Arts and Crafts of the Cook Islands
There was never a large variety of musical instruments, but now their number has dwindled to slit gongs and shell trumpets. The ancient shark-skin covered drum, formerly so important in festivals and religious ritual, has been supplanted by a copy of the English drum covered with cowhide and beaten with a drum stick during demonstrations of native dances. With the exception of carved specimens of the slit gong, shark-skin covered drums, and two ancient shell trumpets, I have failed to locate old specimens of Cook Island musical instruments in museums and private collections.
Following Hornbostel (43), the instruments may be classified as autophones (slit gongs), aerophones (flute, shell trumpet), and membranophones (shark-skin drum). There were no cordophones or string instruments.
Autophones are instruments which are struck directly, the sound being produced without the aid of a medium. These instruments, which are often wrongly referred to in literature as drums, are called slit gongs here for want of a better term. There are two varieties of slit gongs, the small type still in use and the larger, carved form preserved in museums. They were evidently used throughout the island group.
The small slit gong (tokere) was made from a section of a branch of tamanu or miro. Part of the surface was trimmed fiat longitudinally and a slit cut along it. Through the slit the interior was hollowed out. In most gongs, the long middle slit does not extend quite to the ends but is separated from short end slits by septums of unremoved wood. A different note is obtained by changing the beat from the middle to the ends. A gong with end slits was figured by me (70, p. 355) for Aitutaki, and one of each variety from Rarotonga is shown in figure 162.
Figure 162.—Rarotongan slit gongs (Bishop Mus., a, b: C8865; c, d: C8864). a, side view: total length, 500 mm.; 1, main slit, 340 mm. long, 32 mm. wide; 2, 2, end septums, 45 mm. long; 3, 3, end slits, 37 and 39 mm. long, same width as main slit; 4, beating stick (C8866): length, 341 mm.; middle diameter, 21 mm.; pointed at both ends, b, end view: height, 105 mm.; greatest width, 103 mm.; end slits roughly cut out below upper opening, c, side view: total length, 388 mm.; 1, single slit, 310 mm. long; width, 30 mm. but worn under in middle from beating; 2, 2, end septums or walls, 41 and 43 mm. long; no end slits; 3, beating sticks (C8867): length, 330 mm.; middle diameter, 14 mm. d, end view: height, 111 mm.; greatest width, 109 mm.
Figure 163.—Mangaian carved gong (British Mus., L.M.S. 488): length, 647 mm.; height, 220 mm.; greatest width, 207 mm. a, upper flat surface defined at ends by raised bands (3) and at sides by horizontal lines of carved panels, width 80 mm.; opening into interior hollow is triangular at each end (1, 1) with the bases outward, sides slightly curved and apices connected by a narrow slit (2) which is 127 mm. long; base of left triangle is 52 mm. and length from base to apex, 195 mm.; right triangle base is, 56 mm. and length, 200 mm.; parts of upper surface not occupied by opening are uncarved. End bands (3, 3), encircling gong, are 50 mm. wide, corresponding to thickness of end walls and are divided by thin slits into small rectangular spaces for carving. End knobs (4), projecting at a lower plane than end bands, are 28 in number, evenly spaced, forming small rectangular surfaces 13 mm. long, b, side view: same numbering as a. c, end view: shows triangular section of knobs (4) with apices inward, also black painted circle for ornament, d, enlargement of one end: end knobs (1) with upright K and notched bar motif; raised end panel (2) divided by two vertical slits into three panels which in turn are divided into small rectangular spaces by horizontal slits, each space carved into two K-motifs facing each other with stems horizontal. Body (3) of gong is divided into horizontal panels (4) by wide grooves (5); panels are subdivided into small spaces by cross slits; each rectangular space is carved with double K-motif with stem horizontal as in end panels (2); bottoms of dividing grooves (5) are stained in black to form light lozenges of natural wood color.
Large gongs were made from a section of tree trunk trimmed to a flat upper surface from which the interior was hollowed out through two long triangular openings connected by a median narrow slit. At the ends, raised bands correspond in width to the thickness of the end walls and encircle the gong, but these bands are flattened on the upper surface. Beyond the raised bands, short, regularly spaced projections extend around the circumference on the same lower plane as the body of the gong. The British Museum gong is selected as the type specimen (fig. 163).
The Peabody Museum gong is larger than that in the British Museum, but it conforms with the type specimen in general principles. The slit opening is similar; end raised bands and end projections are present; and the end bands and body of the gong, except the upper surface, are divided into panels carved with the K-motif. There are slight variations in the carving of the end bands, the wide painted grooves between the horizontal panels on the body are absent, and the K-motifs are vertical instead of horizontal (fig. 164). The outer ends of the gong are painted as in figure 163, c, but the figures are elliptical instead of circular.page 266
Figure 164.—Mangaian carved gong (Peabody Mus., Cambridge, 53509). a, view from above, showing triangular openings with connecting slit; total length, 700 mm.; height, 272 mm.; greatest width, 270 mm. End projections (1): length, 18 mm.; width, 12 mm.; triangular in section, with apices inward and length from base to apex, 28 mm.; 37 projections at each end. End bands (2) are divided into four vertical panels (see b). Body (3) of gong is divided into horizontal panels by narrow slits without wide grooves as in figure 163 and horizontal panels are again divided by vertical slits into small rectangular spaces for carving with K-motif. b, carving details on one end of drum: end projecting knobs (1) carved with upright K and notched bar: end band (2-5) divided into four panels; outer panel (2) at slightly higher plane than knobs, carved with zigzag lines and with triangles at outer edge cut down to level of knobs; two middle panels (3, 4) at higher plane with rectangular spaces carved with upright double K-motif; inner panel (5) at lower level carved in same way as outer panel (2). Body (6) of gong with rectangular spaces carved with upright double K-motif.
The Christchurch gong is similar to the type gong in having wide grooves between the horizontal panels of the body, the K-motif horizontal on the end band and body, and a single upright K on the end projections. It has a slight variation in four vertical panels on the end bands instead of three.
I examined the Leipzig gong in 1933 through the glass of the case on a Sunday when I could not get closer access. However, the sketches made show some interesting variations from the type specimen. The opening on the fiat upper surface is the same as in the type, forming two end triangles with straight outer bases, curved sides, and a mesial narrow slit between the apices. The page 267end projections are carved with a double upright K but the end bands are painted except at one end where the flat upper surface is lined into small rectangular spaces only partly carved. The body is divided into horizontal panels extending the full length of the body and separated by wide grooves which resemble those of the type gong in that they are painted black to show up spaced lozenges in the natural wood color. The panels, however, are not divided vertically into rectangular spaces but are simply notched and enhanced with incised lozenges. The carving is not completed and it is evident that the instrument was acquired before the carver finished his work. (See figure 165.)
Figure 165.—Mangaian gong (Grassi Mus., Leipzig), a, shows side and upper surface: 1, triangular opening; 2, median slit of opening; 3, raised end band, divided into rectangular spaces and only partly carved; 4, end projections. b, enlargement of carved motif on right end band (a,3). c, enlargement of carving on end projections showing upright opposing K-motif. d, horizontal panels (1, 1) on body of gong, symmetrically notched on both sides and enhanced with incised lozenges with small excavated lozenges in center; wide groove (2) between notched panels painted black to leave single row of unstained lozenges, e, oblique cross motifs with thick arms with rounded ends arranged in alternating rows and painted black. f, two alternating rows of oblique cross motif fused together by joining their adjacent arms. Painted motifs (e, f) are on end bands (except part carved) and on upper surface on either side of opening.
The Rarotongan gong differs from the Mangaian form in that the straight base of the triangular openings are curved on the round, the ends are simply crisscrossed with oblique lines, and the body is not carved (fig. 166).
A slit gong in the British Museum (L.M.S. loan, 1890) has been figured by Edge-Partington (24, 1-16-5). The slit opening is wide and of even width throughout. End rims are present but short and not divided into knobs. The page 268two edges of the slit and the upper surface of the ends which close the slot opening are carved. The under surface is plain but both sides are carved in rectangular areas at end end leaving a middle rectangle uncarved. On each of the middle plain areas are round depressions worn by the beating sticks. The main carving motif is the deep herringbone pattern characteristic of the Austral Islands, to which it probably belongs. This gong approaches the Mangaian gong in beauty and workmanship.
Figure 166.—Rarotongan carved gong, Ngatangiia village. Shows upper surface and side: upper opening shows curved ends (1, 1) and a longer and wider slit (2) than in Mangaian gong; curved incised lines present on either side of ends of opening. Ends are decorated with crossed lines (3) corresponding in position to raised bands of Mangaian gongs and extending around gong except on upper surface. End knobs (4) are cut in at bases, are not triangular in section, and are uncarved.
Aerophones or wind instruments were bamboo flutes and shell trumpets. There were probably toy trumpets made of coconut leaf twisted into a cone by spiral turns, but I obtained no information concerning them.
Bamboo flutes (ko'e) obtained their name from the native name for bamboo (ko'e). These have long been abandoned, and I know of no authentic specimens preserved in museums. Our knowledge of them is derived from references in literature and in native songs. Gill (28, pp. 232, 233), in the story of the Mangaian hero Ngaru, records that Ngaru went to the sky land of Taumareva "where fruits and flowers grew profusely, and the inhabitants of which excelled in flute-playing." In a note to the above statement, he says that the flute was "a. piece of bamboo pierced with three holes, and blown through the nose." In Aitutaki (70, p. 354), I was told that the septa at the nodes of the bamboo other than those at the top were pushed out with a stick and that four holes were made. However, I prefer Gill's earlier statement that there were three holes and that the instrument was played presumably with the nose at the top hole and the fingers at the other two.
Figure 167.—Mangaian shell trumpets (British Mus., a, E.P. 35; b, no number). a, length, 15.75 inches, shows hole (1) side and also a patch (2.) of shell below blowing hole, patch being stuck on with gum, probably breadfruit gum. b, a smaller trumpet: length, 7.1 inches, a hole bored through edge of large aperture and a short length of sennit tied through it.
Figure 168.—Aitutaki drum (Auckland Mus., 987). a, wooden body (1): height, 495 mm.; greatest diameter near top, 407 mm.; bottom diameter, 350 mm.; interior partition, 146 mm. from bottom. Upper opening edge (2) shows against skin cover; inner diameter, 222 mm. Shark-skin cover (3) in one piece covers upper opening and overlaps from 75 to 100 mm.; single strip of shark skin, 64 mm. wide, placed under outer edge of cover to reinforce part where holes are cut through both layers, 12 mm. from edge, 32 to 39 mm. apart, 32 holes in all. Rectangular slots (4): height, 64 mm.; width, 25 mm.; 25 mm. apart; 57 mm. from bottom edge, 21 slots in all; lower rim, 37 mm. thick. Sennit lashing (5, 6) in pairs, one descending (5), the other ascending (6). Pandanus leaf strip (7) wrapped around bottom rim to hide lashing turns and knots and kept in place by cord (8) passing through slots and around bottom rim. b, drum head fixation: 1, lower edge of shark-skin cover showing triangular hole cut through it; 2, inner strip of shark skin projecting below lower edge of cover, showing in triangular aperture of cover, and with small round hole for passage of sennit; 3, lower part of slot corresponds to cover holes above; 4, lower rim of drum. Sennit is cut into lengths, each over twice length between cover and lower rim. First length of sennit (5) was fixed temporarily on left, passed through holes in cover and underlying strip, and descended to slot below where it passed back around rim (4) to which it was temporarily tied. Another sennit length (6) was temporarily tied to lower rim to left of first length (5) and in same slot. It ascended on left, passed over cover edge, under loop formed by first length (5), and turned to right where it passed through next cover holes to descend to slot below to be tied to lower rim. This procedure was followed with all cover holes working from left to right. Each new length was tied to lower rim in same slot in which previous length ended, and on its left. It ascended over cover edge, under previous loop, passed to next cover hole on right, passed down through it under cover edge, descended to slot below and was tied to lower rim. When last length passed down through last cover hole and was tied to rim, first length (5) was detached from its temporary attachment, and its first half passed back to left to be passed down under loop of last length and to be fastened below to lower rim on left of last length in same slot. As there were only 21 lower slots, to 32 cover holes, 11 slots took two pairs of cords but this did not alter technique of fastening. Each pair of sennit lengths in same slot was drawn as taut as possible and retied to remove slack. Beginning end of left cord was usually left long and after drawing left cord taut and taking a turn around rim, end was passed back in a loop (7) around pair which were drawn downward and forward to further tighten them. End was then tied to rim or to surplus end of its fellow pair. Knots were on back or lower surface of rim, and these as well as turns around rim were concealed by covering front, back, and lower edge with strips of pandanus leaf, as shown in a,7. c, section of lower rim, showing how loop (7), by drawing sennit pair downward and forward from its previous position indicated by dotted lines, further tautens braid pairs.
The drum (pa'u) in its Polynesian form, was present in all the Cook Islands. It was made from a section of trunk from the following trees: tamanu, tou, breadfruit, puka, or 'utu. The round block was hollowed out from both ends to leave a wooden partition a little less than a third from the lower end. A circular piece of shark skin was fitted over the upper opening and lashed down by lengths of sennit which were drawn taut and fastened to the lower end. The drum was placed upright on the ground, and the shark-skin membrane was beaten with the bare hands.
The only Cook Islands drums of an old pattern that I know of are four from Aitutaki; two in the Auckland Museum, New Zealand; one in the Cranmore Museum, Kent, England; and one in the Peabody Museum, Cambridge. One of those in the Auckland Museum was presented to me in Rarotonga, but its origin in Aitutaki was attested to by the donor and reliable native informants. When I subsequently saw a number of drums without shark-skin covers in Aitutaki, they conformed in shape and certain details to the Auckland drum. Failing specimens from other islands, the Aitutaki drums may be accepted as illustrating the technique of the group. The Auckland drum (no. 987) already described (70, pp. 357-360) is selected as the type specimen, for I was enabled to identify the other two drums by means of it. The information concerning it is repeated here in figure 168, and the painted pattern on the body of the drum is repeated in figure 169.
Though shape, size, and decoration by carving and painting must have varied in drums of different islands, it is probable that the lashing technique remained the same throughout the group. This technique involved two problems; the upper fixation to the shark-skin cover and the lower fixation to some part of the woodwork.
The upper fixation of the Aitutaki drum was made by the sennit lengths passing directly through holes in the shark-skin cover, though they pass through a loop of the previous length before entering the cover hole. This method may therefore be termed the direct method to distinguish it from an page 272indirect method in which a separate length of material is threaded through the holes around the circumference of the cover and in which the tautening braids are looped over this intermediate element without themselves passing through the cover holes. The lower fixation, through lower rectangular slots around the bottom rim of the drum, may be termed the single slot fixation.
The second Auckland drum (no. 8865) is similar in shape to the type specimen, but the shark-skin cover does not overlap so much. The upper fixation is direct but from right to left. The lower fixation is by the single slot method which includes the tautening loops, but there is no pandanus covering over the lower rim. The rather rough painted pattern in black is shown in figure 170.
Figure 170.—Painted design on Aitutaki drum (Auckland Mus., 8865): lower end of drum showing slots (1-1) with oblique crosses between slots and above, two sets of oblique lines crossing at ends and with short curved lines on either side of upper crossings; also oblique crosses between upper sets of crossings.
The Cranmore drum resembles the type specimen closely in shape, cover overlap, direct upper fixation from left to right, lower single slot fixation with tautening loops, and pandanus covering of lower rim. Faint traces of a painted pattern cannot be made out in detail.
A fourth drum in the Peabody Museum, Cambridge, Massachusetts, is probably from Aitutaki, but it offers some differences that must be considered. The resemblances are shape, cover overlap, and upper direct fixation from left to right. The lower fixation, however, is by a double slot method in which there are two rows of alternating rectangular slots near the bottom rim. The braid lengths enter the upper slots and evidently pass obliquely to the lower slots, hence the fixation turns are around the wood between the two rows of slots. Thus the bottom rim is left clear, and it is carved on the outer side with triangular motifs. The wood between the upper row of slots is carved with a leaf pattern. There are traces of triangular and lozenge motif in faded blue paint on the body of the drum (fig. 171).
Figure 171.—Aitutaki drum (Peabody Mus., Cambridge, 86,236): a, height, 672 mm.; upper opening diameter, 240 mm.; bottom diameter, 330 mm.; dividing septum, 265 mm. from bottom: upper set of 15 rectangular slots (1), 75 mm. high and 37 mm. wide; top edge of slots, 197 mm. from bottom edge: lower set of 15 rectangular slots (2), 50 mm. high and 30 mm. wide; ends of lashing braids (3) wound round between slot rows: lower edge (4) clear of lashings. b, top of drum, showing direct lashing with flat braid worked from left to right. c, lower end showing alternating slots of upper (1) and lower (2) rows with incised leaf design (3) between upper slots and opposed triangles (4) forming hour glass effect between lower slots and bottom edge. Faint designs in faded blue paint on body show rows of triangles with bases touching on same line and also small lozenges.
It is difficult to decide whether the double slot fixation is merely a variation of Aitutaki technique or whether, carrying as it does the clearance of the lower rim and the presence of carving, it constitutes a distinct difference due to the technique of another island in the Cook group.
The musical instruments, though few, played an important role in social and religious life. The bamboo flute was the least important, but young men were able to express love messages by means of its plaintive notes.
In Rarotongan tradition (65, vol. 28, p. 121), the great ancestor Tangiia was said to have obtained the sacred dances (eiva) from the gods in 'Avaiki and with them various material things. Among them were the pa'u (drum), the pu (shell trumpet) and the ka'ara (slit gong). These were placed in the care of Taote, and on the return voyage the drum was beaten (kua rutu te pa'u) and the trumpet was sounded (kua 'akatangi te pu) at each island they reached. On arriving at Kuporu, it was found that the shell trumpet had been left behind at Uea, so the ship returned for it.
The shell trumpet was used to make announcements and assemble people. It served as a military bugle in warning warriors to mobilize. Gill (30, p. 54) states that Vaeruarau, king of Mangaia, kept his royal war conch shell in the marae of Ariana. On a certain occasion, he sounded it long and loud. "The warriors of the neighborhood started out of their sleep at this summons, adjusted their war dresses in the moonlight, and assembled forthwith." No one but the reigning sovereign could use this royal trumpet, which was given to Gill and is probably one of the two in the British Museum. Gill (30, p. 229) also states that Rongo, the principal god of Mangaia, was represented by a shell trumpet used only by the king and kept near the entrance of the national god house.
The shark-skin drums were of even greater importance, as they were used in both civil and religious observances. A drum named Tangimoana (To-sound-at-sea), famous in Rarotongan tradition, was held to have been used on the great marae of Taputapuatea in Raiatea and when John Williams (81, p. 88) first visited Rarotonga, he was asked the whereabouts of the drum. The drum was important to the religious ceremonies connected, with the installation of the military dictator of Mangaia. An old Mangaian song (28, p. 262) says, "E uru tupu ariki te 'apai o te pa'u e" (The installation of a high chief was the purpose for which the drum was used).
In addition, the beating of the drum announced peace. Gill (28, p. 298) writes:
And now the famous drum of peace, expressly made for this solemn occasion, would be beaten; or, strictly speaking, would be heavily struck with the tips of the fingers. A feast occupied the attention of the warriors and chiefs between the presentation of the bits of ears and the drumming. The performance first took place on the marae of Rongo; a procession was now formed of all the victorious tribes, headed by the king and the hereditary drum-beater, who carried the big drum. This object of mysterious reverence was simply part of a tree, dug out at one end with stone adzes; the aperture being covered with a piece of shark's skin. Each relative of the hereditary drum-player carried a small page 275drum, to increase the volume of sound, thus assuring fugitives hiding in the rocks and thickets that better days were dawning…. At a certain point all the males of the kingly families united their voices, and all the drums sent forth their agreeable, although monotonous accompaniment.
The slit gong was used to accompany songs and dances. In Mangaia, it was particularly important in performances given to honor the dead some months after death had taken place. Gill (28, p. 270) states, "The accompaniments of this performance were the great wooden drum, called 'the awakener' (kaara), and the harmonicon. Sometimes the 'paû' was added. The musical instruments were called into use between each song; in the case of the 'showery' songs the great drum accompanied the grand chorus." Here Gill distinguishes correctly between paû[pa'u] or shark-skin drum and the slit gong or kaara[ka'ara] which he erroneously terms the great wooden drum. His translation of "the awakener" for kaara might be correct if the word consisted of ka ara (ara, to awake) but the word with the glottal closure is ka'ara and hence has a different root from ara.
It is curious that, though a number of the Mangaian ka'ara slit gongs have been preserved, no specimens of the more important shark-skin drums of Mangaia have been located. Such drums would surely have been carved in the characteristic Mangaian technique and so be readily identified.