, and without any attempt to distinguish anatomical elements (pl. 14, B). It is composed page 349
Figure 218.—Sennit and feather god (British Mus., L.M.S., 36). a, full figure: thick sennit covered with folded strip of coconut stipule folded longitudinally to roughly form rounded head (1), long body (2), arms (3) 167 mm. long, and legs (4) 130 mm. long; waist width, 121 mm.; total length, 555 mm.; outer folds covered with thin bark cloth from 9 to 14 mm. wide; head and neck formed by constricting with cord and so with arms and legs. b, tropic bird tail feathers (1) with small red feathers (2) tied to quill of long feathers. c, coir cord holder (1) with small red feathers (2), split black (3) and tropic bird tail feathers (4). d, sennit holder (1) with small red (2) split back (3), tropic bird (4), and wing and tail feathers of parakeet (5). e, sennit holder (1) with feathers similar to those in c but upper end (6) of holder projects above red feathers; other end doubled back so that turn of spiral thread (7) fixes holder to sennit length. f, spiral thread (7) takes turn around some of feather holders (1). g, ring carrier (1) with seven cord holders (2) each carrying red feathers (3) and tropic bird feathers (4). h, ring carrier, g, hunched up to form rosette which will be kept in position by spiral thread as in e. i, tuft of human hair, 120 mm. long, with spaced seizing by single coir fiber to form eyelet hole.

Figure 218.—Sennit and feather god (British Mus., L.M.S., 36). a, full figure: thick sennit covered with folded strip of coconut stipule folded longitudinally to roughly form rounded head (1), long body (2), arms (3) 167 mm. long, and legs (4) 130 mm. long; waist width, 121 mm.; total length, 555 mm.; outer folds covered with thin bark cloth from 9 to 14 mm. wide; head and neck formed by constricting with cord and so with arms and legs. b, tropic bird tail feathers (1) with small red feathers (2) tied to quill of long feathers. c, coir cord holder (1) with small red feathers (2), split black (3) and tropic bird tail feathers (4). d, sennit holder (1) with small red (2) split back (3), tropic bird (4), and wing and tail feathers of parakeet (5). e, sennit holder (1) with feathers similar to those in c but upper end (6) of holder projects above red feathers; other end doubled back so that turn of spiral thread (7) fixes holder to sennit length. f, spiral thread (7) takes turn around some of feather holders (1). g, ring carrier (1) with seven cord holders (2) each carrying red feathers (3) and tropic bird feathers (4). h, ring carrier, g, hunched up to form rosette which will be kept in position by spiral thread as in e. i, tuft of human hair, 120 mm. long, with spaced seizing by single coir fiber to form eyelet hole.

entirely of lengths of sennit doubled back at the ends. No stipule strips or white cloth were used, but some of the inner lengths were wrapped in pandanus leaf kept in position by spiral turns of a fine thread. The outer braid lengths were covered on the outer side with small red parakeet feathers put on in transverse rows and kept in position with a spiral thread in exactly the same technique as the braid pendants in the wooden god (fig. 216, g). The lower end was hollow. Small white and small yellow feathers were attached to the braid in narrow bands. Here and there split black feathers, parakeet wing and tail feathers, and tropic bird tail feathers were included under the spiral turns of the general fixation thread. Adjacent outer braids were fixed side by side by a special thread. (See figure 219.)
page 350
Figure 219.—Large sennit feather god, Atiu (British Mus., L.M.S. coll., 572). a, total length 450 mm., circumference at top (1) 330 mm., at middle 520 mm., lower end (2) 415 mm.; foundation sennit thick and coarse, ranging from 10 mm. wide and 4 mm. thick to 25 mm. wide and 12 mm. thick; strands of fiber used in plies bound together by single coir fiber before plaiting; some of internal braid lengths covered with feathers on one side in same manner as outer sennit, evidently due to more braid being covered with feathers than required for outer lengths and hence covered in. b, top end: sennit lengths doubled back at top end (1) and closely bound with fine cord; feathers attached to top bend; three lengths (2) have free ends 420 mm. long which projected beyond end but were turned back to outside of body and tied to it with trade string (evidently came loose). c, lower end: sennit lengths end in free ends not connected with each other for a distance of 421 mm., and last 85 mm. had feathers attached on both sides of sennit by spiral method; interior turns forming solid body evidently ended some distance from lower end, while insertion of a vine loop (1) with its ends overlapping for 9 mm. kept outer layer of sennit in form of a hollow cylinder, the free braids attached to circumference of hoop by threads. d, technique of connecting outer feather-covered length of sennit: two neighboring lengths (1, 2) with continuous spiral turns here made from left to right; a thread (3) passes down between braids and at intervals of 43 to 47 mm. under two adjacent turns of feather fixation thread (4, 5) and around standing part to form an overhand knot which is drawn taut; a thread passed down every interval between sennit lengths for whole circumference, and all outer sennit lengths were bound together.

Figure 219.—Large sennit feather god, Atiu (British Mus., L.M.S. coll., 572). a, total length 450 mm., circumference at top (1) 330 mm., at middle 520 mm., lower end (2) 415 mm.; foundation sennit thick and coarse, ranging from 10 mm. wide and 4 mm. thick to 25 mm. wide and 12 mm. thick; strands of fiber used in plies bound together by single coir fiber before plaiting; some of internal braid lengths covered with feathers on one side in same manner as outer sennit, evidently due to more braid being covered with feathers than required for outer lengths and hence covered in. b, top end: sennit lengths doubled back at top end (1) and closely bound with fine cord; feathers attached to top bend; three lengths (2) have free ends 420 mm. long which projected beyond end but were turned back to outside of body and tied to it with trade string (evidently came loose). c, lower end: sennit lengths end in free ends not connected with each other for a distance of 421 mm., and last 85 mm. had feathers attached on both sides of sennit by spiral method; interior turns forming solid body evidently ended some distance from lower end, while insertion of a vine loop (1) with its ends overlapping for 9 mm. kept outer layer of sennit in form of a hollow cylinder, the free braids attached to circumference of hoop by threads. d, technique of connecting outer feather-covered length of sennit: two neighboring lengths (1, 2) with continuous spiral turns here made from left to right; a thread (3) passes down between braids and at intervals of 43 to 47 mm. under two adjacent turns of feather fixation thread (4, 5) and around standing part to form an overhand knot which is drawn taut; a thread passed down every interval between sennit lengths for whole circumference, and all outer sennit lengths were bound together.

Mitiaro Gods
The specimens of Mitiaro gods in English museums conform to one type, consisting of a dome-shaped head, a body carved into vertical rows of small arches, and a lower spatulate staff. Five of these neatly carved objects are in the British Museum. One, together with three fragments, in the Cambridge University Museum, and two with one fragment in the Oldman collection. Three fragments are in the Pitt-Rivers Museum, Oxford. Of the British Museum specimens, three were acquired with the London Missionary Society collection, and two were obtained from the Sheffield Literary and Philosophical Society. One of the London Missionary Society specimens, both of those page 351in the Sheffield collection, and the Cambridge specimen were minus the lower staff end. The three Cambridge fragments, which are broken-off staff ends, and the three similar fragments in the Pitt-Rivers Museum probably belong to some of the incomplete specimens. Identification of locality rests on two specimens in the British Museum which bear the old numbers 106 and 107; both were numbered 49 in the London Missionary Society catalog with the description, "Idols from Metiaro." (Throughout the writings of John Williams (81) Mitiaro is spelled Metiaro.) The Cambridge specimens also were obtained from Sheffield, and it is known that the Sheffield specimens were presented to the local Philosophical Society by George Bennet, Esquire, of that city. Bennet visited the South Seas with Reverend D. Tyerman in 1821 at the instance of the London Missionary Society. Williams (81, p. 92) records giving a number of Aitutaki religious objects to the deputation, and it is evident that the objects presented to the Sheffield Society were specimens collected from the Cook Islands by the London Missionary Society after the inhabitants of each island embraced the new religion. No doubt exists therefore
Figure 220.—Mitiaro gods, head section. a: 1, vertex; 2, panels; 3, transverse bar, uncarved; 4, neck; 5, first arches of middle section (Cambridge University Mus., 104). b: greatest width 36 mm., neck width 21 mm., transverse bar (3) with single zigzag pattern (British Mus., 7050). c: greatest width 53 mm., height 85 mm., neck width 24 mm., transverse bar (3) with single zigzag pattern; shows upper ends of coir decoration attached to middle section; transverse lines (4) (British Mus., L.M.S., 107). d: greatest width 41 mm., height 73 mm., neck width 17 mm.; transverse bar (3) with double zigzag pattern (British Mus., + 2060). e: greatest width 60 mm., height 75 mm., neck width 36 mm., panels 57 mm. high and 18 mm. at widest part; transverse bar (3) with double zigzag pattern; vertex with knob on top (British Mus., L.M.S., 106). f: greatest width 69 mm., height 68 mm., neck width 33 mm.; five panels instead of four, panel height 50 mm., panel width 12 mm.; transverse bar (3) with double zigzag pattern; neck incised with transverse lines (4) (British Mus., L.M.S., no number).

Figure 220.—Mitiaro gods, head section. a: 1, vertex; 2, panels; 3, transverse bar, uncarved; 4, neck; 5, first arches of middle section (Cambridge University Mus., 104). b: greatest width 36 mm., neck width 21 mm., transverse bar (3) with single zigzag pattern (British Mus., 7050). c: greatest width 53 mm., height 85 mm., neck width 24 mm., transverse bar (3) with single zigzag pattern; shows upper ends of coir decoration attached to middle section; transverse lines (4) (British Mus., L.M.S., 107). d: greatest width 41 mm., height 73 mm., neck width 17 mm.; transverse bar (3) with double zigzag pattern (British Mus., + 2060). e: greatest width 60 mm., height 75 mm., neck width 36 mm., panels 57 mm. high and 18 mm. at widest part; transverse bar (3) with double zigzag pattern; vertex with knob on top (British Mus., L.M.S., 106). f: greatest width 69 mm., height 68 mm., neck width 33 mm.; five panels instead of four, panel height 50 mm., panel width 12 mm.; transverse bar (3) with double zigzag pattern; neck incised with transverse lines (4) (British Mus., L.M.S., no number).

page 352 that these objects were made in Mitiaro with native tools before European influence affected the native crafts (pl. 14, D, E).

The dome-shaped head is fashioned and hollowed out to form a thin curved vertex, supported by equidistant vertical panels elliptical in form and coming to points above and below. The lower ends of the panels join the solid part just above the constricted neck between the head and the middle section. In hollowing out the interior of the head, transverse bars are left to connect the middle of adjacent vertical panels. The panels are plain, but the transverse bars, except in one specimen, are carved in zigzag patterns on raised bars which may be single or double. In one of the Oldman specimens the upper dome part has been broken off. Of the seven gods with domes, six have four panels and one has five. For details of the head see figure 220.

The transverse bar is the only part with decorative carving, except transverse lines on the upper part of the neck in figure 220, c, f. The height of the head ranges from 68 to 85 mm.; the greatest width is 36 to 69 mm. Except for minor details the heads conform to one pattern of structure.

In the middle body section, the wood in the round was cut down longitudinally
Figure 221.—Middle section formation of separate arches: a, wood shaped in the round with equidistant parts (1) marked off to form ridges, and intermediate parts (2) to be removed, leaving a central column (3) to which ridges remain attached. b, intermediate parts (a,2) removed, leaving five ridges (1) attached to central column (3). c, arches formed by piercing a hole (4) of shape shown through ridge (1) and cutting out a piece (5) also of shape shown. Arch itself is more definitely denned by cutting out triangular pieces (6, 6) to outer side of arch, defining arch and forming outward extensions of upper and lower pillars which support it. Pillars are further denned by removing angular pieces (5) which provide projecting points to pillars. Thus, as shown on lower arch, we have arch (7), an upper (8) and lower (9) pillar with outer extensions (10) which also project away from arch. Each ridge is treated in similar manner. One ridge (11) omitted from drawing for clarity. d, outer surface of each arch (7) is trimmed into two inclined planes which form a mesial edge. Outer extensions (10) of pillars which form rectangular surfaces are carved to form fused lozenge-chevron motif with chevron of each pillar toward its own arch. Completed arches form horizontal groups as well as longitudinal series. Thinness of central column (3) is plainly seen and accounts for number of specimens broken.

Figure 221.—Middle section formation of separate arches: a, wood shaped in the round with equidistant parts (1) marked off to form ridges, and intermediate parts (2) to be removed, leaving a central column (3) to which ridges remain attached. b, intermediate parts (a,2) removed, leaving five ridges (1) attached to central column (3). c, arches formed by piercing a hole (4) of shape shown through ridge (1) and cutting out a piece (5) also of shape shown. Arch itself is more definitely denned by cutting out triangular pieces (6, 6) to outer side of arch, defining arch and forming outward extensions of upper and lower pillars which support it. Pillars are further denned by removing angular pieces (5) which provide projecting points to pillars. Thus, as shown on lower arch, we have arch (7), an upper (8) and lower (9) pillar with outer extensions (10) which also project away from arch. Each ridge is treated in similar manner. One ridge (11) omitted from drawing for clarity. d, outer surface of each arch (7) is trimmed into two inclined planes which form a mesial edge. Outer extensions (10) of pillars which form rectangular surfaces are carved to form fused lozenge-chevron motif with chevron of each pillar toward its own arch. Completed arches form horizontal groups as well as longitudinal series. Thinness of central column (3) is plainly seen and accounts for number of specimens broken.

page 353to provide equidistant ridges radiating from a central thin column, which in some of the smaller figures is as thin as 8 mm. in diameter. Four figures had five ridges, but one (fig. 220, f) had six ridges for the upper half, reduced to five for the lower half. Another (fig. 220, e) had seven ridges throughout. The ridges were undercut to form a series of arches, each arch with two pillars and each separated from its neighbors, except in the specimen with seven ridges. In specimens with separated arches the ridges were each formed into four arches and each arch corresponded in its respective level with the arches formed on other ridges. The method of carving the arches and their ornamentation is shown in figure 221. It is evident that the object of the craftsman was to form a series of radiating arches with ornamental cleats, both of which were for the attachment of feather decoration. No evidence is apparent that this technique ever represented a conventionalized human form.
Figure 222.—Middle section, formation of continuous arches. a, column of wood cut into radiating ridges which number seven: one ridge is shown with holes pierced through but none cut through to surface; holes numbered 1 correspond to those for true arches (fig. 221, c,4), and alternate holes (2) correspond to those which reach surface in separate arch technique (fig. 221, c, 5). b, wood over alternate holes (1) of a shaped to form true arches by a technique similar to that in separate arches (fig. 221, c,7), but outer extensions of pillars remain fused because other alternate holes were not cut through to surface at points 2. Carving motif retains the chevron toward true arch but has a broad distal end instead of a pointed one and becomes a triangle instead of a lozenge. Carving motifs of adjacent motifs are separated from each other by a square notch at 2, and artist utilized square cut to form base of a triangle motif and to keep up proportion of parts; an angular notch on either side results in a double triangle with chevron fused to apex of triangle next to it. In series shown on ridge (3) facing observer, each true arch corresponding to those marked 1 on side has two pillar motifs with chevron toward arch. In uppermost arch, upward projection of first pillar is free and is carved with lozenge-chevron motif (fig. 221, d), but downward free projection of lower pillar of lowest arch retains general double triangle-chevron motif. Total length of carved ridges, 220 mm. (British Mus., L.M.S. 106).

Figure 222.—Middle section, formation of continuous arches. a, column of wood cut into radiating ridges which number seven: one ridge is shown with holes pierced through but none cut through to surface; holes numbered 1 correspond to those for true arches (fig. 221, c,4), and alternate holes (2) correspond to those which reach surface in separate arch technique (fig. 221, c, 5). b, wood over alternate holes (1) of a shaped to form true arches by a technique similar to that in separate arches (fig. 221, c,7), but outer extensions of pillars remain fused because other alternate holes were not cut through to surface at points 2. Carving motif retains the chevron toward true arch but has a broad distal end instead of a pointed one and becomes a triangle instead of a lozenge. Carving motifs of adjacent motifs are separated from each other by a square notch at 2, and artist utilized square cut to form base of a triangle motif and to keep up proportion of parts; an angular notch on either side results in a double triangle with chevron fused to apex of triangle next to it. In series shown on ridge (3) facing observer, each true arch corresponding to those marked 1 on side has two pillar motifs with chevron toward arch. In uppermost arch, upward projection of first pillar is free and is carved with lozenge-chevron motif (fig. 221, d), but downward free projection of lower pillar of lowest arch retains general double triangle-chevron motif. Total length of carved ridges, 220 mm. (British Mus., L.M.S. 106).

page 354

One of the specimens (fig. 220, e) shows a departure from the general pattern, in that it has seven ridges on the middle part instead of five and in that the arches are continuous instead of separated. The upper ends of the top arches have free extensions carved with the lozenge-chevron motif, but below them the carving motif changes to a triangle and chevron combination which extends over the uncut arches (fig. 222 and pl. 14, E).

From the above discussion, it is clear that the square cuts between adjacent motifs (fig. 222, b,2) mark the boundaries between adjacent arches but that the wood below the notches has not been cut away to make a dintinct division as in the separate arch technique. Each ridge has four true arches, corresponding with the number in the objects made with the separate arch technique.

The use of the triangle with the chevron as an art motif, although probably influenced by the technique of the continuous arches, is not entirely due to it, as evidenced by the appearance of a similar motif on a specimen in which the arches are separated (fig. 224, c, 2).

The slenderness of the central column and arch pillars is exemplified by one of the Sheffield specimens, in which many of the arches are broken off (fig. 223, a). The manner in which decoration is applied to the lower end is shown in figure 223, b.

Figure 223.—Mitiaro gods, broken arches and feather decoration. a, central column (1), 8 mm. in diameter; perfect arch (2) 38 mm. between upper and lower pillars, and broken pillars (3) of other arches on same level may be seen; lozenge-chevron motif (4) used (British Mus., 2060). b, only specimen of series which has sennit braid decoration still attached to it; feathers and braid with transverse tufts of fiber are placed vertically in grooves between carved ridges and then tied in position by transverse turns of thin white tapa (1); black feathers (2) project about 145 mm. above tapa; braid with cross tufts, which may be seen in middle of figure, projects beyond binding for 160 mm. (British Mus., L.M.S., 107).

Figure 223.—Mitiaro gods, broken arches and feather decoration. a, central column (1), 8 mm. in diameter; perfect arch (2) 38 mm. between upper and lower pillars, and broken pillars (3) of other arches on same level may be seen; lozenge-chevron motif (4) used (British Mus., 2060). b, only specimen of series which has sennit braid decoration still attached to it; feathers and braid with transverse tufts of fiber are placed vertically in grooves between carved ridges and then tied in position by transverse turns of thin white tapa (1); black feathers (2) project about 145 mm. above tapa; braid with cross tufts, which may be seen in middle of figure, projects beyond binding for 160 mm. (British Mus., L.M.S., 107).

page 355

The lower staff part of the gods is a continuation of the central column which expands in one diameter in some specimens to a spatulate form, whereas it remains round in others (fig. 224).

Figure 224.—Lower staff ends of Mitiaro gods. a, staff (1) commences immediately below lowest arches (2) and widens, but part of length evidently cut off, as shown by cuts on either side with a middle fracture of wood; length 97 mm., width at top 14 mm., width, widest part, 32 mm. (British Mus., L.M.S., no number). b, other view of a: top thickness 13 mm., bottom thickness 12 mm. c, narrow view of staff (1) which has a distinct shoulder (3) below last set of arches (2); shoulder 22 mm. in diameter, just below it staff is 12 mm. in diameter and at lower end 9 mm.; on side not shown, staff is 14 mm. in diameter below shoulder and 20 mm. at lower end; lower part of staff also cut off (British Mus., L.M.S., 106).

Figure 224.—Lower staff ends of Mitiaro gods. a, staff (1) commences immediately below lowest arches (2) and widens, but part of length evidently cut off, as shown by cuts on either side with a middle fracture of wood; length 97 mm., width at top 14 mm., width, widest part, 32 mm. (British Mus., L.M.S., no number). b, other view of a: top thickness 13 mm., bottom thickness 12 mm. c, narrow view of staff (1) which has a distinct shoulder (3) below last set of arches (2); shoulder 22 mm. in diameter, just below it staff is 12 mm. in diameter and at lower end 9 mm.; on side not shown, staff is 14 mm. in diameter below shoulder and 20 mm. at lower end; lower part of staff also cut off (British Mus., L.M.S., 106).

The total length of the two unbroken specimens in the British Museum (London Missionary Society 106 and ?) is 412 and 393 mm. respectively, but in both these the lower end of the staff part is cut off. The four broken specimens ranged in length from 215 to 233 mm. The head height of the series of six ranged from 68 to 85 mm.

Applied Decoration

Feathers and sennit decoration were undoubtedly applied to all the Mitiaro gods when dressed for religious ritual, but of the seven gods with heads intact, only two bore traces of former decoration. Ellis (25, vol. 2, frontispiece 4) figured a Mitiaro god with feathers and sennit attached to it, but he gave no explanation in the text.

Though complete detail of the decorative technique of the arches is lacking, the decoration of the lower staff ends is fully illustrated by the three lower fragments in the Cambridge Museum. The Pitt-Rivers specimens bear the following curious description: "3 atuas,—Symbolically carved pieces of wood, page 356with feathers, cocoa fibre strings, etc. attached, supposed to attract the beneficent spirits who take up their abode in them. E. Howard Esq. 1888."

The sennit decoration of the staff ends consists of long lengths of coir cord and braid with transverse or oblique tufts of coir fiber caught under a ply at regular intervals. These were tied with transverse turns to the upper end of the staff just below the last arch, and the lashing was covered with a strip of thin white tapa also lashed with transverse turns of fine thread. Such transverse tufts with a two-ply cord were used in the carved slabs of Aitutaki (figs. 203, b, d and fig. 204). I identified the staff ends with the sennit decoration at Cambridge because the fracture had occurred through the fine central rod of the middle section and the upper part of two fragments carried pillars and arches with the typical Mitiaro technique and the lozenge-chevron motif. It is probable that some of the lower fragments belonged to some of the gods
Figure 225.—Staff ends with tufted sennit decoration (fragments in Cambridge University Mus., numbered group Z6045, but each bears an old London Missionary Society number). a (107): length of wooden part 135 mm., lower end 14 mm. by 8 mm.; one long cord with cross tufts with lashing of turns of thin white bark cloth commencing just below last broken arches. b (106): length of wooden part 130 mm.; lower pillars of arches left with lozenge-chevron motif; two lengths of tufted braid each 540 mm. long; tuft technique different than a, in that they are added as plies, and free ends of each braid are ornamented with a tassel (fig. 226); wrapping of white bark cloth covers part attached to staff. c (105): length of wooden part including lowest arch 142 mm.; one cord and one braid with transverse tufts each 530 mm. long; two tufts of fiber slightly overlapped and cord or braid ply crosses over overlapped part, this technique creating longer tufts; tuft on one side 80 mm. long, on other side 130 mm. long; no bark cloth wrapping, which enables technique of lashing cords and braids to staff to be seen; staff end of cord or braid is without tufts, and a knot (overhand) is tied a short distance from end; cord and braid ends laid against staff just below lowest arch, and a, lashing of transverse turns of continuous thread are made above knot and then carried below knot.

Figure 225.—Staff ends with tufted sennit decoration (fragments in Cambridge University Mus., numbered group Z6045, but each bears an old London Missionary Society number). a (107): length of wooden part 135 mm., lower end 14 mm. by 8 mm.; one long cord with cross tufts with lashing of turns of thin white bark cloth commencing just below last broken arches. b (106): length of wooden part 130 mm.; lower pillars of arches left with lozenge-chevron motif; two lengths of tufted braid each 540 mm. long; tuft technique different than a, in that they are added as plies, and free ends of each braid are ornamented with a tassel (fig. 226); wrapping of white bark cloth covers part attached to staff. c (105): length of wooden part including lowest arch 142 mm.; one cord and one braid with transverse tufts each 530 mm. long; two tufts of fiber slightly overlapped and cord or braid ply crosses over overlapped part, this technique creating longer tufts; tuft on one side 80 mm. long, on other side 130 mm. long; no bark cloth wrapping, which enables technique of lashing cords and braids to staff to be seen; staff end of cord or braid is without tufts, and a knot (overhand) is tied a short distance from end; cord and braid ends laid against staff just below lowest arch, and a, lashing of transverse turns of continuous thread are made above knot and then carried below knot.

page 357 in the British Museum. (See figure 225.) The technique of the tufted cords and braids is shown in figure 226, a-c. Some of the braid lengths end in a form of coir tassel shown in figure 226, d, e.

Though the coir tufts are placed horizontally across the cord or braid in the transverse method, the ends of the tufts hang down obliquely in much the same manner as in the oblique method and thus give a branched appearance. The oblique technique with sennit was used in Tahitian fly whisks. Stolpe (64, p. 41) states that in the cord ornaments of the Lille slabs (fig. 204) red feathers were attached to both and one had yellow feathers, but they were not present in the Mitiaro ornaments examined by me.

Figure 226.—Sennit tuft and tassel techniques. a, transverse method with cord: cord contains two plies (1, 2) and with every second twist a tuft of coir fiber (3) is placed between plies; overlapping method with two tufts of fiber is shown by tufts 4, 5. b, transverse method with braid: braid with three plies (1, 2, 3) includes a tuft (4) by placing it transversely under every second turn of ply that comes into middle position from left; overlapping tufts may also be used to give extra length to tufts. c, oblique method with braid only: after commencing braid, a tuft (1) is added to left-hand ply as it comes into middle, and as it is crossed by right-hand ply, a tuft (2) is added from that side; following left ply (3) and right ply (4) are twisted into position without adding any tufts, but tufts from left (5) and from right (6) are added on next two twists from their respective sides; thus tufts are added alternately from either side with every second twist on respective sides. d, tassel technique, first stage: braid (1) left at commencement end with fibers of the three plies (2) free and unknotted to form part of tassel; coir fibers 220 mm. long laid around braid end with middle point coinciding with end of unfrayed braid and lashed to braid by lashing (5) of a coir fiber in transverse turns. In figure, the near fibers left out for clarity. Thus added fibers (3) above lashing mingle with those of braid plies, whereas other half (4) hang down below lashing (5). e, tassel technique, second stage: lower half of fibers turned upward to join upper half and coir fiber tied with transverse turns in lashing (6); length of tassel from inner lashing (d,5) is half length of added fibers, namely 110. mm.

Figure 226.—Sennit tuft and tassel techniques. a, transverse method with cord: cord contains two plies (1, 2) and with every second twist a tuft of coir fiber (3) is placed between plies; overlapping method with two tufts of fiber is shown by tufts 4, 5. b, transverse method with braid: braid with three plies (1, 2, 3) includes a tuft (4) by placing it transversely under every second turn of ply that comes into middle position from left; overlapping tufts may also be used to give extra length to tufts. c, oblique method with braid only: after commencing braid, a tuft (1) is added to left-hand ply as it comes into middle, and as it is crossed by right-hand ply, a tuft (2) is added from that side; following left ply (3) and right ply (4) are twisted into position without adding any tufts, but tufts from left (5) and from right (6) are added on next two twists from their respective sides; thus tufts are added alternately from either side with every second twist on respective sides. d, tassel technique, first stage: braid (1) left at commencement end with fibers of the three plies (2) free and unknotted to form part of tassel; coir fibers 220 mm. long laid around braid end with middle point coinciding with end of unfrayed braid and lashed to braid by lashing (5) of a coir fiber in transverse turns. In figure, the near fibers left out for clarity. Thus added fibers (3) above lashing mingle with those of braid plies, whereas other half (4) hang down below lashing (5). e, tassel technique, second stage: lower half of fibers turned upward to join upper half and coir fiber tied with transverse turns in lashing (6); length of tassel from inner lashing (d,5) is half length of added fibers, namely 110. mm.

The decorations attached to the body of the British Museum specimen (fig. 223, b) consisted of a length of loose braid with about ten cross tufts, a length of sennit with four branches to which feathers were attached, loose long feathers, and one set of feathers with their quill ends seized into an eyelet hole. The feathers were laid in the vertical grooves between the arches with the tip ends directed upward. The branched braid and tufted braid were in different grooves and the whole was kept in position by transverse turns of white bark cloth. It is probable that this decoration was readjusted in London page 358after the lower staff was broken off. The specimen with continuous arches (fig. 222, b) had feathers attached to the lozenge-chevron cleats at the upper ends of the top arches. The thread attaching the feathers continued down and passed around some of the arches. Some of the other arches had the remnants of a two-ply thread around them showing that feathers had been attached to them. (See figure 227.)

Figure 227.—Feather decoration, Mitiaro gods (a, b: British Mus., L.M.S., 107; c-e: British Mus., L.M.S., 49). a, four pieces of braid (1-4) successively combined by fraying out lower ends and combining them as plies with others, forming branched figure combined into one braid (5); small bunch of black feathers (6) tied together at quill ends (7) added to upper braid ends by lashing (8); upper ends of braid (9) frayed out to form part of decoration; some branches without feathers but transverse turns of coir fiber remain around end of branches, showing feathers had been in position; feathers restored in figure. b, black feathers with quill ends doubled into loop (1), seized with single coir fiber, end passed around between two parts (2) and tied to protruding commencement end with reef knot. c, side view of top end of top pillar showing further shaping of lozenge-chevron motif (1) which forms a cleat. d, side view showing feather bunch (2) lashed to cleat (1) with turns around constricted neck of lozenge (3) and turns around chevron end (4). upper view of d, showing turns around lozenge neck (3) and figure-of-eight turns around arms of chevron (4), lashing material of single coir fiber (5) which passes on to be used on lower arches.

Figure 227.—Feather decoration, Mitiaro gods (a, b: British Mus., L.M.S., 107; c-e: British Mus., L.M.S., 49). a, four pieces of braid (1-4) successively combined by fraying out lower ends and combining them as plies with others, forming branched figure combined into one braid (5); small bunch of black feathers (6) tied together at quill ends (7) added to upper braid ends by lashing (8); upper ends of braid (9) frayed out to form part of decoration; some branches without feathers but transverse turns of coir fiber remain around end of branches, showing feathers had been in position; feathers restored in figure. b, black feathers with quill ends doubled into loop (1), seized with single coir fiber, end passed around between two parts (2) and tied to protruding commencement end with reef knot. c, side view of top end of top pillar showing further shaping of lozenge-chevron motif (1) which forms a cleat. d, side view showing feather bunch (2) lashed to cleat (1) with turns around constricted neck of lozenge (3) and turns around chevron end (4). upper view of d, showing turns around lozenge neck (3) and figure-of-eight turns around arms of chevron (4), lashing material of single coir fiber (5) which passes on to be used on lower arches.

Carving Motifs

From the practical use of the arches and pillar projections, it is evident that the geometrical patterns were largely influenced by utility in lashing. The simple lozenge-chevron motif was probably brought about by notching the sides of the pillar projection and bifurcating the inner end to form the chevron shape suitable for figure-of-eight lashing turns. The outer ends were trimmed into a point for balance, and the figure thus assumed the lozenge-chevron form. In some projections, the square end was not trimmed off and the form remained as triangle-chevron. Other minor variations consisted in the trimming of the upper surface of the arch. These and the two forms of bar ornamentation across the clefts in the dome-shaped head are shown in figure 228.

page 359
Figure 228.—Forms of carving on Mitiaro gods (a-e, arches and pillars; f, g, cross-bar on head sections): a, simple form of lozenge (1) and chevron (2) on pillar projections in which free point of lozenge is always directed outward, and chevrons directed inward toward arch; in trimming upper surface of arch to a mesial edge, a lozenge-shaped figure (3) is left on middle of arch. b, with simple lozenge-chevron motifs on pillar projections, and middle of arch (3) left as transverse band with slight points corresponding to mesial edges. c, upper projection (1) of top arch in god with continuous arches (fig. 222, b) is pointed, whereas lower end (4) is square-cut owing to shallow groove (5) separating it from next motif; triangle motif formed is treated with lateral notches (6) for lashing purposes, and a notched triangle-chevron results; middle part of arch (3) is trimmed to an even narrow band. d, blunt triangle (1) is used but extra-side projections (4, 4) formed in addition to chevron (2), resulting in a triangle-double chevron motif; middle of arch (3) rounded off without any projections. e, upper pillar projection (1) notched (6, 6) to form notched triangle-chevron motif, while lower projection (2) had second notches (7) in addition to first notches (6), forming double notched triangle-chevron motif. f, single zigzag band. g, double zigzag band.

Figure 228.—Forms of carving on Mitiaro gods (a-e, arches and pillars; f, g, cross-bar on head sections): a, simple form of lozenge (1) and chevron (2) on pillar projections in which free point of lozenge is always directed outward, and chevrons directed inward toward arch; in trimming upper surface of arch to a mesial edge, a lozenge-shaped figure (3) is left on middle of arch. b, with simple lozenge-chevron motifs on pillar projections, and middle of arch (3) left as transverse band with slight points corresponding to mesial edges. c, upper projection (1) of top arch in god with continuous arches (fig. 222, b) is pointed, whereas lower end (4) is square-cut owing to shallow groove (5) separating it from next motif; triangle motif formed is treated with lateral notches (6) for lashing purposes, and a notched triangle-chevron results; middle part of arch (3) is trimmed to an even narrow band. d, blunt triangle (1) is used but extra-side projections (4, 4) formed in addition to chevron (2), resulting in a triangle-double chevron motif; middle of arch (3) rounded off without any projections. e, upper pillar projection (1) notched (6, 6) to form notched triangle-chevron motif, while lower projection (2) had second notches (7) in addition to first notches (6), forming double notched triangle-chevron motif. f, single zigzag band. g, double zigzag band.

Maukean Gods

In the London Missionary Society collection in the British Museum there are two objects labeled as the god Tangaroa from Mauke. Unfortunately their origin is uncertain. One is a decorated wooden staff (pl. 16, B) and the other is a sennit bundle covered with bark cloth and decorated with feathers. The wooden object is described in the catalog as a "God in the form of a flat staff, expanded at the head and pierced to represent a conventional human figure. The greater part of the staff is covered with plaited sennit with remains of feather decoration." Another catalog entry read: "Tangaroa from Maute with feathers and fibre sennit appended to him." Maute was the spelling used by the Missionaries for Mauke at the time the gods were collected. The "conventional human figure" is a statement that is not supported by the appearance of the object.

Many mistakes have been made in the original catalog of the Missionary Museum, and these errors have accompanied the collection to the British Museum where doubtless additional errors were made. An outstanding example is the label on a Mangarevan image stating that it was the son of page 360
Figure 229.—Wooden god (British Mus., L.M.S., 66?): total length, 812 mm.; divided into upper perforated part in three sections, lower staff part covered with sennit. a, front: upper section (1) with pointed end, curved sides, and two large openings; wood on sides and middle, perforated with spaced small holes; base width, 121 mm. and base thickness, 29 mm.; second section (2) with sides curved from narrow upper junction and four large openings as shown; five wooden limbs and base perforated with spaced small holes; bunch of split black feathers lashed to outer limb (8) on left and feather lashing devoid of feathers also on right limb (9); middle limb has mesial edge from meeting of two planes; third section (3) with straight parallel sides and two longitudinal slots with lower ends covered by sennit wrapping; transverse lashing (10) of two-ply cord across three limbs for attaching white tropic bird tail feathers. Sennit wrapping in form of cuff (4) extends downward for 205 mm. with lower part covered by a band of white bark cloth (5) covered with red and yellow parakeet feathers secured by spiral turns of fine thread; wrapped part, 56 mm. in diameter at top, 70 mm. in middle, and 56 mm. at bottom. Below wrapping, transverse turns of thick braid (6), 8 mm. wide and 5 mm. thick, extend downward for 180 mm. Remaining 60 mm. of staff covered with black bark cloth (7). b, back: sennit cuff (4) is wrapped over inner layer of longitudinal lengths of thick sennit, 8 mm. wide, which is laid against wood and thickened in middle; ends of cuff (5, 5) made to meet on back and laced together with a length of sennit passing through loops on either end of cuff; cuff gapes at upper end and shows part of the sennit inner layer (6). c, cuff technique: cuff, made separately and applied after completion, consists of longitudinal elements or warps each composed of two strands of sennit, 2.5 mm. wide, which are looped at upper ends (5) to form four elements in each, warp; below free loops (5), compound warps appear to be seized with other sennit material; weft element (6) consists of single sennit length (6) which passes around first warp (1), crosses two warps (1, 2), around second (2), across next two warps (2, 3), around third (3), and so continues to cross two warps in front and around right element of pair; above technique with one weft element constitutes wrapped technique and differs from twined work which has two weft elements. d, successive rows of wrapped work are continued for width of cuff, each succeeding row being made close to one above as shown in the six rows (1-6) until depth of cuff is reached; warps (7). e, feather work: tropic bird tail feathers (1) are decorated in front at quill end with six tiers of small red parakeet feathers (2), commencing above and ending below with transverse lashing turns (3). f, back view, showing spiral turns of thread to keep red feathers in position; quill so covered is 41 mm. long.

Figure 229.—Wooden god (British Mus., L.M.S., 66?): total length, 812 mm.; divided into upper perforated part in three sections, lower staff part covered with sennit. a, front: upper section (1) with pointed end, curved sides, and two large openings; wood on sides and middle, perforated with spaced small holes; base width, 121 mm. and base thickness, 29 mm.; second section (2) with sides curved from narrow upper junction and four large openings as shown; five wooden limbs and base perforated with spaced small holes; bunch of split black feathers lashed to outer limb (8) on left and feather lashing devoid of feathers also on right limb (9); middle limb has mesial edge from meeting of two planes; third section (3) with straight parallel sides and two longitudinal slots with lower ends covered by sennit wrapping; transverse lashing (10) of two-ply cord across three limbs for attaching white tropic bird tail feathers. Sennit wrapping in form of cuff (4) extends downward for 205 mm. with lower part covered by a band of white bark cloth (5) covered with red and yellow parakeet feathers secured by spiral turns of fine thread; wrapped part, 56 mm. in diameter at top, 70 mm. in middle, and 56 mm. at bottom. Below wrapping, transverse turns of thick braid (6), 8 mm. wide and 5 mm. thick, extend downward for 180 mm. Remaining 60 mm. of staff covered with black bark cloth (7). b, back: sennit cuff (4) is wrapped over inner layer of longitudinal lengths of thick sennit, 8 mm. wide, which is laid against wood and thickened in middle; ends of cuff (5, 5) made to meet on back and laced together with a length of sennit passing through loops on either end of cuff; cuff gapes at upper end and shows part of the sennit inner layer (6). c, cuff technique: cuff, made separately and applied after completion, consists of longitudinal elements or warps each composed of two strands of sennit, 2.5 mm. wide, which are looped at upper ends (5) to form four elements in each, warp; below free loops (5), compound warps appear to be seized with other sennit material; weft element (6) consists of single sennit length (6) which passes around first warp (1), crosses two warps (1, 2), around second (2), across next two warps (2, 3), around third (3), and so continues to cross two warps in front and around right element of pair; above technique with one weft element constitutes wrapped technique and differs from twined work which has two weft elements. d, successive rows of wrapped work are continued for width of cuff, each succeeding row being made close to one above as shown in the six rows (1-6) until depth of cuff is reached; warps (7). e, feather work: tropic bird tail feathers (1) are decorated in front at quill end with six tiers of small red parakeet feathers (2), commencing above and ending below with transverse lashing turns (3). f, back view, showing spiral turns of thread to keep red feathers in position; quill so covered is 41 mm. long.

page 361 Oro, the principal god of the Society Islands (77, pp. 462, 464). The identification as Mauke, therefore, cannot be accepted simply from the catalog ascription. The only method left for confirmation or disagreement is analysis of the technique used. The object, and the technique of the sennit and feather decoration are shown in figure 229. The feathers used were split black, tropic bird tail feathers and red and yellow parakeet feathers.

The decoration of the quills of tropic bird tail feathers with small red feathers is a common practice in the neighboring islands and the addition of extra tiers is merely a matter of degree and not a new principle. An identical technique occurs in Tahiti. The use of sennit to form a textile cuff for wrapping around a staff, however, is something entirely new and has not been recorded for the neighboring islands of Atiu and Mitiaro or for the other islands of the group, though Aitutaki shares certain carving motifs with Atiu and Mitiaro. From the available study material, I should say that the cuff technique on the god under discussion is entirely foreign to the Cook Islands. For further discussion as to the origin of this unique object, see pages 469-473.

Figure 230.—God attributed to Mauke (British Mus., L.M.S., 66 or 99): cylindrical body (1), composed of thick five-ply coir braid, dark red in color, folded into lengths of 100 mm. to reach a diameter of 30 mm.; evidently unbraided coir fiber, which projects at one end, was also added inside to thicken body; braid foundation wrapped with white bark cloth and tied at each end with fine oronga thread; split black feathers tied in bunches and seized with single coir fiber added at each end; upper end ornamented also with three red tropic bird feathers.

Figure 230.—God attributed to Mauke (British Mus., L.M.S., 66 or 99): cylindrical body (1), composed of thick five-ply coir braid, dark red in color, folded into lengths of 100 mm. to reach a diameter of 30 mm.; evidently unbraided coir fiber, which projects at one end, was also added inside to thicken body; braid foundation wrapped with white bark cloth and tied at each end with fine oronga thread; split black feathers tied in bunches and seized with single coir fiber added at each end; upper end ornamented also with three red tropic bird feathers.

The second god attributed to Mauke is labeled, "idol or fetish" and also "Tangaroa from Mauke (?)." The note of interrogation as well as the correct spelling of Mauke raises doubt as to the locality. (See figure 230.)

page 362

The use of folded sennit to form a body has been described for neighboring Atiu. There is nothing special in the technique that can be used for or against its attribution to Mauke. A similar technique also occurs in Mangaia (fig. 232) and in some unlocalized specimens (figs. 249 and 250). The object may therefore be accepted as belonging to the Cook Islands—probably to Mangaia but possibly to Mauke, for the missionaries who collected from the other islands were not likely to omit Mauke.

Mangaian Gods

The names of the gods of Mangaia have been recorded by Gill (33, pp. 331-333). Material representatives of 13 gods worshipped by the various tribes were kept in a national god house named Te Kai'ara, situated between the important temples of 'Akaoro and Ara'ata in the Kei'a district (76, p. 172). A caretaker looked after the god house, changed the clothing of the gods (white thick cloth, tikoru), and cooked food for them every evening. Neither the gods nor their caretaker were disturbed during the changing fortunes of intertribal war.

During the exile of the skilled carver, Rori, the god house was accidently burned down and its contents destroyed. Rori subsequently emerged from exile and was commissioned to carve a set of gods for the new god house. Rongo was represented by a shell trumpet, but the remaining 12 gods are said to have been carved from wood. They were Motoro, Tane-papa-kai, Tane-ngaki-au, Tane-i-te-'utu, Tane-kio, Te A'io, Te Kura'aki, Utakea, Turanga, Teipe, Kereteki, and Tangiia. Rori carved all these except his own god, Teipe, which was carved by his friend Tapaivi. A deified ancestor named Vaeruarau also was carved in wood by Rori and placed in the national god house. The image was removed because of sickness among his followers, and he was hidden in the rocks by some who remained faithful (33, pp. 86, 87).

The London Missionary Society established the Tahitian missionary, Davida, in Mangaia in 1823. Gill (33, pp. 333, 334) recounts what happened to the Mangaian gods as follows:

Soon after the introduction of Christianity it happened that one of the regal family was taken dangerously ill. Whilst the heart was softened by affliction, the parent was struck with the glaring absurdity of professed adherents to Christianity keeping up this idol-shrine with its daily oven of food. What should be done with the idols? It was decided by the king, Parima, Simeona, and other leading men of the day, to surrender these dangerous things to Davida. To the horror of the heathen, but to the great joy of the Christian party, the whole thirteen were carried in triumphal procession to the house of Davida by the sea. The wrappings were thrown away, and for the first time since they were carved by Rori they were exposed to the vulgar gaze. Soon after Messrs. Williams and Platt paid an opportune visit to the island; and they proceeded to Raiatea laden with page 363these idols, which were eventually deposited in the museum of the London Missionary Society, and are now in the British Museum.

The day the idols were removed, the house in which they had been kept was set on fire; the maraes all over the island were desecrated, the little houses in which the deity was supposed to be invisibly present were burnt, the great stone idol of Rongo at the sea-side, where human sacrifices were offered, was smashed to atoms, and (what is much to be regretted) the magnificent native mahogany (tamanu) trees were set on fire, on account of their supposed connection with idolatry.

As the tracing of these gods and the material of which they were made is important, I quote further details from Gill concerning individual gods.

Wooden Gods

Motoro. Next to Rongo, Motoro, the god of the ancient Ngariki tribe, was the most powerful. Gill (33, p. 246) says, "That idol named Motoro, is a rude representation of the human form, carved in ironwood." In writing of Mangaia, Gill (33, p. 204) says, "Rori obtained some of her [an aunt of the carver Rori] beautiful hair to adorn the then newly carved image of Motoro. At the period of the surrender of the idols to Mr. Williams the hair of this woman was still on it."

Red feathers brought from Tahiti by Ume were also used for decoration according to Gill (33, p. 235). "Beautiful red parakeet feathers, brought to this island by his grandfather, and concealed with Rori's other treasures in the raei, were used by him to adorn Motoro, to the great admiration of men of that day."

Some Aitutaki men, after collecting red feathers at Manuae, were wrecked on Mangaia and gave some feathers to their hosts. Gill (33, p. 242) says: "The beautiful red feathers presented to those who had so kindly entertained them were collected and put on their god Motoro. When that idol was given up to Mr. Williams these identical red feathers adorned it."

The place of Motoro in the god house is thus indicated by Gill (28, p. 28):

Motoro, Kereteki, and Utakea were represented by iron-wood idols in the god-house of the king. On entering that rude reed hut, the dwelling place of the chief divinities of Mangaia, the first idol was Rongo, in the form of a trumpet shell; next came the honoured Motoro, the guide of daily life; then came Tane and ten other objects of worship; amongst which were Kereteki and Utakea.

As to Motoro's later abode, Gill (31, p. 110) says, "… Motoro, whose ugly form may be seen in the Museum of the London Missionary Society."

Tane. Of Tane's position in the god-house, Gill (28, p. 28) says: "The iron-wood idol called Tane merely, was asserted to represent the fifth son of Vatea; and yet was only third in order of dignity.… Of the innumerable objects of fear and worship, only thirteen were admitted to the honour of a place in this rude Pantheon as national gods."

page 364

Gill (28, p. 107) continues:

Amongst the thirteen principal gods of Mangaia which at the establishment of Christianity were surrendered to the Missionaries were four bearing the name of Tane. They were simply pieces of iron-wood carved roughly into human shape, once well wrapped up in numerous folds of the finest native cloth. Of these four Tanes three—Tane-Ngakiau, Tane-i-te-ata, and Tane-kio—were considered to be inferior to the first, who was usually called Tane, sometimes, however, Tane Papa-kai, i.e. Tane-piler-up-of-food. In order of rank Tane came after Rongo and Motoro, the chief deities of Mangaia.

In this list, Tane-i-te-ata is substituted for Tane-i-te-'utu of the first list. Concerning Tane-Ngakiau, Gill (28, p. 31) says: "This uncomfortable god had a carved ironwood form, and was one of the thirteen gods of Mangaia now in the Museum."

Te Kuraaki. Of this god, Gill (28, p. 31) says: "Yet the carved ironwood idol remained in the Pantheon until 1824, when it was surrendered to Messrs. Williams and Platt."

Tiaio [Te A'io]. Gill (28, p. 30) says: "Rori's life was spared by Manaune, expressly that he might carve the rough iron-wood representation of Tiaio, which, with the rest, now quietly reposes in the Society's museum."

Utakea. A native convert, reported by Gill (31, p. 105), spoke of "our god being Utakea, long since taken away with the rest by "Wiliamu" (Rev J.Williams)."

Teipe. In addition to the god carved by Rori's friend to form one of the 13 deposited in the god-house, another kept in the curtained off part of the house of the priest Keu, is alluded to by Gill (33, p. 153) as "this uncouth ironwood god."

Tangiia. Gill (28, p. 28) says, "Tangiia, the fourth son of Vatea, was the last in regard to dignity and order."

Speaking of Tangiia, the Rarotongan ancestor, Gill (28, p. 24) states: "It is needful to distinguish this Tangiia, who is unquestionably an historical character, from the mythical Tangiia descended from Vatea, and one of the gods of Mangaia, whose iron-wood form is deposited in the museum of the London Missionary Society."

It is evident from Gill's accounts that all 13 gods were given to Mr. Williams and sent to the London Missionary Society Museum in London. Four of them-Motoro, Tane-Ngakiau, Tiaio (Te A'io), and Tangiia-are definitely mentioned as being in the Missionary Museum. The material transferred from the Missionary Museum to the British Museum in 1890 is contained in one large case. There are only three wooden artifacts labeled "District gods from Mangaia", and unfortunately their names were not recorded. They consist of an upper carved slab, a rounded shaft, and an expanded base (pl. 14, F-H). In form and carving they are very similar; and, in all probability, they are some of those made by Rori. However, by no stretch of imagination can they be regarded as "a rude representation of the human form." One wonders what page 365became of the other nine, carved in ironwood, that constituted the 12 wooden gods of the Mangaian Pantheon.

Priests kept a representative of their particular gods in a screened-off part of their dwelling houses (33, p. 153). Various groups had family gods of lesser note. Some were natural objects such as stones, as the great god Rongo was represented in the god-house by a shell trumpet. Other religious symbols were formed of coconut leaf, feathers, human hair, sennit, and tapa.

The specimens available for study consist of simple objects, compound objects, a spiked wooden artifact, carved slabs, and ceremonial adzes.

Water-worn Stone

A water-worn dark stone, rounded in shape, collected by Gill, and deposited in the British Museum, bears the following inscription: "God from Mangaia, answering to the household gods of the Romans but not usually kept in the house—in some shady, secret place." This was evidently a family god and it was figured by Edge-Partington (24, I-17-5). (See figure 231, a.)

Figure 231.—Mangaian religious symbols. a, stone symbol (British Mus., 9939): length, 88 mm.; width, 79 mm.; height, 60 mm. b, palm-leaf symbol (British Mus., 9940): tip of coconut leaf with seven leaflets; length, 425 mm.; a piece of flat sennit, 4 mm. wide, with four fibers in each ply, is tied around leaflets with a reef knot forming two loops like a bow tie.

Figure 231.—Mangaian religious symbols. a, stone symbol (British Mus., 9939): length, 88 mm.; width, 79 mm.; height, 60 mm. b, palm-leaf symbol (British Mus., 9940): tip of coconut leaf with seven leaflets; length, 425 mm.; a piece of flat sennit, 4 mm. wide, with four fibers in each ply, is tied around leaflets with a reef knot forming two loops like a bow tie.

Palm-leaf Symbol

Gill (33, pp. 149, 150) writes as follows: "In heathenism no canoe ever ventured over the reef at Mangaia to fish without first fastening to its bows the fisherman's god. This consisted merely of the extremity of a coco-nut frond secured with fine-plaited sennit tied into a bow." It was the symbol of the god Mokoiro, and the hereditary office of Ruler of Food (76, p. 117) was combined with that of the priests of Mokoiro. It was part of the office to equip each canoe in the fishing season (September to December) with this symbol. Sometimes a fleet of 200 canoes assembled, and each was provided with a symbol prepared beforehand. The priests distributed the symbols and then gave the signal to start. A specimen collected by Gill and now in the British Museum is shown in figure 231, b. It was figured by Edge-Partington (24, I-1-1).

page 366
Tapa Cloth Symbols

An alternate form of fishing charm, representing Mokoiro, was distributed by the Ruler of Food (76, p. 146). It consisted of bark cloth (autea) folded into cones termed poani (plug) of which four were placed in different parts of the fishing canoe to plug up the four quarters from which the winds blew.

Tapa and Feather Gods
Tapa symbols were formed into gods with a body formed of folds of sennit, covered with tapa, and decorated with feathers. One in the Cambridge University Museum, with arms and legs roughly indicated, bears an old label with the following inscription: "Tangua a mighty god from Mangaia. Said
Figure 232.—Tapa and feather gods, Mangaia. a: length of body and legs, 330 mm.; middle body width, 100 mm.; two arms (1, 1), 50 and 60 mm. long; legs (2, 2): body of folded lengths of sennit and wrapped with white tapa and tied with spiral turn of fine thread; tropic bird tail feathers kept in position by thread lashings (Cambridge Univ. Mus., Z.6094). b: smaller than Cambridge figure and without arms; length of body and legs 142 mm.; body width 22 to 24 mm.; legs (2, 2), 36 and 34 mm. long and 11 and 10 mm. thick; body formed of sennit folds and wrapped with continuous strip of white tapa which covers up end in transverse turn, descends in oblique turns, and divides at bottom to wrap each leg separately; bunches of split black feathers seized with single coir fiber, arranged around top outside tapa, and tied with turns of fine oronga thread; red and green feathers added to body in layers from above down with spiral thread; small black feathers on body at back; tropic bird feathers attached at sides to lean forward; tufts of small red feathers tied to legs (British Mus., L.M.S., no number). c, formation of body of b with thick sennit, 9 mm. wide, doubled three times at top (1) and twice at bottom of body, with two free ends (2) projecting down to form legs.

Figure 232.—Tapa and feather gods, Mangaia. a: length of body and legs, 330 mm.; middle body width, 100 mm.; two arms (1, 1), 50 and 60 mm. long; legs (2, 2): body of folded lengths of sennit and wrapped with white tapa and tied with spiral turn of fine thread; tropic bird tail feathers kept in position by thread lashings (Cambridge Univ. Mus., Z.6094). b: smaller than Cambridge figure and without arms; length of body and legs 142 mm.; body width 22 to 24 mm.; legs (2, 2), 36 and 34 mm. long and 11 and 10 mm. thick; body formed of sennit folds and wrapped with continuous strip of white tapa which covers up end in transverse turn, descends in oblique turns, and divides at bottom to wrap each leg separately; bunches of split black feathers seized with single coir fiber, arranged around top outside tapa, and tied with turns of fine oronga thread; red and green feathers added to body in layers from above down with spiral thread; small black feathers on body at back; tropic bird feathers attached at sides to lean forward; tufts of small red feathers tied to legs (British Mus., L.M.S., no number). c, formation of body of b with thick sennit, 9 mm. wide, doubled three times at top (1) and twice at bottom of body, with two free ends (2) projecting down to form legs.

page 367 to be the progenitor of the present inhabitants and deities. Human sacrifices were offered." It was collected by John Williams. Tangua did not figure in the Mangaian list but an examination of the museum entry book showed that the name was Tangiia. Tangiia, the fourth son of Vatea, had a place in the god house; but, as he was made of ironwood, the tapa figure was evidently another form or its identification is due to a misplaced label. The person who wrote the label evidently confused Tangiia the god with Tangiia the Rarotongan ancestor, a mistake that Gill warned about (p. 364).

A similar tapa and feather bundle, without arm projections, is in the London Missionary Society case but no exact locality is given. There is little doubt that the two objects came from the same island, probably Mangaia. The followers of Tangiia were practically exterminated in the tribal wars. Though an ironwood form of the god could survive in the god house, it is doubtful whether a tapa and sennit bundle would have lasted in such a fresh condition through the years after the tribe became extinct. Probably the bundles are minor family gods and the label of the real ironwood Tangiia was applied to the Cambridge specimen by mistake. (See figure 232.)

Sennit Roll

The direct representation of Mokoiro consisted of a small roll of sennit which was hidden in a cave. Gill (33, p. 31) states: "It had no human likeness whatever, and therefore found no place in the king's idol-house."

The god Tane-kio was originally represented by finely plaited coconut fiber. Gill (33, pp. 56, 57) recounts how the god was expelled from Tahiti because a number of his followers wasted away. His priest, Ue, sealed the symbol in a coconut shell and set it adrift in the sea to seek a new home. Ue followed by canoe and located his god in Mangaia, where he built the marae of Maungaroa to him (76, p. 168). Later, a representation of Tane-kio was carved in wood by Rori and lodged in the national god house. Possibly the sennit roll referred to by Gill was a sennit foundation.

Triton Shell

Speaking of the place in the god house of Rongo, the principal god of Mangaia, Gill (33, pp. 331, 332) says: "His chief representative was a triton shell, used only by the king, and deposited at the entrance. It is now preserved in the London Missionary Society case at the British Museum."

The Triton shell was in the form of a trumpet, as it was used by the king to summon his warriors together. Two shell trumpets attributed to Mangaia are in the Missionary case and though there is no information concerning them, one of them is probably the symbol of Rongo. (See figure 167.)

page 368
Wooden Spiked God
A god in the Missionary case in the British Museum resembles an unwieldy weapon bristling with spikes and ornamented with human figures. It has been figured by Read (56, pl. 13, F) and by Edge-Partington (24, II-8-3). Read states that the name and locality are lost and Edge-Partington attributes it to eastern Polynesia. Somehow, I mistook the locality for Mangaia, and I hoped that because of the human figures, human hair, and red feather decoration that it might be Motoro. However, the technique is so different from that used by Rori on the three district gods that its locality remains doubtful. Having
Figure 233.—Spiked god (British Mus., L.M.S., 287), upper section. a, front: 1, upper row of five straight points, one broken; top width 168 mm.; 2, horizontal bar carrying upper and lower row of five curved spikes, some broken, width 155 mm., thickness 97 mm.; 3, second transverse bar connected at sides and middle with upper bar (2) and carrying heads of three spaced figures; heads with shaped brows and nose, eyes and mouth formed by rectangular depressions, heads pointed except middle one; 4, bodies with pendent arms poorly defined and straight legs which fuse with bar below; 5, lowest bar which supports figures and also two rows of six divergent curved spikes, width 147 mm., thickness 109 mm.; 6, neck formed by narrowing in from sides; showing turns of sennit braid tied around it, width 76 mm., thickness 54 mm.; 7, pair, of front spikes curved at ends and framing upper part of single figure (8) with similar, head to upper figures. b, back: similar in all respects to front but one of upper figures (3) has head broken off; 6, mass of braid material tied to neck; 7, downward point from curved spike; 8, figure similar to one on front; 9, legs of single figure. c, left side: numbering same as a: bodies of upper row of figures (4) separated by width of bars, lower single figure in profile; length from top spikes to neck 400 mm.

Figure 233.—Spiked god (British Mus., L.M.S., 287), upper section. a, front: 1, upper row of five straight points, one broken; top width 168 mm.; 2, horizontal bar carrying upper and lower row of five curved spikes, some broken, width 155 mm., thickness 97 mm.; 3, second transverse bar connected at sides and middle with upper bar (2) and carrying heads of three spaced figures; heads with shaped brows and nose, eyes and mouth formed by rectangular depressions, heads pointed except middle one; 4, bodies with pendent arms poorly defined and straight legs which fuse with bar below; 5, lowest bar which supports figures and also two rows of six divergent curved spikes, width 147 mm., thickness 109 mm.; 6, neck formed by narrowing in from sides; showing turns of sennit braid tied around it, width 76 mm., thickness 54 mm.; 7, pair, of front spikes curved at ends and framing upper part of single figure (8) with similar, head to upper figures. b, back: similar in all respects to front but one of upper figures (3) has head broken off; 6, mass of braid material tied to neck; 7, downward point from curved spike; 8, figure similar to one on front; 9, legs of single figure. c, left side: numbering same as a: bodies of upper row of figures (4) separated by width of bars, lower single figure in profile; length from top spikes to neck 400 mm.

page 369 written up the object in detail, I include the description here for the purposes of record.

The artifact may be divided into an upper specialized section with human figures and a long shaft with transverse sets of curved spikes. The total length is 2,886 mm., and it is carved from one piece of wood. The lower end has been cut off so that originally it was longer. The upper section with human figures and spikes ends in a neck to which is attached a composite cord carrying a strange assortment of shells and pieces of coral. Below the neck, single human figures and curved spikes commence the upper end of the shaft (fig. 233). The upper end of the shaft, with single human figures back and front and with lateral spikes is shown in figure 234.

Figure 234.—Spiked god, uppermost single human figure. a, front: human figure carved on front of staff with head (1) showing details similar to upper row of figures, with short downward knobs (2) for arms and two projecting bent legs (3); upper part of figure framed by two long incurved spikes (4); upper pair of front spikes (5, 5) curved upward and lower pair (6, 6) curved downward. b, back: figure of same form as front figure and showing posterior upper (5) and lower (6) spikes. c, left side: showing lateral incurved spikes (4, 4) and heads (1) projecting from staff with pointed heads curving upward toward curved spikes above; pointed chins also point downward like spikes; upper limbs (2, 2) project well forward to balance upper lateral spikes (5, 5); abdomen (7, 7) projects forward; legs (3, 3) form forward and downward projections which fuse with spikes (8, 8) which balance lower lateral spikes (6, 6).

Figure 234.—Spiked god, uppermost single human figure. a, front: human figure carved on front of staff with head (1) showing details similar to upper row of figures, with short downward knobs (2) for arms and two projecting bent legs (3); upper part of figure framed by two long incurved spikes (4); upper pair of front spikes (5, 5) curved upward and lower pair (6, 6) curved downward. b, back: figure of same form as front figure and showing posterior upper (5) and lower (6) spikes. c, left side: showing lateral incurved spikes (4, 4) and heads (1) projecting from staff with pointed heads curving upward toward curved spikes above; pointed chins also point downward like spikes; upper limbs (2, 2) project well forward to balance upper lateral spikes (5, 5); abdomen (7, 7) projects forward; legs (3, 3) form forward and downward projections which fuse with spikes (8, 8) which balance lower lateral spikes (6, 6).

The staff is rectangular and four-sided with single figures on the back and front and balanced on the sides with two pairs of long incurved spikes with their bases opposite the heads of the figures, whereas opposite the lower part of the body the sides are balanced by upper and lower pairs of diverging spikes.

Below the upper single figures the two pairs of diverging lateral spikes are continued, and on the same level single pairs of diverging spikes take the place page 370of figures on the front and back. After a number of these sets of spikes, a middle set of front and back human figures displaces two sets of single diverging spikes. The middle set of figures is followed by three sets of spikes, which are followed by a set of front and back figures. (See figure 235.)

Figure 235.—Spiked god, middle and lower single figures. a, front middle figure: details of figure same as upper but head (1) balanced by lateral upper pair (2) whose points balance upper end of pointed head and lateral lower pair (3) whose points are somewhat balanced by pointed chin; upper and lower lateral pairs balance lower part of body with upper points (5) somewhat balanced by pointed arm projections (4) while lower points (7) are balanced as before by points (6) below legs. b, left side of lowest figure: details same as a, but balance of upper and lower points of head (1) with lateral spikes (2, 3) more obvious while same applies to lower part of body. c, lowest plain section: showing sets of spikes on same level.

Figure 235.—Spiked god, middle and lower single figures. a, front middle figure: details of figure same as upper but head (1) balanced by lateral upper pair (2) whose points balance upper end of pointed head and lateral lower pair (3) whose points are somewhat balanced by pointed chin; upper and lower lateral pairs balance lower part of body with upper points (5) somewhat balanced by pointed arm projections (4) while lower points (7) are balanced as before by points (6) below legs. b, left side of lowest figure: details same as a, but balance of upper and lower points of head (1) with lateral spikes (2, 3) more obvious while same applies to lower part of body. c, lowest plain section: showing sets of spikes on same level.

The lowest figure was followed by five sets of spikes where the staff was 34 by 36 mm. in two thicknesses. Here the four surfaces began to narrow, and at a distance of 42 mm. the two thicknesses had diminished to 28 mm. each. The plain surfaces below the last spikes were incised with parallel oblique lines. Evidently a long plain shaft had been sawn off.

It is obvious from the figures that the pointed vertices and chins of the figure were so carved as to balance the points of neighboring curved spikes which formed the dominant motif of the artifact. The arms and legs were similarly affected.

The composite cord with its ornaments after Edge-Partington (24, II-8-3) and methods of attaching feather holders under seizing turns is shown in figure 236. It is evident, from the great number of feather holders on the terminals and even in the main cords, that the feather decoration must have helped considerably in making the spike god an imposing sight.

page 371
Figure 236.—Composite cord with shell and coral ornaments. a, composite cord (1) of number of lengths of sennit tied around neck (2) of spike god and projects out from back; sennit lengths kept together by close seizing for 76 mm., when three broken ends (3) issue from beneath seizing which is 30 mm. thick; seizing continues for 57 mm., when three more short ends (4) were freed, two of them amalgamated to form one cord, thickness of seizing at 4 diminished to 23 mm.; seizing below 4 lessened in thickness to 20 mm. and 68 mm. farther on, six cords liberated, two single and two pairs formed by dividing two thicker cords, one (5) which is short, tied around a shell (6); main cord (7) continues for 370 mm., upper part seized with braid 4 mm. wide and lower part with cord 2 mm. wide, end finished off in loops (8) 300 mm. long; other cords or braids support shells (9) which were tied to cords by a separate transverse lashing while others support pieces of coral (10); some of cords supporting heavy shells snapped off and ends retied; six tufts of hair with seized eyelet holes, five frizzy (11) and one straight (12) were tied to various parts with what appeared to be trade string. b, straight terminal of some cords, seized with feather holders of a doubled coir fiber (1) with spiral tie (2) of single coir fiber in position, but feather dropped out. c, looped terminal of some cords with feather holders as in b. d, some feather holders of single coir fiber (1) with spiral tie (2) in position, whereas others (3) tie and feathers have disappeared. e, spiral tie of single fiber showing buried commencement end (1) with spiral turns (2) and finishing end (3) twisted over last turn and pushed down under it.

Figure 236.—Composite cord with shell and coral ornaments. a, composite cord (1) of number of lengths of sennit tied around neck (2) of spike god and projects out from back; sennit lengths kept together by close seizing for 76 mm., when three broken ends (3) issue from beneath seizing which is 30 mm. thick; seizing continues for 57 mm., when three more short ends (4) were freed, two of them amalgamated to form one cord, thickness of seizing at 4 diminished to 23 mm.; seizing below 4 lessened in thickness to 20 mm. and 68 mm. farther on, six cords liberated, two single and two pairs formed by dividing two thicker cords, one (5) which is short, tied around a shell (6); main cord (7) continues for 370 mm., upper part seized with braid 4 mm. wide and lower part with cord 2 mm. wide, end finished off in loops (8) 300 mm. long; other cords or braids support shells (9) which were tied to cords by a separate transverse lashing while others support pieces of coral (10); some of cords supporting heavy shells snapped off and ends retied; six tufts of hair with seized eyelet holes, five frizzy (11) and one straight (12) were tied to various parts with what appeared to be trade string. b, straight terminal of some cords, seized with feather holders of a doubled coir fiber (1) with spiral tie (2) of single coir fiber in position, but feather dropped out. c, looped terminal of some cords with feather holders as in b. d, some feather holders of single coir fiber (1) with spiral tie (2) in position, whereas others (3) tie and feathers have disappeared. e, spiral tie of single fiber showing buried commencement end (1) with spiral turns (2) and finishing end (3) twisted over last turn and pushed down under it.

Carved Slabs with Pedestals

The three "district gods of Mangaia, names now lost" in the Missionary case in the British Museum are thick wooden slabs with projecting ridges carved into pillars and arches and pierced between the ridges with lozenge-shaped holes. Below the carved upper sections, the slabs diminish to rounded shafts, which terminate in expanded carved pedestals. (See plate 14, F-H.)

As the districts of Mangaia were occupied by different tribes with their own tribal gods, these figures represent three of the tribal gods taken from the national god house in the district of Kei'a and given to Williams and Platt. page 372They are survivors of the 11 tribal gods carved by the master craftsman, Rori. Though the original coverings of thick white tapa were torn off by the zealous Mangaian converts to Christianity, the gods are well preserved. The sennit and feather ornamentation of the shafts still remains, but there is no trace of the feathers that were originally attached to the carved arches.

Figure 237.—Technique of carved slabs, Mangaia. a, cross section of rectangular slab; shaded sections regularly spaced are parts to be removed. b, cross section through lozenge holes; with shaded parts in a removed, slab divided into four ridges (1) in front and back and two (2) at ends; shaded portions show sections through middle of fore-and-aft lozenge holes (3) and end holes (4) between two end ridges which run into adjacent fore-and-aft holes; ridges average about 22 mm. in width. c, upright view: showing longitudinal ridges (1) and end ridges (2); ridges are square-cut and of even thickness. d, notches (1) cut in ridges and so spaced as to leave alternating smaller (2) and larger (3) rectangular spaces on outer surface of ridges; smaller segments formed solid pillars while larger sections were perforated (4) from side to side to form arches (5); notches on all ridges cut on same level so as to form horizontal series of pillars and arches as well as vertical sets; lozenge-shaped holes (6) between ridges obscured by angle of drawing, but top section through holes shows three fore-and-aft holes and end holes as in b.

Figure 237.—Technique of carved slabs, Mangaia. a, cross section of rectangular slab; shaded sections regularly spaced are parts to be removed. b, cross section through lozenge holes; with shaded parts in a removed, slab divided into four ridges (1) in front and back and two (2) at ends; shaded portions show sections through middle of fore-and-aft lozenge holes (3) and end holes (4) between two end ridges which run into adjacent fore-and-aft holes; ridges average about 22 mm. in width. c, upright view: showing longitudinal ridges (1) and end ridges (2); ridges are square-cut and of even thickness. d, notches (1) cut in ridges and so spaced as to leave alternating smaller (2) and larger (3) rectangular spaces on outer surface of ridges; smaller segments formed solid pillars while larger sections were perforated (4) from side to side to form arches (5); notches on all ridges cut on same level so as to form horizontal series of pillars and arches as well as vertical sets; lozenge-shaped holes (6) between ridges obscured by angle of drawing, but top section through holes shows three fore-and-aft holes and end holes as in b.

A similar figure is in the Peabody Museum, Cambridge, Massachusetts, but many of the arches are broken off and the shaft is bare of decoration. A figure in the Oldman collection differs from the other four in the treatment of arches and carving. The technique by which the upper carved part was formed is shown in figure 237. The notching of the ridges into pillars and arches formed small rectangular areas that were carved by further notching to form geometrical patterns. The pillars and arches are better seen in figure 238.

The four similar slabs follow a common technique as regards the form of arch and pillar, their arrangement in horizontal series at front and back and ends, and the lozenge-shaped holes on the same level as the arches. The tops of the pillars and arches are all carved with combinations of bar and chevron motifs, the pattern being the same on each slab but varying slightly in different page 373slabs. All have four front and back ridges except one (L.M.S. coll., 30) which has three. All have two ridges at the ends except the largest (L.M.S. coll., 32), which has three and, consequently, two sets of lozenge-shaped holes in the two inter-ridge spaces. The arches in each ridge range from 7 to 10. Dimensions in millimeters of the carved slabs are shown in table 3.

Figure 238.—Carving of pillars and arches (British Mus., L.M.S., 31). a, front: pillars and arches show up in side view of end arch; 1, 1, spaced notches; 2, pillars; 3, arch section; 4, arch perforation; 5, actual arch; 6, lozenge-shaped hole; three sets of holes between four front ridges, each horizontal set on same level as arches; four front ridges (7) shown with geometrical carving, smaller areas on pillars and larger on arches. b, end, showing carving on two end ridges (8, 8) and side views of neighboring front and back ridges (7, 7); carving on pillar (9) forms a bar and chevron which unite to form a K, while opposite edge of pillar space defined by a bar notched in middle; in wider space above arch (10) pattern is bisymmetrical with a notched bar at each end and a free chevron separated from the bar with the open angles toward each other; end holes (6) also opposite the arches and communicate with nearest fore-and-aft holes.

Figure 238.—Carving of pillars and arches (British Mus., L.M.S., 31). a, front: pillars and arches show up in side view of end arch; 1, 1, spaced notches; 2, pillars; 3, arch section; 4, arch perforation; 5, actual arch; 6, lozenge-shaped hole; three sets of holes between four front ridges, each horizontal set on same level as arches; four front ridges (7) shown with geometrical carving, smaller areas on pillars and larger on arches. b, end, showing carving on two end ridges (8, 8) and side views of neighboring front and back ridges (7, 7); carving on pillar (9) forms a bar and chevron which unite to form a K, while opposite edge of pillar space defined by a bar notched in middle; in wider space above arch (10) pattern is bisymmetrical with a notched bar at each end and a free chevron separated from the bar with the open angles toward each other; end holes (6) also opposite the arches and communicate with nearest fore-and-aft holes.

Table 3.—Measurements of Carved Slabs
L.M.S. 30 L.M.S. 31 L.M.S. 32 Cambridge , 1354
total length 857 900 1,385 1,130
upper section length 390 330 645 518
upper section top width 131 205 145
upper section thickness 83 138 75
upper section bottom width 91 114 100
upper section bottom thickness 99 75
page 374

All the slabs diminish in width and thickness from the top down, but, the bottom arches and pillars expand in both diameters to more nearly approach the dimensions of the rounded shaft below them. Variations in the top and bottom of the carved section are shown in figure 239. The top end of the Cambridge figure is identical with figure 239, c, but only one of the free arches is intact. The lower end is similar to figure 239, e.

Figure 239.—Tops and bottoms of carved slabs. a, top front (L.M.S., 30): carving commences with arches (1) above which ridges (2) are cut down for short distance to top; top border plain with slight concavities (3) between ridges (4); this slab has three front ridges (4), middle ridge shown without pillars and arches. b, top front (L.M.S., 31): similar to a with four front ridges (4). c, top front (L.M.S., 32): above top complete arch (1), a pillar (2) projects beyond outer plane of ridge, with free arch (3) projecting upward and terminating in a pillar (4) with an inner projecting free point (5); all ridges similarly treated and both pillars with arch carved. d, bottom front of a: no outward projection of arches beyond general plane; feather decoration (1) projecting upward from top end of shaft. e, bottom front of b: two arches (1, 2), without intervening pillar, project beyond plane of last normal arch (3) also without intervening pillar; thickened part denned by two lowest arches (1, 2) and lozenge-shaped holes in ridge interspaces of two lowest arches do not pierce through to opposite side, whereas lozenges (4) opposite last normal arch (3) go right through; transverse seizing (5) of shaft with feather holders (6) denuded of feathers projecting upward. f, end view of e, showing greater projections of two lowest front and back arches (1, 2) as compared with lowest normal arch (3); two end ridges with intervening lozenges (4), lowest two not pierced through. g, bottom front of c, showing outward projection of two lowest pillars (1, 1) and intervening arch (2) which define thickened lower part of slab: four front ridges with three sets of lozenge-shaped holes (4) of which those in thickened section on level of lowest arch (2) are not pierced through. h, end view of g: outward projection of lowest pillars (1, 1) and arch (2), inter-ridge lozenges on level of lowest cut (2) not pierced through.

Figure 239.—Tops and bottoms of carved slabs. a, top front (L.M.S., 30): carving commences with arches (1) above which ridges (2) are cut down for short distance to top; top border plain with slight concavities (3) between ridges (4); this slab has three front ridges (4), middle ridge shown without pillars and arches. b, top front (L.M.S., 31): similar to a with four front ridges (4). c, top front (L.M.S., 32): above top complete arch (1), a pillar (2) projects beyond outer plane of ridge, with free arch (3) projecting upward and terminating in a pillar (4) with an inner projecting free point (5); all ridges similarly treated and both pillars with arch carved. d, bottom front of a: no outward projection of arches beyond general plane; feather decoration (1) projecting upward from top end of shaft. e, bottom front of b: two arches (1, 2), without intervening pillar, project beyond plane of last normal arch (3) also without intervening pillar; thickened part denned by two lowest arches (1, 2) and lozenge-shaped holes in ridge interspaces of two lowest arches do not pierce through to opposite side, whereas lozenges (4) opposite last normal arch (3) go right through; transverse seizing (5) of shaft with feather holders (6) denuded of feathers projecting upward. f, end view of e, showing greater projections of two lowest front and back arches (1, 2) as compared with lowest normal arch (3); two end ridges with intervening lozenges (4), lowest two not pierced through. g, bottom front of c, showing outward projection of two lowest pillars (1, 1) and intervening arch (2) which define thickened lower part of slab: four front ridges with three sets of lozenge-shaped holes (4) of which those in thickened section on level of lowest arch (2) are not pierced through. h, end view of g: outward projection of lowest pillars (1, 1) and arch (2), inter-ridge lozenges on level of lowest cut (2) not pierced through.

page 375

The shaft extending from below the lowest arches to the expanded pedestal is round in all four artifacts, with the diameter ranging from 44 to 56 mm. The Cambridge shaft is bare but the other three are covered with sennit. In two, the sennit is wound in close transverse turns throughout their length but the largest (L.M.S. coll., 32) has close transverse turns around the upper end of, the shaft and the beautiful multiple lozenge pattern termed inaere for the rest in the same technique used on house rafters (fig. 18). At different levels, feathers, human hair, and coir holders from which the feathers have fallen out were introduced in rows and fixed by the transverse turns of sennit which pass over and conceal their lower ends. (See figure 240.)

Figure 240.—Decoration of shafts (L.M.S., coll.: a,30; b,31; d,32): a, shaft seized with flat sennit 2.5 mm. wide for 45 mm. and bunches of split black feathers seized with single coir fiber caught under turn of shaft seizing; feathers, including occasional tropic bird white tail feathers, directed upward as in figure 239, d; three succeeding rows spaced successively 61, 58, and 69 mm. apart with effect shown (1); lowest feather row (2); 16 mm. below, two tufts of human hair (3) caught under seizing; at lower end whole black feathers in two rows pointing upward and downward with quills together were tied in position over shaft seizing by four turns of sennit (see fig. 241, a). b, cord holders (1) projecting 44 mm. above seizing turn of fine braid in four groups of 5 or 6 holders, succeeded lower down by short single holders (2); seizing braid changes to coarser, 3 mm. wide, which lower down fixes 4 sets of 7 holders (3). c, lower end of b: lower sennit-covered end of shaft covered by turns of doubled thick braid 5 mm. wide, wrapped with human hair of brownish color (1); top end buried under subsequent turns and lower end (2) pushed through under last turn; pedestal (3). d, lower end of shaft with multiple lozenge inaere design (1) worked in fine sennit; upper end (not shown) with sennit seizing supporting coir cord holders in groups of four; four lengths of thick sennit 4 mm. wide were seized together with human hair and looped around upper part of shaft, hair on hanging part had evidently unwound.

Figure 240.—Decoration of shafts (L.M.S., coll.: a,30; b,31; d,32): a, shaft seized with flat sennit 2.5 mm. wide for 45 mm. and bunches of split black feathers seized with single coir fiber caught under turn of shaft seizing; feathers, including occasional tropic bird white tail feathers, directed upward as in figure 239, d; three succeeding rows spaced successively 61, 58, and 69 mm. apart with effect shown (1); lowest feather row (2); 16 mm. below, two tufts of human hair (3) caught under seizing; at lower end whole black feathers in two rows pointing upward and downward with quills together were tied in position over shaft seizing by four turns of sennit (see fig. 241, a). b, cord holders (1) projecting 44 mm. above seizing turn of fine braid in four groups of 5 or 6 holders, succeeded lower down by short single holders (2); seizing braid changes to coarser, 3 mm. wide, which lower down fixes 4 sets of 7 holders (3). c, lower end of b: lower sennit-covered end of shaft covered by turns of doubled thick braid 5 mm. wide, wrapped with human hair of brownish color (1); top end buried under subsequent turns and lower end (2) pushed through under last turn; pedestal (3). d, lower end of shaft with multiple lozenge inaere design (1) worked in fine sennit; upper end (not shown) with sennit seizing supporting coir cord holders in groups of four; four lengths of thick sennit 4 mm. wide were seized together with human hair and looped around upper part of shaft, hair on hanging part had evidently unwound.

The upper surfaces of the circular pedestals join the lower end of the shafts at an obtuse angle, and slope outward and slightly downward to reach its upper circumference, the diameter of which is a little less than that of the bottom. The pedestal is divided into ridges by removing wedge-shaped pieces so that the ridges radiate from the shaft center. The ridges are carved into pillars and arches with the same technique as the upper ridges, and the outer page 376surfaces of pillars and arches are carved with the same geometrical motifs as those in the upper section. A transverse bar was usually left about the middle between the ridges, which may be notched with triangles on both edges or with a sunken lozenge. The under surface of the pedestal was hollowed out so that the lower ends of the carved ridges formed legs around the circumference. (See figure 241.)

Figure 241.—Pedestal of carved slabs, Mangaia (L.M.S. coll.: a,30; b,31; c,32). a, pedestal concealed by feather ornamentation; height 54 mm., base diameter 72 mm.; divided into 7 radiating ridges, carved into two pillars with intervening arch; outer surface of top pillar (1) with K-motif, top of arch below it with two K's facing each other, and lower pillar concealed by feathers. b, base diameter 85 mm.; nine ridges each with two arches (1, 2) and no intervening pillar; sunken lozenge in inter-ridge spaces on level of upper arches, transverse bar with upper and lower edges notched with three triangles on level between two arches; lower arch free through cutting out part of under surface; sennit wrapped with human hair (3) wound around shaft. c, diameter of shaft (1) 50 mm., diameter of base (2) 100 mm., about 12 ridges each with two pillars (3, 4) and intervening arch (5); sunken lozenge in inter-ridge space on same level as arch; continuous lozenge pattern (inaere) on shaft (6).

Figure 241.—Pedestal of carved slabs, Mangaia (L.M.S. coll.: a,30; b,31; c,32). a, pedestal concealed by feather ornamentation; height 54 mm., base diameter 72 mm.; divided into 7 radiating ridges, carved into two pillars with intervening arch; outer surface of top pillar (1) with K-motif, top of arch below it with two K's facing each other, and lower pillar concealed by feathers. b, base diameter 85 mm.; nine ridges each with two arches (1, 2) and no intervening pillar; sunken lozenge in inter-ridge spaces on level of upper arches, transverse bar with upper and lower edges notched with three triangles on level between two arches; lower arch free through cutting out part of under surface; sennit wrapped with human hair (3) wound around shaft. c, diameter of shaft (1) 50 mm., diameter of base (2) 100 mm., about 12 ridges each with two pillars (3, 4) and intervening arch (5); sunken lozenge in inter-ridge space on same level as arch; continuous lozenge pattern (inaere) on shaft (6).

The Cambridge specimen (1354) has six ridges with two arches each, the lower one free (fig. 241, b). All except one of the lower free arches are broken off. On this slab, there are a number of loops made with fine sennit around the wood between the lozenge-shaped holes and under the adjacent arch. It may be safely assumed that these loops formerly attached feathers in the inter-ridge spaces. The lozenge-shaped holes thus assisted in the fastening of feather decoration, and from this position the feather decoration was less likely to obscure the carving patterns on the arches and pillars.

The carving motifs used on pillars and arches were formed on small rectangular surfaces by cutting out small triangular or rhomboidal pieces to leave portions of the original surface in the forms of transverse bars and chevrons. Both bars and chevrons were separated by an intervening space or brought together to form a geometrical figure like the letter K. Owing to the larger area left above the arch, the patterns made on the longer space contained more elements than on the pillars. Sometimes the bar was notched in the middle and sometimes the outline of the K was formed without cutting out the small triangles to define the arms of the K. Another variation occurred when a wide page 377bar was formed into a triangle through trimming the arch to form a mesial edge. The variations and combinations of the bar and chevron with their use on the pillars and arches are shown in figure 242.

Figure 242.—Carving motifs and combinations. a, motifs: 1, bar; 2, notched bar; 3, chevron; 4, bar and chevron; 5, K; 6, unnotched K; 7, trimmed bar and chevron. b, combinations on pillars: 1, K; 2, K and bar; 3, K and notched bar; 4, unnotched K and notched bar; 5, paired K's; 6, paired K's repeated. c, combinations on arches: 1, chevron; 2, paired chevrons; 3, bar-chevron and bar; 4, notched bar, chevron, and notched bar; 5, K and bar; 6, paired K's; 7, pair trimmed bar and chevron. When bar and chevron are separated by a space they are treated as two distinct elements but when point of chevron touches bar, they are treated as a compound (bar-chevron). When point of chevron fuses with bar, it becomes a true K. Distinction between the three forms is due to size of two notches between bar and chevron. Mesial line running from middle of any K or chevron is formed by an edge, due to cutting out intervening material in a downward and outward direction on either side of mesial line. In trimmed bar (c,7) triangular form of bar is due to trimming arch into a mesial edge beyond carving pattern.

Figure 242.—Carving motifs and combinations. a, motifs: 1, bar; 2, notched bar; 3, chevron; 4, bar and chevron; 5, K; 6, unnotched K; 7, trimmed bar and chevron. b, combinations on pillars: 1, K; 2, K and bar; 3, K and notched bar; 4, unnotched K and notched bar; 5, paired K's; 6, paired K's repeated. c, combinations on arches: 1, chevron; 2, paired chevrons; 3, bar-chevron and bar; 4, notched bar, chevron, and notched bar; 5, K and bar; 6, paired K's; 7, pair trimmed bar and chevron. When bar and chevron are separated by a space they are treated as two distinct elements but when point of chevron touches bar, they are treated as a compound (bar-chevron). When point of chevron fuses with bar, it becomes a true K. Distinction between the three forms is due to size of two notches between bar and chevron. Mesial line running from middle of any K or chevron is formed by an edge, due to cutting out intervening material in a downward and outward direction on either side of mesial line. In trimmed bar (c,7) triangular form of bar is due to trimming arch into a mesial edge beyond carving pattern.

It is seen from the carving technique of pillars, arches, and geometrical motifs that there is nothing to suggest any derivation from a human figure or to justify the "degradation" that some writers have applied to this form of design.

The Oldman carved slab (54, vol. 47, pl. 4, no. 432) resembles the four just discussed in having an upper section divided into vertical ridges with arches and lozenge-shaped perforations between the ridges, a rounded staff, and an expanded circular pedestal. It differs in that the arches are connected without intervening columns. Each wide surface has three ridges, and the sides have one each. The ridges are carved to form alternate wide and narrow arches, eight of each in each ridge. The inter-ridge perforations are on the same levels as the wide and narrow arches (pl. 14, I)

page 378

The wide and narrow arches are complementary to each other in the carving pattern. The wide arch is carved with the bar and the chevron motifs and the narrow arch with the notched bar. The notched bar is formed by notching the upward projection of the pillar between the lower end of the wide arch and the upper end of the narrow arch, but the notch is continued downward to perforate the narrow arch into two longitudinal divisions which resemble legs. At first sight, the carving appears distinct from that of the preceding four slabs but a careful analysis shows that the difference is merely one of degree. (See figure 243.)

Figure 243.—Carved slab (Oldman coll., 432): length 43 inches; upper carved part, height 17.25 inches, top width 7 inches, thickness 4 inches; eight ridges, three front and back, one at each side; brown hard wood. a, upper end: upper edge broken in places; three carved ridges (1) on one surface and one carved ridge (2, 2) at each side; lozenge-shaped perforations between ridges. b, part of one front ridge (1), side ridge (2), and lozenge perforations (3); ridges carved into main pillars (4, 4) with arch (5, 5) showing chevron flange (6, 6) at upper part, supported below by secondary pillar (7, 7) which with secondary arch (8, 8) is grooved in middle line, order of carving motifs from front being plain bar (4), chevron (6), arch (5), notched bar (7), and grooved arch 8). Pillars and arches with accompanying carving motifs are repeated throughout length of each ridge eight times. c, lower end of carved portion: showing end of three front ridges (1) and two side ridges (2) bounded by a transverse band (3) with sunken lozenges; below band, single arches (4) correspond to line of upper ridges; wood is curved in to form a shoulder (5) which meets rounded shaft (6). d, pedestal flared out with a number of carved ridges which are broken, details not discernible from photograph.

Figure 243.—Carved slab (Oldman coll., 432): length 43 inches; upper carved part, height 17.25 inches, top width 7 inches, thickness 4 inches; eight ridges, three front and back, one at each side; brown hard wood. a, upper end: upper edge broken in places; three carved ridges (1) on one surface and one carved ridge (2, 2) at each side; lozenge-shaped perforations between ridges. b, part of one front ridge (1), side ridge (2), and lozenge perforations (3); ridges carved into main pillars (4, 4) with arch (5, 5) showing chevron flange (6, 6) at upper part, supported below by secondary pillar (7, 7) which with secondary arch (8, 8) is grooved in middle line, order of carving motifs from front being plain bar (4), chevron (6), arch (5), notched bar (7), and grooved arch 8). Pillars and arches with accompanying carving motifs are repeated throughout length of each ridge eight times. c, lower end of carved portion: showing end of three front ridges (1) and two side ridges (2) bounded by a transverse band (3) with sunken lozenges; below band, single arches (4) correspond to line of upper ridges; wood is curved in to form a shoulder (5) which meets rounded shaft (6). d, pedestal flared out with a number of carved ridges which are broken, details not discernible from photograph.

The upper end of the upper section is broken in parts and the exact finish is not clear. The lower end is narrowed laterally and thickened antero-posteriorly into a rounder section. Its upper boundary is formed by a narrow band carved with sunken lozenges resembling the technique on the pedestal of one of the other slabs (fig. 241, b).

The middle staff part is smooth and circular in section, but it was probably ornamented originally like the three slabs in the British Museum.

The lower circular pedestal is much broken and it appears as if a row of arches had been broken off. (See figure 243, d and plate 14, I.) For comparison of Mangaian and Mitiaro gods, see p. 467.

page 379
Images in Human Form

Although Gill states that Motoro was "a rude representation of the human form, carved in ironwood" and conveys the impression that the other district gods carved by Rori had a similar form, there is no support from the objects available for study. The three district gods in the British Museum and the similar object in the Peabody Museum, Cambridge, show no trace of the human form. It may be that Gill was so obsessed with the idea that the geometrical motifs on these objects represented human figures that he used the term "human form" somewhat loosely. It is true that the spiked god (figs. 233-235) has definite human figures, but the figure cannot be definitely attributed to Mangaia. The Oldman figure (fig. 243) does have motifs on the arches which may be regarded as conventional human forms, but by no stretch of imagination can we accept the whole figure as representing a human form.

What seems to have been a definite human form was the stone symbol of Rongo, which Gill (33, p. 267) describes as follows:

A long block of stone, rudely shaped like a man, was also regarded as an inferior representation of this Polynesian Mars. Many years ago, when the people embraced Christianity, this huge stone idol was utterly defaced, and the fragments form part of the stonework of the church at the principal village.

Gill (33, p. 331) further states:

A block of stone, shaped like a man, and covered with cloth and coarse sacrifice-nets, was set up at his marae O-Rongo for the convenience of worship and sacrifice; where also stood a smaller image named Little Rongo, or "Rongo-of-the-red-tongue."

Human sacrifices, after being offered on the inland marae of 'Akaoro, were carried in a coarse sennit net to the shore marae of Orongo where they were offered to Rongo. The body was then thrown into the bushes at the back of the marae and the net was cast over the statue of Rongo where it was left to decay. Rongo was thus represented by two stone images on the marae of Orongo and by a shell trumpet in the national god house. What the images were like we can never know.

With the exception of these two sporadic stone images, Mangaia has followed the principle adopted by Atiu, Mitiaro, and probably Mauke in repsenting their gods by carved stands decorated with sennit, feathers, and human hair.

Ceremonial Adzes

Adzes lashed with the Mangaian triple triangle pattern (fig. 110) to shafts of various forms, carved with the K-motif are widely distributed through the museums of the world. Though often labeled "Hervey Islands" or even "Austral Islands", they are peculiar to the one island of Mangaia. A general page 380impression prevails that they were used in religious ritual connected with the making of peace, hence they have been popularly termed "peace axes." The only authoritative statement concerning the religious significance of such adzes is given by Gill (31, p. 224):

The stone adzes were secured to their wooden hafts by means of fine sinnet, itself esteemed divine. It was fabled that the peculiar way in which the natives of Mangaia fasten their axes was originally taught them by the gods. A famous god, named Tanemataariki, i.e., Tane-of-the-royal-face, was considered to be enshrined in a sacred triple axe, which symbolized the three priestly families on the island, without whose aid the gods could not be acceptably worshipped. Tane-of-the-royal-face was one of the very few much-respected gods not surrendered to the missionaries, but hidden in caves. All trace of this interesting relic of heathen antiquity is now lost. The shape of a god-adze differed at the back from those used by artisans in being rounded underneath.

If by triple adz it is meant that three adz heads were attached to one shaft, the technical problem of lashing must have been interesting. The god-adzes, as Gill terms them, evidently had the median ridge at the back rounded off thus making the adz thin and unsuitable for working wood. Several such adz heads are known. If they were all "god-adzes" there must have been a number of single adzes in use.

The number of adzes with carved shafts now preserved seems to be out of all proportion to the needs of the limited population of Mangaia. Dodge (21) describes 25 adzes with carved shafts in the Peabody Museum, Salem; and Giglioli (27) describes 33 in his private collection. The reasons for this large number is due to their unique appearance, which aroused the interest of collectors. The Mangaian craftsmen satisfied this demand by carving wooden hafts and mounting old adz heads upon them. Hence the supply continued although the original functions of the adzes ceased after the acceptance of Christianity.

On my last visit to Mangaia in 1929, Taniera Tangitoru was the sole survivor of a succession of craftsmen who had for a century made carved adzes for the foreign markets. In the early part of the foreign production period the craftsmen no doubt adhered to stone age technique in the forms of the hafts, carving patterns, lashing design, and types of adz heads used, albeit they employed steel tools on the woodwork. In the course of time, however, craftsmen without the careful training of their immediate forbears were influenced more and more by the desire to please their foreign buyers. With steel tools, changes in the shape of the hafts were readily produced, and new carving motifs were introduced to create a greater appeal to the limited market. As the production of artifacts was merely a side issue, the details of carving finish and lashing technique received less attention. The later products have cruder carving, and the lashings indicate clearly that the unique Mangaian form of the triple-triangle design had been forgotten. As the supply of adz heads appropriate to the carved hafts fell short, any available adz head was used. page 381Thus, adz heads are now scarce in Mangaia, as compared with the neighboring islands. The two daughters of Taniera Tangitoru sent him adz heads from Rarotonga to fit to the Mangaian hafts that he made. On occasion Tangitoru was so hard pressed for original stone adz heads that he shaped pieces of rock with a grindstone to complete his adzes, but the old triangular pattern involved so much work that the stone was ground down like an ax with a bevel from front and back.

It is evident that a large number of the Mangaian carved hafts in collections were not made to function in the native culture and that form and carving were affected by influences foreign to that culture. Though interesting in that they show the changes since European contact, later artifacts have confused the determination of the original Mangaian forms. A careful inquiry is needed to determine what adzes in museums date back to the period immediately following missionary contact in 1823. Lacking sufficient information, an analysis will be made of the material now available. From their material form the carved adzes may be broadly classified into three main classes according to the form and carving of their shafts.

Class 1 consists of adzes with shafts that could be manipulated with one hand. Professor Giglioli (27), from information supplied by European and native residents in the Cook Islands, gave the names of such adzes as toki tamaki (war adzes), toki kaika'a (weapon adzes), and toki a rore. In the last name, rore is a misspelling for Rori. He was a skilled carver, and the form of Mangaian carving is attributed to him. In the name toki-a-Rori (adz-of-Rori) we have an acknowledgment that Rori began the application of the carving technique to the shafts of adzes that could be used by hand. Giglioli (27) states that such adzes were used as weapons and as the insignia of chiefs. Gill (31, p. 224) states that "some of their adzes were intended for dispatching their foes", and he relates that in the battle of Araeva (31, p. 69) a great stone adz was upraised to slay a man named Eke. The use of adzes as weapons is thus supported and the terms toki tamaki and toki kaika'a are descriptive of this particular function. The term toki-a-Rori brings out the fact that the shaft was carved. That such adzes should be used as property by chiefs, conforms to Polynesian custom. The highly carved hafts of Maori jade adzes were also used as weapons and property. A number of adzes with usable carved shafts have been preserved. Those I have seen appear to have been carved with metal tools, but they no doubt represent an old form. The shaft and foot resemble those of working adzes (fig. 111), the only difference being the complete carving of the shaft which is consequently somewhat thicker. The shaft is divided into narrow panels by transverse grooves or deeply cut lines which encircle the shaft. The panels are carved with combinations of the bar and chevron motifs resembling the K-figures on the carved slabs representing district gods. In the district gods, however, the carving surfaces on the pillars page 382and arches are short and admit of only one or two K-motifs, whereas the longer panels on the adz shafts allow repetition of the carving motifs. In one specimen the orthodox division into panels was not used and a continuous lozenge pattern took the place of the K-pattern (fig. 244, c). This may be regarded as a post-missionary development. The shafts are generally round in section, but some of the more modern specimens may have four or more longitudinal surfaces. These adzes have no religious significance. (See figure 244.)

Figure 244.—Carved hafts, class 1. a, K figures in repeated pairs, horizontal to long axis of panels (Wesleyan Univ., Middletown, Conn., 374). b, shaft trimmed to eight longitudinal surfaces of equal width, but notched transversely into panels; double K figure on each surface, vertical to long axis of panel (Auckland Mus., N. Z.). c, round shaft without panels and with continuous lozenge pattern covering whole surface (Otago University Mus., N. Z., D.093).

Figure 244.—Carved hafts, class 1. a, K figures in repeated pairs, horizontal to long axis of panels (Wesleyan Univ., Middletown, Conn., 374). b, shaft trimmed to eight longitudinal surfaces of equal width, but notched transversely into panels; double K figure on each surface, vertical to long axis of panel (Auckland Mus., N. Z.). c, round shaft without panels and with continuous lozenge pattern covering whole surface (Otago University Mus., N. Z., D.093).

Class 2 includes shafts which are enlarged forms of class 1. They could not be used, and were never intended to be used, with one hand. In cross-section the shafts range from round to square, and their length usually exceeds greatly that of the working adzes. Some are massive and very long. One in the Cambridge University Museum, figured by Sayce (58, frontispiece), is 6 feet 3 inches long. The diameter of the shaft is practically uniform throughout except at the lower end where the transverse panels project slightly. In some shafts, small, square-cut projections or legs were formed on the lower end and the outer surfaces of these were carved. Transverse narrow panels were formed on the whole length of the shafts, and the dividing grooves were sometimes fairly deep and wide, resembling the deep grooves between the ridges page 383in the carved slabs representing gods. In shafts with a rectangular cross-section, the grooves were disconnected at the four corners, which were undercut to allow communication between neighboring grooves on the same level. In some shafts, longitudinal grooves made the carved panels vertical instead of horizontal. The carvings on the panels were variations of the K-motif, as in class 1. The foot and the lashing design were also similar to those of class 1. Two examples of class 2 were figured by Williams (81, p. 225), which shows that the form is old. I believe that the religious symbols were made in this form. (See figure 245.)

Figure 245.—Ceremonial adzes, class 2. a, octagonal shaft with double K-motif on each surface, side edge of each surface on middle of K-stem (1): expanded base with two upper serrated edges (2) followed by continuous lozenge pattern (3) and terminal legs with single K-motif (4) (Rome Mus. Preist., 1073). b, total length, 1,718 mm. (67.5 inches); shaft octagonal, 66 mm. in diameter; transverse panels separated by deep grooves; carved with fused K-motif (1): longitudinal slit (2) on shaft; expanded base (3) with three carved bands; feet with single K-motif (4); base diameter, 103 mm.; base height, 95 mm. (Fuller coll., 261). c, long square shaft with deep transverse grooves between carved panels with fused K-motif (1); grooves do not reach side edges formed by narrow notched strip; just above base expansion, one panel is grooved longitudinally into four divisions carved with double K (2); expanded base divided into two bands by deep groove and bands carved (3); legs also carved (4) (Fuller coll., lot 654).

Figure 245.—Ceremonial adzes, class 2. a, octagonal shaft with double K-motif on each surface, side edge of each surface on middle of K-stem (1): expanded base with two upper serrated edges (2) followed by continuous lozenge pattern (3) and terminal legs with single K-motif (4) (Rome Mus. Preist., 1073). b, total length, 1,718 mm. (67.5 inches); shaft octagonal, 66 mm. in diameter; transverse panels separated by deep grooves; carved with fused K-motif (1): longitudinal slit (2) on shaft; expanded base (3) with three carved bands; feet with single K-motif (4); base diameter, 103 mm.; base height, 95 mm. (Fuller coll., 261). c, long square shaft with deep transverse grooves between carved panels with fused K-motif (1); grooves do not reach side edges formed by narrow notched strip; just above base expansion, one panel is grooved longitudinally into four divisions carved with double K (2); expanded base divided into two bands by deep groove and bands carved (3); legs also carved (4) (Fuller coll., lot 654).

page 384
Figure 246.—Ceremonial adzes, class 3, with pedestals, vary in height; b is 17 inches high, g38 inches high. a, round shaft (1) completely carved with continuous lozenge pattern; round pedestal (2) with upper edge carved in rectangular projections with horizontal double K-motif (3); perforated with square holes surrounded by horizontal and vertical panels of horizontal double K-motif (4); rest of pedestal carved with horizontal rows of fused K's repeated (5): leg knobs with horizontal double K (6) (Auckland Mus., N.Z., 14486). b, round shaft with continuous lozenge pattern (1); round pedestal (2) with upper notched band carved with horizontal double K (3); upper wider (4) and lower (5) bands carved with continuous lozenge pattern like shaft; rest of pedestal divided into vertical panels alternately carved with horizontal double K (6) in horizontal rows and with small excised triangles (7) arranged in rows of three in three vertical subdivisions in each panel; leg knobs carved with horizontal double K (8) and some with excised triangles (9) (St. Augustine's College, Canter-bury Canterbury ). c, round shaft with continuous lozenge pattern (1); round pedestal (2) with upper notched border with single pairs of horizontal K's (3), followed by two horizontal rows, horizontal K's fused into geometrical pattern (4); also bottom band (5) with two rows similar to upper band (4); between bands (4, 5) pedestal is cut with wide vertical grooves forming narrow vertical panels carved with single pairs of horizontal K's (6); leg knobs with single pairs of horizontal K's (7) but stem of lower K usually left out (Otago University Mus., N.Z., D.36, 761). d, round shaft with continuous lozenge pattern (1); round pedestal (2) carved to form narrow vertical and oblique panels; upper border (3) carved with two horizontal rows of horizontal K's fused into geometrical pattern (4), similar double row on lower border (5) and single row on narrow oblique (6) and horizontal (7) panels; vertical panels (8) carved with repeating horizontal fused K (9); leg knobs with middle fused K (10) (Otago University Mus., D.36, 797). e, round shaft with continuous lozenge pattern (1) but lower part with vertical rows of triangles (2) in fives with bases down; pedestal (3) quadrangular with eight square perforations on each surface; upper edge of pedestal notched with rectangular knobs carved in horizontal double K (4); except for areas around holes, body carved with horizontal rows of upright fused K (5) with limbs of opposing K's not meeting; rectangular space around each hole carved with vertical rows of triangles with apices cut off by adjoining bases, part above hole with bases upward (6) and part below with bases below (7); leg knobs with horizontal double K (8) with stem of lower K not plain (Auckland Mus., 1299). f, octagonal shaft (1) with each vertical surface divided into three parts by horizontal lines, each division carved alternately with continuous lozenge pattern or vertical rows of excised small triangles, also upper part of neck carved with triangles: rectangular pedestal (2) with five rows of three rectangular perforations; upper border of pedestal carved with single horizontal row of fused K (3) bounded below by single row of excised triangles on side (4); wide upper band of continuous lozenge pattern (5) and also lower band (6); four intermediate horizontal bands between sets of holes carved with three rows of fused K-motif (3); small rectangular panels between holes on same row carved with close rows of triangles (7); leg knobs with horizontal double K (8) (Otago University Mus., D.29, 1356). g, octagonal shaft with continuous lozenge pattern (1) continued down to first band (3) on square, pedestal (2) and lower band (4); pedestal with five horizontal slits forming four narrow panels with single row of vertical fused K's (5); lower part with six vertical slits running into lowest horizontal slit and forming five short vertical panels with free upper end and carved with single row of horizontal fused K's (6); leg knobs with middle fused K carved horizontally (7) Peabody Mus., Salem, E.22, 453). h, Giglioli's peace adz or toki mahia: octagonal shaft (1) with each surface carved in vertical lines of small incised triangles: rectangular pedestal (2) with upper projecting band (3) with upper border (4) notched into rectangular projections on sides and carved with two triangles (5), body with horizontal rows of triangles as shown and lower border (6) notched similarly to upper border (4, 5); lower projections band (7) treated similarly to upper band (3) but some of lower projections not notched on sides; intermediate part (8) divided into three sections, upper one carved in vertical rows of small triangles and lowest section (9) carved with horizontal rows of larger triangles; narrow plain part (10) below lower band (Rome Mus., Preist., 1492).

Figure 246.—Ceremonial adzes, class 3, with pedestals, vary in height; b is 17 inches high, g38 inches high. a, round shaft (1) completely carved with continuous lozenge pattern; round pedestal (2) with upper edge carved in rectangular projections with horizontal double K-motif (3); perforated with square holes surrounded by horizontal and vertical panels of horizontal double K-motif (4); rest of pedestal carved with horizontal rows of fused K's repeated (5): leg knobs with horizontal double K (6) (Auckland Mus., N.Z., 14486). b, round shaft with continuous lozenge pattern (1); round pedestal (2) with upper notched band carved with horizontal double K (3); upper wider (4) and lower (5) bands carved with continuous lozenge pattern like shaft; rest of pedestal divided into vertical panels alternately carved with horizontal double K (6) in horizontal rows and with small excised triangles (7) arranged in rows of three in three vertical subdivisions in each panel; leg knobs carved with horizontal double K (8) and some with excised triangles (9) (St. Augustine's College, Canter-bury ). c, round shaft with continuous lozenge pattern (1); round pedestal (2) with upper notched border with single pairs of horizontal K's (3), followed by two horizontal rows, horizontal K's fused into geometrical pattern (4); also bottom band (5) with two rows similar to upper band (4); between bands (4, 5) pedestal is cut with wide vertical grooves forming narrow vertical panels carved with single pairs of horizontal K's (6); leg knobs with single pairs of horizontal K's (7) but stem of lower K usually left out (Otago University Mus., N.Z., D.36, 761). d, round shaft with continuous lozenge pattern (1); round pedestal (2) carved to form narrow vertical and oblique panels; upper border (3) carved with two horizontal rows of horizontal K's fused into geometrical pattern (4), similar double row on lower border (5) and single row on narrow oblique (6) and horizontal (7) panels; vertical panels (8) carved with repeating horizontal fused K (9); leg knobs with middle fused K (10) (Otago University Mus., D.36, 797). e, round shaft with continuous lozenge pattern (1) but lower part with vertical rows of triangles (2) in fives with bases down; pedestal (3) quadrangular with eight square perforations on each surface; upper edge of pedestal notched with rectangular knobs carved in horizontal double K (4); except for areas around holes, body carved with horizontal rows of upright fused K (5) with limbs of opposing K's not meeting; rectangular space around each hole carved with vertical rows of triangles with apices cut off by adjoining bases, part above hole with bases upward (6) and part below with bases below (7); leg knobs with horizontal double K (8) with stem of lower K not plain (Auckland Mus., 1299). f, octagonal shaft (1) with each vertical surface divided into three parts by horizontal lines, each division carved alternately with continuous lozenge pattern or vertical rows of excised small triangles, also upper part of neck carved with triangles: rectangular pedestal (2) with five rows of three rectangular perforations; upper border of pedestal carved with single horizontal row of fused K (3) bounded below by single row of excised triangles on side (4); wide upper band of continuous lozenge pattern (5) and also lower band (6); four intermediate horizontal bands between sets of holes carved with three rows of fused K-motif (3); small rectangular panels between holes on same row carved with close rows of triangles (7); leg knobs with horizontal double K (8) (Otago University Mus., D.29, 1356). g, octagonal shaft with continuous lozenge pattern (1) continued down to first band (3) on square, pedestal (2) and lower band (4); pedestal with five horizontal slits forming four narrow panels with single row of vertical fused K's (5); lower part with six vertical slits running into lowest horizontal slit and forming five short vertical panels with free upper end and carved with single row of horizontal fused K's (6); leg knobs with middle fused K carved horizontally (7) Peabody Mus., Salem, E.22, 453). h, Giglioli's peace adz or toki mahia: octagonal shaft (1) with each surface carved in vertical lines of small incised triangles: rectangular pedestal (2) with upper projecting band (3) with upper border (4) notched into rectangular projections on sides and carved with two triangles (5), body with horizontal rows of triangles as shown and lower border (6) notched similarly to upper border (4, 5); lower projections band (7) treated similarly to upper band (3) but some of lower projections not notched on sides; intermediate part (8) divided into three sections, upper one carved in vertical rows of small triangles and lowest section (9) carved with horizontal rows of larger triangles; narrow plain part (10) below lower band (Rome Mus., Preist., 1492).

page 385

Class 3 may be termed pedestal adzes, because the haft has an enlarged pedestal forming the lowest section. Giglioli (27) gives the following names for the various parts: head [foot], poro; column [shaft], ma'a; base [pedestal], uma. The foot is similar in shape to those in classes 1 and 2. The shafts are usually round or elliptical, but some are shaped to form four or more longitudinal surfaces defined by distinct edges. The shaft flares out from the con-page 386stricted neck of the foot and continues downward with a gradually increasing diameter until it meets the pedestal at a distinct shoulder, which is due to the greater diameter of the pedestal. The lashing is the same as that in the other two classes and in working adzes. (See figure 109.)

The pedestal may follow the general cross-section of the shaft, but it is usually quadrangular and thus contrasts with the general rounded form of the shaft. The most characteristic feature of the pedestal is the presence of rectangular holes, which divide the surfaces into smaller horizontal and vertical panels. The panels in turn, may be subdivided by deeply incised horizontal or vertical lines into panels carved with the K-motif. As in adzes of class 2, the lower end of the pedestal is cut with square notches to form short projecting legs, which are also carved with the K-motif.

The narrow panels on the pedestals are carved with the bar and chevron motif, as are the corresponding panels in classes 1 and 2; but a new motif in the form of rows of incised triangles is sometimes introduced. The surface of the thick shafts is not divided by grooves, clefts, or holes; and, on this uninterrupted field, the geometrical motifs are combined to form a pattern of continuous lozenges that could not be developed on the narrow panels of classes 1 and 2. Rows of triangles are sometimes used on the shafts as well as on the pedestals.

With the use of steel tools and with the probable weakening of craft conservatism as regards form, the stimulus of trade led to great variety in the pedestal forms. I have not yet seen an example of this class that dates back to the period of early missionary contact. It is significant that though Williams (81, pp. 224, 225) speaks of "the exquisite carving of the handles of their stone axes", the two examples that he figures belong to class 2. It seems reasonable to suppose that if the more spectacular pedestal forms had been available in his time, he would have figured one of them. I believe that the long pedestal form with the carving motifs of lines of triangles and continuous lozenges has been developed in the post-missionary period and that it never had religious significance. Some of the common forms are shown in figure 246.

Giglioli's Classification

Giglioli (27), from his study of his collection of 33 carved adzes and from correspondence with natives and Europeans in the Cook Islands, classified adzes into five groups which are summarized as follows:

1.Working adzes with uncarved hafts.
2.Adzes with carved shafts that were easy to handle. (This corresponds to my class 1.)
3.Adzes with carved shafts not usable for practical purposes. For the carving he adopts Gill's term of tikitiki tangata and hence terms the adzes toki tikitiki. They are enlarged forms of his class 2 which being carved, have as much right to the term toki page 387 tikitiki (carved adz) as his class 3. He states that they "are considered as funerary monuments connected with the cult of the ancestors." (This curious statement will be dealt with later.) He divides his class 3 into two:
a.Typical and complete with shaft divided into three parts of which the lowest is the pedestal (my class 3).
b.Those without a pedestal.
4.Adzes with shaft in form of a square pyramid with carvings of triangle motifs (m'o mango; sharks' teeth) and with at least two faces carved with combinations representing different tribes. Giglioli termed them peace adzes named toki mahia. He stated that they were scarce and that he had the only sample in Europe. It was made for him by a Mangaian named Miringa-tangi. I knew Miringa-tangi personally and both his contemporaries, and I regard him as a very unsound authority. He and Taniera Tangitoru carved adz hafts for trade but they had no old patterns to copy. The adz figured by Giglioli (27, p. 8) shows a modern form of my class 3 with poor carving and lashing. I cannot accept this adz as forming a special class (fig. 246, h). The question of peace adzes is discussed on page 395.
5.Sacred adzes. This is a purely hypothetical group based on the statement by Gill (31, p. 224) already quoted concerning the god Tane-mataariki. Giglioli quoted Gill and said that he understood that the three stones were now hafted separately on monumental hafts (toki tikitiki) of different types. He then described the large flat adz heads with notched ridges as being Tane-mataariki of which he had one. One in the British Museum has been described (fig. 84), but the shaft upon which it is mounted is a long shaft (my class 2) and saw marks dispose of the question of the antiquity of the hafting. (For further discussion see page 394.)
Development of Geometrical Patterns

District gods. The long panels or ridges were vertical and the division into pillars and arches provided small rectangular surfaces on which the bar motif was used to define one end of the surface and the chevron motif to decorate the remaining area (fig. 242, b,1). When space permitted, another bar bounded the other end of the surface (fig. 242, b,2, 3; c,3, 4, 5). When further space or artistic inclination prompted, a second chevron and bar were introduced for symmetrical balance (fig. 242, b,5, 6; c,6, 7). The open side of the second chevron faced the open side of the first chevron. In forming the open side of the chevron, the carving implement was pressed down vertically toward the outer edges to deepen the incisions. In removing the surrounding wood to make the chevron stand out, the implement cut from the mesial longitudinal line in a slanting direction downward and outward on one side to define one limb of the chevron. The process was repeated on the other side. This technique resulted in a longitudinal mesial edge extending from the open angle of the chevron to the middle of the next motif, whether it was a bar or a chevron. Thus, when two chevrons were used, the mesial edge extended between their open angles. The mesial edge was not a carving motif, but because this technical edge comes out as a distinct line in rubbings it has been erroneously regarded as part of the carving design. In the district gods, the bars and chevrons lie in a horizontal direction, owing to the vertical direction of the carved surfaces; and the technical edge extending between the open angles of the chevrons is vertical.

page 388

Adz shafts of classes 1 and 2. The panels on these shafts are horizontal and of a similar width to that of the short surfaces on district gods. The bar and chevron motifs, thus lie vertically. The chevrons are produced by exactly the same technique as that used in the district gods. consequently the technical mesial edges which extend between the open angles of adjoining chevrons are horizontal. The carving technique is identical in the two groups, though the panel direction is vertical in one and horizontal in the other. On the adz shafts, however, the bar and chevron motifs are repeated on each panel to cover the longer surface.

Figure 247.—Carving patterns on ceremonial adzes. a, double K-motif. b, double K-motif separated by grooves. c, fused K with vertical limbs of back-to-back K's combined in one; showing two panels separated by a fairly wide groove. d, three panels with fused K-motif separated by narrow incised line. e, lozenge pattern obtained by eliminating transverse groove between panels and vertical elements of K-motifs. f, common multiple lozenge pattern in which limbs of K-motif are arranged in line. g, narrow band pattern in which stems of K's are horizontal and limbs do not meet. h, triangular motifs with apex touching base of next triangle. i, triangular motifs in which apices are cut off by succeeding bases.

Figure 247.—Carving patterns on ceremonial adzes. a, double K-motif. b, double K-motif separated by grooves. c, fused K with vertical limbs of back-to-back K's combined in one; showing two panels separated by a fairly wide groove. d, three panels with fused K-motif separated by narrow incised line. e, lozenge pattern obtained by eliminating transverse groove between panels and vertical elements of K-motifs. f, common multiple lozenge pattern in which limbs of K-motif are arranged in line. g, narrow band pattern in which stems of K's are horizontal and limbs do not meet. h, triangular motifs with apex touching base of next triangle. i, triangular motifs in which apices are cut off by succeeding bases.

The most developed combination of the bar and chevron motifs possible on the short surfaces on the district gods is the short balanced bar and chevron combination shown in figure 242, b,5; c,6. This pattern was used and repeated on the adz shafts until the length of the panel was filled in. The combination of bar and chevron formed the K-figure, and in the double combination the ends of the arms of the K's may or may not meet on the edges of the panel (fig. 247, a). In some panels, the combination of two K's facing each other is repeated as a distinct pattern. The combinations are separated by a vertical groove between the vertical limbs of the K's, which are back to back (fig. 247, b). The more common technique, however, was to omit the vertical groove and allow one vertical element to serve the two back-to-back K's. In other words, the two vertical limbs of the back-to-back K's are fused into one (fig. 247, c). This combined figure, which may be called "fused K's", is a repeating motif, for it forms the right stem of one pair of K's and the left of page 389another pair. It is the most common motif on the longer panels of the carved hafts. In such panels the design commences on the left with a normal K, which is followed by a series of fused K's and ends on the right with a reversed K. The effect of the oblique arms of the K's meeting at the upper and lower edges of the panel is to enclose a lozenge-shaped space with a horizontal edge extending between the side angles of the lozenge.

The pattern set up on one horizontal panel is repeated on the succeeding panels, as shown in figure 247, c. Two panels separated by a wide groove are shown in the figure. In some adz shafts the wide groove is replaced by an incised line which brings the patterns on adjacent panels close together (fig. 247, d). With the close approximation of panels, the small excised triangles between the oblique and vertical elements of the K's of an upper and lower panel practically fuse into larger triangles; and these triangles on either side of the joined vertical elements of the K's form an alternating set of lozenges divided by vertical bars and by horizontal incised lines defining the panels. In the adz shafts of classes 1 and 2, in which the conventional panels were maintained, the pattern in figure 247, d was as far as progressive development could go.

With a few exceptions, the pedestals in class 3 were marked off into panels, and the patterns developed on the panels of groups 1 and 2 were, of necessity, followed. On the shafts, however, the carvers departed from the panel technique; and, with the disappearance of horizontal grooves and dividing lines, the tendency toward fusion (fig. 247, d) became complete in the common pattern shown in figure 247, e, f. Here it is evident that the idea of alternating sets of excised lozenges suggested in figure 247, d took definite form. The panel dividing lines which marred the new set of lozenges were eliminated, and for further balance and symmetry the vertical limbs of the K's also were left out; but to retain the alternating effect the two triangles forming the new series of alternating lozenges were excised laterally from a mesial vertical edge. Thus the alternating set of lozenges adopted the technique of the original sets except that the two halves were cut laterally from a vertical mesial edge, whereas in the original lozenges the two halves were cut upward and downward from a mesial horizontal edge. It may be said that the vertical edges are on the site of the vertical elements of the K's which were completely eliminated in the new development. What remained of the fused K-motif were the oblique arms on either side, which fused to form continuous oblique raised lines running in opposite directions and enclosing lozenges. In other words, the original bar motif was abandoned and the normal and reverse chevrons were retained.

An introduction to the common multiple lozenge pattern was figured by Stolpe from an adz haft in Neuchatel (fig. 247, e).

page 390

A common pattern used on the front surface of the foot of pedestal hafts is shown in figure 247, g, but it is merely another arrangement of the combined bar and chevron motifs in which the ends of opposing chevrons do not meet.

The preceding discussion disposes of the various forms and combinations of the bar and chevron motifs. It will be noted that, except for straight grooves which are used to divide panels, the craftsmen made use of excised triangles to separate the bar and chevron motifs used in the patterns.

Excised triangles as a distinct motif were used in class 3 on both the pedestal and the column. They were arranged in rows, either vertical or horizontal, with the apex of one triangle touching the middle of the base of the next triangle (fig. 247, h). This arrangement is characteristic of the later. Mangaian work. A variation of the triangle pattern occurs when the apices of the triangles are overlapped by the bases of adjoining triangles, giving a "herring bone" appearance (fig. 247, i). Gill (31, p. 223) termed this technique "the sharks' teeth pattern" (nio mango). The word nio should be written ni'o (tooth) as it has the glottal h, being written niho in other dialects.

The square of rectangular holes used only on the pedestals of group 3 are termed ai tuna (eel borings) by Gill (31, p. 223). Giglioli (27) quotes Colonel Gudgeon as stating that they were called rua (hole) and rua matangi (hole of the wind). According to Giglioli, when the holes are on all four sides of the pedestal and on the same level (as they always are), the meaning of ruamatangi is "mouth of the four winds." This is too free a translation. Taniera Tangitoru stated that they were rua tangeo (tabernacles of the god Tangeo). Tangeo is not present in the list of Mangaian gods, and it is difficult to see how an unknown god could be located in objects which were originally associated with the god Tane-mata-ariki.

In the district god slabs, a marked feature of the technique is the presence of holes in the spaces between the ridges. I believe that the holes in the pedestals are a carry-over in technique and are rendered easier by the use of steel tools. In the district gods the holes are lozenge shaped, but from their position they do not interfere with the arrangement of the carved surfaces on the ridges. In the pedestals, however, they were made square or rectangular to avoid interference with the plotting out of vertical and horizontal panels as a field for carving. In the district gods the holes functioned in the tying on of feathers, but in the adz pedestals they were purely ornamental. It is evident that the original function of the holes was forgotten and that the later carvers invented new functions, probably to satisfy the questions of foreigners. The fact that they were given two functions and two sets of names indicates that rationalization had taken place.

The grooves or clefts in the district gods served to divide the slab surface page 391into ridges. This technique was applied to the adz shafts to form panels, and though they underwent a change of direction from the vertical to the horizontal, their function remained essentially the same. On the wider pedestals the clefts were sometimes vertical. On the adz shafts the clefts sometimes became mere incised lines, and in the continuous lozenge pattern they disappeared altogether. Though the names applied to the holes and clefts in district gods have not been recorded, it is probable that the terms ai tuna (holes) and kavava (clefts) applied by Gill to the carved adzes are a carryover from the terminology of the district gods.

Theories of Development

A myth widely spread throughout Polynesia asserts that the first human being created was Tiki. In most areas Tiki was a male, but in Mangaian myth Tiki was female. An object carved in human form was termed a tiki. In New Zealand the carved figures erected on the front gable ends of houses and the small neck ornaments made of jade are termed tiki, because they represent the human form. In Tahiti the images used by sorcerers were termed ti'i in distinction to the to'o objects representing gods which were not carved in human form. The term tiki was probably derived originally from the human ancestor Tiki, but the tiki objects are merely conventional products of the art developed in the area. They do not represent Tiki, but the ancestor that the Maori desired to commemorate or the evil spirit that the Tahitian sorcerer wished to enlist in his service. This use of the term tiki applies equally well in Mangaia and the other Cook Islands.

In the Cook Islands, the verb to carve is 'akatiki ('aka, causative); and, though it may have originally applied to carving the human form, it came to be used in current speech as applying to any form of ornamental carving. Sometimes the word tiki is reduplicated in the verb, as in 'akatikitiki. Gill (31, p. 223), after referring to the sharks' teeth pattern on Mangaian hafts, stated, "Other figures are each supposed, by a stretch of imagination, to represent a man squatting down (tikitiki tangata)." This statement led Stolpe (64), Giglioli (27), and others to regard the bar and chevron combinations on the Mangaian adz shafts as having developed by progressive degeneration from carvings of the human form.

Stolpe, who deserves the greatest credit for pioneering in the detailed study of the ornamental art of savages based on museum carved specimens, spoiled his chain of reasoning by grouping the Hervey (Cook) and Austral Islands together as one art area with an identical development. He noted that the continuous lozenge pattern was used on Raivavae paddles and on Mangaian adzes. By a carefully selected series of examples he was able to demonstrate that certain geometrical patterns on the Raivavae paddles were developed page 392by processes of elimination and fusion from the female human figures which were carved on the flared proximal ends of the paddles. This sequence of forms was convincing, because they were on objects carved by the craftsmen of the same island. Stolpe was able to demonstrate the disappearance of the head, loss of the forearms and lower legs, and the straightening out of the curved upper arms and thighs. Though this process appears definite enough, in the figures illustrated, I am not sure that in Raivavae the continuous lozenge pattern, which is identical with the Mangaian form (fig. 247, f), did not develop from a simple process of arranging the excised triangle technique in symmetrical form. However, Stolpe goes further in attributing a similar course of development to the geometrical patterns on Mangaian adz hafts. Thus, for the continuous lozenge pattern, which occurs on Austral Islands paddles and on Mangaian adz hafts, he holds that what I have described as the technical edge extending between the open angles of two chevrons represents the trunk and that the two arms of opposing chevrons represent upper and lower limbs (64, p. 26). Of the continuous lozenge pattern he says (64, p. 27, fig. 36, c) that "the fusion between arm-curves and leg-curves is so complete that legs and arms are formed by uninterrupted crossed lines enclosing regular chequers within which we can still always recognize the rudiments of the trunk—that is, the elevated sharp ridge, which in the drawing appears merely as a slender line."

Read (56, pp. 145-147) independently followed a method similar to Stolpe's to show that the Mangaian geometrical patterns were derived from human figures by a process of "degradation." He used a sequence of 16 drawings (56, pl. 14) from artifacts, of which five were from Tahiti and Huahine, four from the Austral Islands, and seven from Mangaia. The human figures that headed the sequence were from Tahiti and Huahine.

The fault of the method used by Stolpe and Read is the fact that the definite human forms had to be imported from Tahiti, Huahine, and the Austral Islands to introduce the sequences because no distinctly human forms could be fitted in from the art of Mangaia itself. Both authors recognized this weakness and relied on the statement of Gill (31, p. 223) that certain patterns on the Mangaian adz hafts "by a stretch of imagination" were supposed to represent the human form in a squatting position (tikitiki tangata). On the evidence of the development from artifacts confined to Mangaia, I do not hesitate to conclude that Gill's informants rationalized the carving as they did the square holes in the pedestals and that they added the qualifying term tangata (man) to tikitiki, the general term for carving. The origin of the tikitiki carving was attributed to Rori. But in the district gods, also attributed to him, we find only the bar and chevron motifs, except in the unique specimen in the Oldman collection (fig. 243). Because the history of the district gods proves them older than page 393any carved adzes, I regard them as providing the fundamental principles of Mangaian carving art, which is based on narrow panels for decoration with the bar, chevron, and K-motifs. This pattern was followed on carved adz hafts paddles, bowls, wooden gongs, a fan handle, and a short club. In all these objects, the origin of the carving is purely geometrical with no trace of derivation from the human form. In the Oldman specimen (fig. 243), there is a departure from the general technique. By trimming the main arches, a rounded curved body is substituted for the flat surfaces of the other orthodox figures; the chevron motif, slightly modified, is retained at one end of the arched body; and this is balanced at the other end by the bisection of a smaller arch longitudinally to give the semblance of legs. This technique gives the appearance of a conventional human body, but I regard it as a development from the more general form shared by four other figures. On the evidence, I believe that Mangaian carving has undergone development from simple geometrical motifs and not "degradation" from the human figure.

The term "K pattern" was first used by Read (56, p. 146). Stolpe (64, p. 55) objected to the term on the grounds that it was applicable only when the pattern was mutilated by kavava (clefts or grooves) and ai-tuna (holes). This objection was based on the erroneous theory of "degradation", implying that the continuous lozenge pattern was older and that the simpler forms were due to the continuous lozenges being broken up by the clefts and holes. Stolpe said that it was better to name ornamental types, if possible, after their derivation and significance. He therefore preferred the term "tiki-tiki ornamentation" as being directly indicative of "the essence of the question", which to him was the human figure. The name preferred by Stolpe indicates the essence of the error, for any name based on derivation and significance will carry with it any mistakes made by the author in his interpretation of origin and meaning. Read's term, based on geometrical appearance, avoids such errors.

Read (56, p. 142) states that in many theories of the development of forms of primitive instruments and deviations, there are two defects. First, a good deal has to be assumed before the theory will bear criticism; second, the series of progressive stages can be traced equally well from either end, and it is impossible to say from internal evidence whether it is a case of development or degradation. Both defects are present in Read's theory of the degradation of Mangaian carving patterns. It is too much to assume that a composite sequence made up from three different groups of islands indicates what happened in one, Read selected the wrong end of the progressive stages and found degradation instead of development.

Function and Symbolism

So much has been written about the function and symbolism of the carved adzes that a detailed criticism of all the statements made would require a page 394volume in itself. It is necessary, however, to traverse some of the theories expounded.

The adzes of class 1 have been assigned definite functions as war adzes and objects of art that marked social distinction. Shafts of a size convenient to handle were carved with the bar and chevron combinations purely as ornamentation to enhance their value, and they had no hidden religious meaning. The fact that motifs similar to those on the district gods were carved on secular objects shows that the carvings were not sacred in themselves. The use of the objects indicated whether they were religious or secular, and the purely ornamental carving patterns were applied to both without restriction.

The large hafts of classes 2 and 3 have led to confused thinking. Haddon (38, p. 80), in discussing a pedestal adz, says "… for some reason or another they have become functionless through increase in the size of the handle, and by reason of the weakness caused by deep carving." What Haddon evidently meant was that these adzes could not be used to cut wood. However, they were made purposely in that form to serve a function other than that of cutting wood. Sayce (58, p. 128) makes a similar error when he writes, "In the Hervey Islands [Mangaia] the stone adz became a god by a gradual transition of associated meaning, and in doing so it developed a handle much too large, too elaborate, and too weak for use." Why should the use of stone adzes be confined to purely economic purposes? When the Mangaian crafts-men made certain adzes as symbols of a god, they mounted them on long elaborate hafts to make them into stationary objects in honor of their deity. The adz continued to function in a form that had developed within the native culture. The supply of basalt was unlimited and the supply of working adzes was not affected.

I have discussed elsewhere (76, pp. 132, 133) the view that large adzes were symbols of Tane-mata-ariki, the god of craftsmen, but I could get no information as to the exact part they played in religious ritual. It is probable that master craftsmen kept a carved adz and went through some ritual with this craft symbol, when occasion warranted. In Tahiti, Tane was the god of craftsmen, who put their working adzes "to sleep" in the temple the night before engaging in some important work and "awoke" them the next morning, both processes being accompanied by a set ritual. In an adz incantation from Aitutaki (70, p. 245), reference is made to putting the adz to sleep and awakening it. In a Mangaian chant (76, p. 134), reference is made to Una, grandfather of Rori, and the placing of his adz in the temple.

Tei Ara'ata ta'au nu, ta'au toki. In the Ara'ata temple is thy coconut, thy adz.
Ite 'ae'aenga i a Tane, Ua 'ui mai nga ta'unga, nga toa e, When timber (Tane) is dressed, Assembled are the expert craftsmen and the warriors,
Mai ki a Una e—. And they come to Una—.
page 395

In a culture in which working adzes underwent a temple ritual to obtain mana from the god Tane, there is little wonder that one area should have developed an adz symbol for the tutelary deity of craftsmen, Tane-mata-ariki.

The type of haft originally used with the religious symbols remains in doubt, but the likely form is that of class 2. In the more elaborate forms of this class, the projecting transverse panels with small legs at the lower end form the prototype of the later pedestal form. Such a pattern could have been suggested by the pedestals of the district gods (fig. 241).

Funerary monuments. Stolpe (64, p. 34), supported by Haddon (38, p. 81) and Giglioli (27), held that some of the carved adzes were funerary monuments connected with the "cult of the ancestors." His theory was based entirely on the label of an adz in a museum at Chambery which stated: "… The stone axe itself had belonged to a chief of Tahiti in Cook's time, and it was, after the owner's death, hafted in this manner that it might be preserved by his family as a remembrance. It was afterwards presented to a French officer who chanced to do the family some great service." Stolpe's drawing (64, p. 28, fig. 41) shows part of the carving, consisting of two horizontal panels with repetition of the fused K-motif (fig. 247, c) and an intervening deep rectangle with vertical rows of the repeated triangle motif (fig. 247, h). The carving is Mangaian, and it would be interesting to know why the family of a Tahitian chief of Cook's time was commemorated with a Mangaian adz haft bearing the triangle motif which was probably not used until after missionary contact in 1823. Stolpe goes on to discuss "ancestor worship" of a form that never existed and that could have no bearing on the case. Polynesian families have preserved ancestral objects as heirlooms, but that they constituted funerary monuments associated with ancestor worship is a theory supported by one erroneous label and may be summarily dismissed.

Peace axes. The theory that the carved adzes are "peace axes" is evidently post-Gill in origin, for Gill makes no reference to it. As already stated, Giglioli (27) bases his account on one modern adz haft carved by Miringa-tangi and information supplied by that untrustworthy individual through Colonel Gudgeon. The adz termed toki mahia had faces (vaero) of the haft carved with what were said to be the emblems (tohu) of the Ngati-akatauira and Ngati-tane tribes and Giglioli concludes that the adz was hafted to celebrate peace between these two tribes. I have pointed out (76, pp. 132, 133) that in the recorded accounts concerning the declaration of peace, no mention is made of the use of adzes as part of the ceremony. As the conquering tribe only ceased warfare when their opponents were so reduced in number that they could offer no further opposition, the use of a peace adz seems foreign to native psychology. Furthermore, there is no historical record of a war page 396between the Ngati-akatauira and the Ngati-tane tribes that could have given rise to Miringatangi's adz.

Tribal emblems. Giglioli (27, p. 2) states that five special patterns of carving were used as tribal emblems. The five tribes using them were Te Manahune, Ngati-tane, Ngati-akatauira, Ngati-ruanuku, and Ngariki. The Ngati-ruanuku is absent from the list of tribes of Mangaia (76, pp. 102, 103), and no mention of it is made by Gill. An analysis of the five carving patterns shows them to be various combinations of lozenges and rows of triangles which I regard as post-missionary in development. On the evidence, the symbolic use of "peace axes" and the development of carving patterns as "tribal emblems" appear to be neo-myths.

Genealogical records. Colley March (51) advanced the absurd theory that the carved hafts were genealogical records. Not the slightest evidence could be found in Mangaia itself in support of any material object having been used to record and to demonstrate lineages. The notched genealogical stick of New Zealand and the knotted braid records of the Marquesas are local developments which have no similar material expression in other parts of Polynesia.

Unlocalized Religious Objects

A number of objects in the London Missionary Society collection are attributed to the Hervey Islands, but no information is given as to the particular islands from which they came. They consist of one wooden and some feather and human-hair objects. They are figured here for record.

Wooden Object

An object with a middle staff and end blades is catalogued, "Hervey Group, Paddle-shaped idol, L.M.S. 55." The two blades are brought to a sharp edge at their ends and the artifact looks more like a digging or weeding implement than like an "idol" (fig. 248).

Figure 248.—Wooden artifact (British Mus., L.M.S., 55). 1, middle staff: length 447 mm.; perfectly round in section with diameter of 35 mm. 2, blade: curved, sharp end edge; length, 156 mm.; greatest width near staff, 105 mm.; thickness, 30 mm. 3, blade: curved sharp end edge, like 2; length 152 mm.; greatest width 152 mm. and thickness 28 mm., chip depressions on blade visible.

Figure 248.—Wooden artifact (British Mus., L.M.S., 55). 1, middle staff: length 447 mm.; perfectly round in section with diameter of 35 mm. 2, blade: curved, sharp end edge; length, 156 mm.; greatest width near staff, 105 mm.; thickness, 30 mm. 3, blade: curved sharp end edge, like 2; length 152 mm.; greatest width 152 mm. and thickness 28 mm., chip depressions on blade visible.

Feather and Hair Objects

The other artifacts consist of some detached feather ornaments, bundles ornamented with feathers and with human hair, and a feather shield.

page 397

Five individual objects tied together with a piece of foreign string were evidently odds and ends that were bundled together in the Missionary Museum. Four of them must have formed parts of the decoration of some more complex object, and the fifth, a doubled length of thick sennit covered with tapa and decorated with feathers, is complete in itself. It is bifurcated at one end into legs and its appearance and technique suggest affinity with the Cambridge Museum form of the Mangaian god Tangiia (fig. 232, a). For the individual objects, see figure 249.

Figure 249.—four feather objects (British Mus., L.M.S.). a, bunch of eight split black feathers (1), fairly stiff, projecting 154 mm. above lashing (2) of single coir fiber with seizing, 18 mm. long, and thickness at upper edge, 5 mm. b, coir cord holder: 1, three-ply cord, 135 mm. long to feather lashing (5); 2, end of coir fibers of cord (1), unraveled and projecting 172 mm. above lashing; 3, split black feathers, 108 mm. long from lashing; 4, outside layer of short white feathers, 37 mm. long above lashing; 5, lashing of fine oronga thread, extending 10 mm. c, pearl-shell holder: 1, unique pearl-shell holder, 80 mm. long, 8 mm. wide, 6 mm. thick, with half holes on one edge where evidently broken off after drilling holes, lower free part 44 mm. to feather lashing (6); 2, three tropic bird tail feathers and long black feathers (3) tied to holder by separate lashing (4); lower layer of short red feathers (5) tied over other feathers by a separate lower lashing (6). d, double braid holder: 1, sennit doubled at lower end, 13 mm. wide, 11 mm. thick; 2, white tapa wound around sennit, which is separated into two ends (3, 3) above and divided tapa wrapped around each limb, 9 mm. wide; 4, unraveled sennit fibers fluffed out for 195 mm. to form part of decoration; 5, tropic bird tail feathers added to each limb; 6, lashing of feathers and tapa to each limb; 7, lashing of lower end of tapa which includes a few small feathers. There was another coir cord holder similar to b, carrying two red and one white tropic bird tail feathers and a few small red feathers mixed with the white (b, 4).

Figure 249.—four feather objects (British Mus., L.M.S.). a, bunch of eight split black feathers (1), fairly stiff, projecting 154 mm. above lashing (2) of single coir fiber with seizing, 18 mm. long, and thickness at upper edge, 5 mm. b, coir cord holder: 1, three-ply cord, 135 mm. long to feather lashing (5); 2, end of coir fibers of cord (1), unraveled and projecting 172 mm. above lashing; 3, split black feathers, 108 mm. long from lashing; 4, outside layer of short white feathers, 37 mm. long above lashing; 5, lashing of fine oronga thread, extending 10 mm. c, pearl-shell holder: 1, unique pearl-shell holder, 80 mm. long, 8 mm. wide, 6 mm. thick, with half holes on one edge where evidently broken off after drilling holes, lower free part 44 mm. to feather lashing (6); 2, three tropic bird tail feathers and long black feathers (3) tied to holder by separate lashing (4); lower layer of short red feathers (5) tied over other feathers by a separate lower lashing (6). d, double braid holder: 1, sennit doubled at lower end, 13 mm. wide, 11 mm. thick; 2, white tapa wound around sennit, which is separated into two ends (3, 3) above and divided tapa wrapped around each limb, 9 mm. wide; 4, unraveled sennit fibers fluffed out for 195 mm. to form part of decoration; 5, tropic bird tail feathers added to each limb; 6, lashing of feathers and tapa to each limb; 7, lashing of lower end of tapa which includes a few small feathers. There was another coir cord holder similar to b, carrying two red and one white tropic bird tail feathers and a few small red feathers mixed with the white (b, 4).

The formation of a neat cylindrical sennit bundle, ornamented with feathers and covered with a lozenge design, is reminiscent of the Atiu technique. The longitudinal folding of thick sennit to form a body is also present in the god Tangiia (fig. 232) attributed to Mangaia. However, the use of red and green parakeet feathers indicates Atiu, because these feathers seem to have been rare in Mangaian work. (See figure 250.)

page 398
Figure 250.—Sennit and feather bundle (British Mus., L.M.S., 63). Bundle formed of thick sennit, 9 mm. wide by 5 mm. thick, folded in lengths of 144 mm., with top and bottom turns (1) kept level; inserted in center of upper end is a plume (2) of five feathers, three narrow split black, one wide split black, and one tropic bird white tail feather, with upward projection of 170 mm.; bundle compressed into cylindrical form by close transverse turns (3) of same sennit; split black feathers and tropic bird white tail feathers (4) arranged around upper periphery and fixed by close transverse turns of coir two-ply cord, and below, another layer of feathers attached in same way (5); piece of black tapa (6), 126 mm. deep, wrapped around cylinder to conceal lower ends of feathers; about 18 mm. below upper edge of tapa, red feathers tied in tufts with two-ply oronga thread or single coir fiber, arranged in rows (7) on outer surface of tapa and fixed by spiral turns of single, thick coir fiber, fresh fibers being joined by reef knots; small green feathers alternated in rows with red; feathers cover a depth of 13 mm. but probably reached to top edge of tapa originally; rest of cylinder, 91 mm. in depth, covered with sennit worked in Rarotongan multiple lozenge design (8) in two colors, natural and red stained; some tapa wrapping (9) projects below lozenge design; diameter of completed cylinder at top, 70 mm. and bottom, 60 mm.

Figure 250.—Sennit and feather bundle (British Mus., L.M.S., 63). Bundle formed of thick sennit, 9 mm. wide by 5 mm. thick, folded in lengths of 144 mm., with top and bottom turns (1) kept level; inserted in center of upper end is a plume (2) of five feathers, three narrow split black, one wide split black, and one tropic bird white tail feather, with upward projection of 170 mm.; bundle compressed into cylindrical form by close transverse turns (3) of same sennit; split black feathers and tropic bird white tail feathers (4) arranged around upper periphery and fixed by close transverse turns of coir two-ply cord, and below, another layer of feathers attached in same way (5); piece of black tapa (6), 126 mm. deep, wrapped around cylinder to conceal lower ends of feathers; about 18 mm. below upper edge of tapa, red feathers tied in tufts with two-ply oronga thread or single coir fiber, arranged in rows (7) on outer surface of tapa and fixed by spiral turns of single, thick coir fiber, fresh fibers being joined by reef knots; small green feathers alternated in rows with red; feathers cover a depth of 13 mm. but probably reached to top edge of tapa originally; rest of cylinder, 91 mm. in depth, covered with sennit worked in Rarotongan multiple lozenge design (8) in two colors, natural and red stained; some tapa wrapping (9) projects below lozenge design; diameter of completed cylinder at top, 70 mm. and bottom, 60 mm.

A cylindrical tapa bundle without an inner core of sennit offers a change of ornamentation in the use of human hair. It is interesting in that it has three spaced lashings with a continuous length of sennit in a technique used on some Mangaian adz shafts and weapons. Short white feathers and a tropic bird tail feather are also used. The absence of red feathers is negative evidence in favor of Mangaia, where red feathers seem to be scarce as compared with their frequent use in Aitutaki and the Atiu group which culturally includes Mauke and Mitiaro. (See figure 251.)

page 399
Figure 251.—Tapa and human hair bundle (British Mus., L.M.S., 64): bundle composed of cylindrical folds of white tapa, 235 mm. long, with black frizzy hair (2) enclosed by upper end of tapa and fixed by transverse turns of sennit at top, middle, and bottom; top diameter 38 mm.; middle diameter 58 mm.; bottom diameter 30 mm.; hair length above bundle 240 mm.; a few white feathers and a tropic bird white tail feather are caught under top lashing; in top lashing (3), sennit commencement end is bent upward and a number of turns made over thumb laid vertically against bundle, sennit after last turn at top is passed downward under turns, thumb removed, and turns drawn taut from below upward which covers and fixes commencement end; sennit is pulled downward to remove slack and a complete top lashing; sennit is car ried down to middle lashing (4) where three turns are made from below upward over thumb, sennit passed down under turns, turns tautened from below upward, and slack removed by pulling sennit free end downward as in upper lashing; bottom lashing (5) is made in exactly same way and sennit, after pulled downward to remove slack, is cut off below lashing. The technique is like that used on some adz shafts (fig. 116, a).

Figure 251.—Tapa and human hair bundle (British Mus., L.M.S., 64): bundle composed of cylindrical folds of white tapa, 235 mm. long, with black frizzy hair (2) enclosed by upper end of tapa and fixed by transverse turns of sennit at top, middle, and bottom; top diameter 38 mm.; middle diameter 58 mm.; bottom diameter 30 mm.; hair length above bundle 240 mm.; a few white feathers and a tropic bird white tail feather are caught under top lashing; in top lashing (3), sennit commencement end is bent upward and a number of turns made over thumb laid vertically against bundle, sennit after last turn at top is passed downward under turns, thumb removed, and turns drawn taut from below upward which covers and fixes commencement end; sennit is pulled downward to remove slack and a complete top lashing; sennit is car ried down to middle lashing (4) where three turns are made from below upward over thumb, sennit passed down under turns, turns tautened from below upward, and slack removed by pulling sennit free end downward as in upper lashing; bottom lashing (5) is made in exactly same way and sennit, after pulled downward to remove slack, is cut off below lashing. The technique is like that used on some adz shafts (fig. 116, a).

A unique feather object in the form of a shield with a looped wooden handle at the back was probably used in dances rather than in religious ritual. The use of lengths of thick sennit covered with feathers on one surface and fastened by continuous spiral turns of thread, has affinity with the technique of Atiu. A braid ornament with cross pieces of coir fiber is also characteristic of the Atiu area. The braid lengths are attached to a wooden frame. Short lengths of thick sennit with slanting feathers secured by spiral turns of a thread are used in an ornament in the Aitutaki headdress (fig. 46, b). (See figure 252.)

A long cylinder partly covered with feathers is unique in having a feather cap held in position by a feather covered collar. The artifact was originally attributed to Tahiti, but is now labeled, "Sandwich deity adorned with red feathers." It certainly does not belong to Hawaii. The technique is reminiscent of the Cook Islands, and the free use of red feathers indicates Atiu or the neighboring islands as its origin. The cap and collar technique, however, is page 400
Figure 252.—Feather shield (British Mus., L.M.S., 579). a, wooden framework three longitudinal rods (1, 2, 3), 715 mm. long; the two lateral rods (1, 3), 7 and 9 mm. thick and middle rod, 13 mm. thick, all slightly thinner at top; middle rod has handle (4) on back, about 93 mm. long and raised 35 mm. above rod plane; five crossbars about 7 mm. thick but lowest 9 mm. thick, spaced against front of rods, and lashed to them with oronga cord. A piece of dark tapa folded to right dimensions lies on front of crossbars between outside rods and upper and lower bars, b, front: eight lengths of sennit, 10-13 mm. wide, 732 mm. long, covered on outer surface with tropic bird red tail feathers, attached by continuous spiral turns of an oronga thread from above downward, over front surface of frame; lower ends of sennit lengths hang below lowest bar and thread passes between individual lengths to keep lower ends together braid ornament (1) with coir side tufts hangs down free from the top bar. c, back nine lengths of braid similar to those in front but covered with white and with black feathers attached to framework, spiral thread fixing feathers also passing around crossbars as it reaches them; sennit lengths at each side edge are attached to lateral rods; braid length in line with the mesial handle (4) is cut off above handle and another short length is attached below it thus leaving handle clear; lower free ends are also connected by a thread passing between their ends. d, some tropic bird tail feathers. (1) lashed at a slant to separate short lengths of sennit (2) which were lashed to longitudinal lengths as extra ornaments; bunches of small red feathers were attached; to quill end of some tail feathers. e, sennit streamer technique: long free tufts of coir fiber (1) were added on either side to braid fibers (2); streamers 660 mm. long.

Figure 252.—Feather shield (British Mus., L.M.S., 579). a, wooden framework three longitudinal rods (1, 2, 3), 715 mm. long; the two lateral rods (1, 3), 7 and 9 mm. thick and middle rod, 13 mm. thick, all slightly thinner at top; middle rod has handle (4) on back, about 93 mm. long and raised 35 mm. above rod plane; five crossbars about 7 mm. thick but lowest 9 mm. thick, spaced against front of rods, and lashed to them with oronga cord. A piece of dark tapa folded to right dimensions lies on front of crossbars between outside rods and upper and lower bars, b, front: eight lengths of sennit, 10-13 mm. wide, 732 mm. long, covered on outer surface with tropic bird red tail feathers, attached by continuous spiral turns of an oronga thread from above downward, over front surface of frame; lower ends of sennit lengths hang below lowest bar and thread passes between individual lengths to keep lower ends together braid ornament (1) with coir side tufts hangs down free from the top bar. c, back nine lengths of braid similar to those in front but covered with white and with black feathers attached to framework, spiral thread fixing feathers also passing around crossbars as it reaches them; sennit lengths at each side edge are attached to lateral rods; braid length in line with the mesial handle (4) is cut off above handle and another short length is attached below it thus leaving handle clear; lower free ends are also connected by a thread passing between their ends. d, some tropic bird tail feathers. (1) lashed at a slant to separate short lengths of sennit (2) which were lashed to longitudinal lengths as extra ornaments; bunches of small red feathers were attached; to quill end of some tail feathers. e, sennit streamer technique: long free tufts of coir fiber (1) were added on either side to braid fibers (2); streamers 660 mm. long.

page 401 unique. The artifact must have been a pleasing sight when the feathers were fresh. (See figure 253.)
Figure 253.—Tapa covered cord bundle with feather cap (British Mus., L.M.S., 576). a, cylindrical body (1) of thick two-ply coir cord folded in lengths of 375 mm., covered with white tapa; near upper end (2), sets of 4 red feathers from wings or tail of parakeet bound together with single coir fiber and kept in position with spiral turns of oronga two-ply thread, feathers with tips downward and lowest layer about 130 mm. from upper end; at top end, small red and yellow feathers set close together; cap of feathers (3) used as cover over upper end of cylinder and kept in position by tapa collar (4). b, upper cap (a, 3) formed of thin white tapa foundation covered with fine net of oronga with closely set small red feathers tied to net meshes and stitched also to tapa foundation; cap folded over upper end of cylinder and tied to cylinder by transverse turns (5) of fine oronga cord about 1 mm. thick. c, collar (4) formed of folded tapa to about 5 mm. wide, with alternate rows of small red and yellow feathers fixed to outer surface of collar by two-ply oronga thread in continuous spiral turns; collar fitted over cord lashing (b, 5).

Figure 253.—Tapa covered cord bundle with feather cap (British Mus., L.M.S., 576). a, cylindrical body (1) of thick two-ply coir cord folded in lengths of 375 mm., covered with white tapa; near upper end (2), sets of 4 red feathers from wings or tail of parakeet bound together with single coir fiber and kept in position with spiral turns of oronga two-ply thread, feathers with tips downward and lowest layer about 130 mm. from upper end; at top end, small red and yellow feathers set close together; cap of feathers (3) used as cover over upper end of cylinder and kept in position by tapa collar (4). b, upper cap (a, 3) formed of thin white tapa foundation covered with fine net of oronga with closely set small red feathers tied to net meshes and stitched also to tapa foundation; cap folded over upper end of cylinder and tied to cylinder by transverse turns (5) of fine oronga cord about 1 mm. thick. c, collar (4) formed of folded tapa to about 5 mm. wide, with alternate rows of small red and yellow feathers fixed to outer surface of collar by two-ply oronga thread in continuous spiral turns; collar fitted over cord lashing (b, 5).