Arts and Crafts of the Cook Islands
Plaiting is such a distinctive and widespread art in Polynesia and the fundamental technique is so similar when applied to different materials and various articles, that it is dealt with here as one section although it groups together artifacts that could be placed under other headings. I have already described the Cook Islands technique in detail (70) and, to avoid unnecessary repetition, I refer the student to this work.
Plaiting was women's work, but men made baskets of coconut leaves in the cultivations when women were not present. Food platters, baskets, mats, fans, eyeshades, and sheets used in house construction for thatch, ridging, and wall protection were made of coconut leaves.
Thatch sheets (Mangaia, kei'a) were made from a length of coconut leaf about six feet long, kept to the same length by a measuring rod. The leaf midrib was split down the middle and the extraneous thick material split off. The leaflets were plaited in check, alternate leaflets being turned opposite to their natural direction to provide crossing elements. The sheets were plaited singly or two were placed together and the corresponding leaflets of each plaited as double wefts. The plaiting was continued outward from the midrib for a few inches and the leaflet ends left free (fig. 20, a). The technical details have been described (70, pp. 7-11).
The ridging sheet (tapatu) was made in lengths of six feet or more, from the full length of a leaf and the two split halves of another length. The full leaf was placed in the middle and the halves on either side with midribs out-page 50ward so that the leaflets crossed those from either side of the middle leaf. The crossing leaflets were plaited in check and their ends left free (fig. 20, b). The technical details are described in my work on Aitutaki (70, pp. 25-29).
The wall sheet ('apuka) was made of two single thatch sheets placed so that the free ends of the leaflets from each sheet crossed each other in the midline. The leaflet ends of each sheet were plaited together in a three-ply braid and the end tail knotted (fig. 20, c). I have previously given details of this technique (70, pp. 31-33). These sheets were tied horizontally to the poles of the wall to keep out draughts. The same technique was used in Samoa (73, pp. 172-174) in the making of carrying sheets.
The general term for basket is kete. Three types are made from green coconut leaves, and one type is made from pandanus.
The o'ini basket (70, pp. 168-171) is a development of one of the platters (fig. 21, b) in which the leaflet ends are plaited to form sides and a handle. It was used for cooked food and fruit (fig. 21, c). This type of basket is said to have been introduced from Tahiti, where the identical form and name are page 51found. It is also made in Samoa (73, p. 205, pl. 14, d,4), where it is said to be a fairly recent introduction.
Figure 21.—Coconut-leaf platters and baskets: a, platter made with two midrib strips (1, 2), plaited in check and knotted (3, 4); b, platter of four midrib strips (1-4), plaited in check, leaflet ends knotted (5-8); c, small basket (o'ini), with leaflet ends braided into handles (1, 1); d, basket plaited from strip of whole leaf and midrib split (1, 2) after completion; e, permanent basket plaited in twill.
The tapora basket (Mangaia, peru) was made from a piece cut off toward the tip of the leaf with an equal number of leaflets on each side of the midrib. One side was plaited like the roof sheet, in check with the leaflets opened out. On completion of one side, the plaiting was continued round the end by taking in the leaflets of the other side and closing the second end with the free leaflets left over on each side. The free ends of the leaflets were plaited in a three-ply braid to close the bottom which was termed tekere (keel). The leaf midrib was then split down the middle in its full length and formed the basket rim (fig. 21, d). The full technique has been described (70, pp. 172-179). These baskets were made in cultivations and were used for carrying home uncooked food such as taro, husked coconuts, and other products. After use they were discarded, as single open leaflets become brittle when dry.
Another kind of tapora (70, pp. 179-181) was made from a single strip of midrib cut twice the required length of the basket. A continuous sheet was page 52plaited in check with the open leaflets and the end closed as it was in the other form. The leaflet ends were plaited in braid to close the bottom, and the midrib edge formed the rim. The two ends of the midrib strip were tied together by a strip of midrib from one end (70, p. 180, fig. 162). This differs from the Samoan technique of twisting a leaflet from one end of the strip around the other end and including it in the plaiting (73, p. 190, fig. 97, b). This basket was used for the same purposes as the other tapora. The Mangaians held that this one was the true o'ini and that the o'ini imported from Tahiti was called po'ini in the Society Islands. However, Willowdean Handy describes the Tahitian form (42, pp. 21-35) under the name of oini.
The kete nikau (70, pp. 181-190) was the best form of coconut-leaf basket. It was made with two strips of leaflet twice the length of the basket. The two midrib strips were laid together and the closed leaflets were twisted over each other to form a firmer rim (70, p. 183, fig. 164). The closed leaflets were plaited in twill as a continuous close sheet and the ends were closed as in the tapora basket but followed the twill technique instead of the check. The bottom was closed in two stages with the braid keel formed on the outside of the basket (fig. 21, e). The double thickness of the closed leaflets made this basket strong, and it was used until it wore out. In Mangaia three sizes of this basket were made, and each was given a distinctive name. The large size (kete rore) was used to convey quantities of cooked foods to feasts such as the takurua. When used to store cooked fish it was termed vai'ata and was hung up on a scaffolding (pa'ata) in the house. An intermediate size (kete takoto) was also used for cooked food (manga maoa). During community undertakings, women brought cooked food in such baskets, and the food was shared (tu'a) among the laborers. Because of the close plaiting, they were also used to hold small fry (ika tauira) of certain fish that came in in shoals. A smaller size (kete 'apua) was a fisherman's basket (kete tautai) which was tied around his waist when he fished in the lagoon or on the reef.
Pandanus-leaf baskets (kete rau) were made with a check plait. The bottom was closed with a neat three-ply braid ('iri), either outside or inside, or with the plaited finish (taviri) used on the edge of floor mats. The baskets were plaited with the dull side of the leaf on the outside. When the bottom was closed on the outside, the basket was turned outside in so that the braid or plaited finish which forms a kind of keel was concealed. Now, the bottom is sometimes stitched together with a sewing machine (70, p. 195, fig. 175). The rim was finished in one of a number of ways: with a three-ply braid ('iri), a straight plaited finish (taviri), or a serrated edge (patara).
Decorative patterns were made by using dyed wefts overlaid on the foundation wefts. Alternating colored and plain wefts in the ordinary check technique page 53were used, but greater elaboration was obtained by changing the check plait into varying combinations of twill usually in horizontal bands or in triangles, squares, or rhomboids arranged on a background of plain check. The full details of the pandanus technique have been described (70, pp. 190-202).
Pandanus-leaf baskets are now used to store clothing, trinkets, and articles introduced after European contact. Many of the colored patterns on the baskets are similar to those used on the sleeping mats. It is probable that European contact was the stimulus to the making of such baskets for home use, gifts, and sale.
Mats (general name, 'ariki), were made of coconut or pandanus leaves.
Coconut-leaf mats (tapakau) were made formerly in all the islands except Mangaia. I saw none in use and unfortunately did not have a sampler made. The tapakau mats of Manihiki (75, p. 113) and Tuamotu differed in details of commencement but were plaited in twill, and it is probable that the Cook Islands tapakau were plaited with a similar technique. They were made for floor coverings, and in sleeping houses they were laid over a covering of grass bedding. Finer sleeping mats of pandanus leaf were laid over them.
Pandanus-leaf sleeping mats were usually termed moenga (moe, to sleep). They were made from a cultivated pandanus with smooth-edged leaves. A number of green leaves were plaited by their tip ends (kauru) into a braid ('iri) and hung up to dry. When dry, the individual leaves were cut from the braid, opened by winding around the left hand, and scraped with a ka'i shell to flatten out the shriveled edges. The leaves were rolled around the finger for a start and other leaves added to make a fair-sized roll (tupe rau 'ara), which was kept in position by a strip of bark or pandanus leaf passed through the central hole and tied around the circumference. These rolls were stored until needed and sometimes were used as objects of trade in stores or sent to such places as Rarotonga where good pandanus is scarce.
In preparation for plaiting, the woman untied the roll and scraped the leaves with a sea shell (ka'i) or a piece of coconut shell (anga ipu). With a sharp point such as the skin spine of a porcupine fish (ivi totara) (now with a point of a safety pin), the woman separated the side edges of the leaf and the midrib, discarding them. Each half leaf was divided into about four wefts with the same sharp point, but a short length of the butt end of each half leaf was left undivided to keep the wefts together and so make the plaiting commencement easier.
Using the individual butt ends, there were two forms of commencement. In the single butt commencement ('atu ta'i), all the butt ends were laid obliquely in the same direction with the wefts directed toward the right (dextrals) so that some of the wefts must be turned toward the left (sinistrals). In the page 54double butt commencement ('atu rua), the butt ends were crossed alternately so that their wefts crossed without any being twisted to the left. Only the double butt commencement was used in Samoa (73, p. 217).
As plaiting proceeded, the left side edge was formed by bending in the lowest sinistral weft with a double turn to bring the weft back into the body as a dextral weft. This was done successively as the sinistrals reached the point above the turned weft below it and so a straight turned edge termed piu was formed at right angles to the bottom commencement edge. Similarly when the required width of the mat was reached, the right side edge was formed by turning back the dextrals to act as sinistrals and so form a right piu edge.
A joining ('ono) was made when the wefts began to narrow toward their tips. The butt ends of a second set of wefts were laid over the tip ends of the first set, and the plaiting was continued with double wefts for a distance sufficient to insure the security of the joining. The free ends of the first set were then dropped, and plaiting proceeded with the second set alone. Other joinings were made in a similar way to procure the length desired. Later the free ends of the wefts were trimmed off at the edge of the first crossing wefts.
When the full length of the mat was reached, the top edge was formed by a special finish termed taviri by which the free ends were turned down to be incorporated in the plaiting. The mat was then reversed, the butt ends split into individual wefts, and the taviri finish applied to them. Full details of the butt commencements, piu edge, 'ono joining, and taviri finish are described elsewhere (70, pp. 108-121).
Plain mats may be made with the above technique with piu sides and taviri ends. These mats were made in check throughout, each weft crossing over and under one alternately.
The good sleeping mats (moenga) are characterized by decorative borders in color (termed pae). The pae was formed by overlaid plaiting in which colored strips were laid on the plain sinistral wefts of the border as they were placed in the shed formed by the separation of the dextral wefts in the various arrangements demanded by the twilled technique of the pattern used.
The material for the colored wefts was the shiny upper layer (aro) of pandanus leaf separated from the dull under part (tua). The leaves were drawn through the fire so they would separate easily and then split from the tip end (kauru) by hand. The separated under part was discarded, and the tip ends of the upper layer, now termed papa, were plaited in a three-ply braid to keep the strip together. The strips were soaked for three to five days in the sea, the connecting braid ('iri) being tied to a stick driven into the sandy bottom. The strips were washed in fresh water and the braid hung up to dry. When dry, each leaf was cut from the braid and folded into convenient lengths of about six inches. In the modern method, the strips are soaked in a solution of coral lime (ngaika) for about three days to help them take the dye.page 55
The colors now used are black and red, but many informants maintain that originally only black was used. To obtain the red color, the strips were soaked in nono juice until they were the right shade. Then they were heated in an earth oven to fix the dye. The color was a dull red, and the method was abandoned for the brighter trade dyes.
Black dye was obtained from the scraped inner bark of the candlenut (tuitui) which was pounded in a bowl and the juice expressed by wringing through a strainer of coconut stipule (kaka). The papa strips were soaked in the dye solution for three days and the color fixed in an earth oven. The strips were then immersed in the black mud of taro swamps for four to five days, washed, and dried. The color soaked through the strip and shows up on both sides.
Black wefts were also obtained from the shiny black outer skin of the plantain ('uatu, Rarotonga; ve'i, Aitutaki). Appropriate strips were peeled from near the base of the trunk and the spongy under material removed with a shell.
The decorated mats were made in two forms, the koviri and paretumu.
The koviri mat was made with turned side edges (piu). When the required length was reached, all the wefts were split into two, to make finer plaiting. The border was then plaited according to the pattern desired, and the colored strips, split to the same width as the mat wefts, were laid in turn on the sinistrals as they were laid in the shed provided for them. In addition to the actual pattern formed by changing the number of wefts in the twill strokes, variety was added by using a number of wefts overlaid with black and then changing to a similar number of reds. The colors thus formed alternating oblique panels of black and red. When the depth of the decorative panel was reached, the edge was formed with the taviri finish. The mat was then reversed, the butt ends which were sufficiently long for the purpose, were split into wefts corresponding to the mat wefts and split again to form border wefts. The technique of the other end was repeated. The koviri mat therefore had plain sides and decorated ends. (See plate 5, C.)
The paretumu mat has decorated borders on all four sides. As the piu turn at the sides could not be used, the left side edge was commenced, the sinistral wefts were allowed to project free beyond it, and fresh dextrals with long butt strips were added successively on the left edge to engage the sinistrals as the plaiting proceeded upward. Similarly, the dextrals on the right side edge were left free to project beyond and new sinistral strips were added as the plaiting proceeded. On the completion of the two end borders, the butt strips of the two sides were treated with the same technique as that of the commencement edge in the koviri mats. (See plate 5, B.)
The color does not show on the under surface of the mat, because the page 56colored strips were laid on the foundation wefts, but the wefts show the geometrical pattern used. For details of the plaiting technique of the mats and the various motifs and patterns used, see my work on Aitutaki (70, pp. 106-159).
Overlaid plaiting in color is used in other islands, but the splitting of the body wefts to form narrower border wefts, their use to form decorative borders in color in a number of geometrical patterns, and the manner in which new wefts are added to the side edges are local developments not found in any other Polynesian group.
Fans (ta'iri) of two types were made from coconut leaflets, a rough fan for ordinary use and a finer one of dressed leaflets.
The rough fan was made of a piece of coconut leaf cut off near the tip end and long enough to form both handle and fan. The butt end was stripped to form the handle and the remaining part, with seven to ten leaflets on each side, provided the body of the fan. Opposing pairs of closed leaflets were crossed to the opposite side, and the plaiting commenced by turning the lowest leaflet on one side upward to lie parallel with the midrib and so cross all the leaflets on the same side in check. The next leaflet was turned up in a similar way until all on the same side had been plaited. The same was done on the other side. When completed, the free dextrals of one side crossed the free sinistrals of the other in the middle of the fan. These free wefts were plaited in check to fill in the middle part between the two oblique sides. When the plaiting had formed an even front transverse edge, the edge was finished by doubling the leaflets back and pushing them under crossing wefts on the body of the fan. The ends were then cut off. The full technique of the plaiting has been described (70, pp. 202-205). (See figure 262.)
Finer fans were made of closed coconut leaflets stripped from young leaves, with the edges trimmed to form narrower wefts about 6 mm. wide. In the best fans the handle was carved on the proximal part and the distal part served as a tang to which the wefts were attached and plaited. (See figure 22.)
Splitting the leaflets into hard and soft wefts gives the fans a ridged appearance. The downward prolongation of the midrib wefts before they were turned upward results in an ornamental effect peculiar to Aitutaki (pl. 4, A, B). The resultant shape of the fans is triangular with the apex on the handle and the base distal. In finishing off the base edge, all the wefts are bent at right angles to their previous course, pushed down under a few crossing wefts and the ends cut off. In bending down the hard midrib wefts, small projections are formed that give the base edge a serrated appearance. The page 57fans described are still made in Aitutaki but the handles seen were plain branches without carving. The lowest leaflet butts were covered by a series of close transverse turns of the binding thread.
Figure 22.—Technique of fine fan, Aitutaki. a, piece of thin cord (1) is tied to handle (2) some distance from distal end; two leaflets (3, 3) are placed with butt ends on either side of handle, and the cord makes turn (4) around them. Another pair of leaflets (5, 5) is placed in position below first pair and fixed with another turn (6) of the cord. In this way, successive pairs are added and fixed with spiral turns of the cord until full number of 16 to 24 pairs of leaflets are added. After last pair is added, cord makes several transverse turns close together to cover butt ends. b, two lowest leaflets (1, 2) are crossed obliquely over what is to be back of handle. c, the two leaflets (1, 2) are carried around to front of handle and again crossed in check. d, turning to the back, next pair of leaflets (3, 4) are treated in same way back and front and so on in succession until all pairs have been dealt with. Handle is thus covered by a check plait up to first attached leaflets. e, front of handle showing oblique crossings with leaflet midribs all on distal side. Each leaflet about 6 mm. wide is split into two wefts, the proximal consisting of the two thicknesses of the soft leaf material and the distal of hard stiff midrib. Commencing on the left with lowest leaflet pair (1, 1′), the near soft weft (1′) is turned up parallel with handle and plaited through wefts above it in check. f, next weft, which is stiff midrib part (1), is bent down a short distance and then turned upward parallel to its soft neighbor (1′) to engage other wefts in check. For next movement, the soft weft (3′) will be turned upward and then the hard weft (3). From here on plaiting is exactly the same as that in rough fans (70, pp. 202-205). g, cut off corner: the hard weft (1), bent in as a dextral, forms a boundary which cuts off the sharp corner usual in Rarotongan fans; other sinistrals on reaching it are doubled back into plaiting. Hard weft (1) is then bent at right angles (2) to form a point, soft weft (3) crosses over it to form interspace with next hard weft (4), which also forms a point, and so with next soft weft (5) and next hard weft (6). Thus, hard wefts form the sawlike basal edge of fan.
Figure 23.—Rarotongan fan handles (a, plain, attributed to Rarotonga but probably from Aitutaki; b-e, carved with two heads in profile, looking outward from midline of junction): a, round wood, uncarved, 28 mm. in diameter at proximal end, and uncovered part, with shoulder (1) where sennit braid wrapping in lozenge pattern commences; total length of handle including tang, 390 mm. (Cambridge Univ. Mus., Z.6100). b, 1, double head with narrow notch to mark division, eyes formed by five incised curves and mouth by four curves; 2, raised curved flange with square hole in depressed interval between it and heads; 3, raised area pierced by large hole forming segment of circle, and with round hole in interval between it and flange 2; handle beyond for 39 mm. seized with fairly fine two-ply coir cord arranged alternately in two turns of a cord dyed red and an undyed cord; length of carved part 76 mm.; greatest width of heads, 38 mm., and thickness, 19 mm. (British Mus., 6851). c, 1, double head with curved hole marking cranial division; eyes formed by wider incisions than b, and mouth formed by single notch on one side and double notches on other; 2, curved flange closer to heads and notched on either side to form zigzag bar; mesial hole between 2 and 1; 3, raised area with proximal side notched, and large square hole with proximal side also notched; round hole between 3 and 2; 4, extra raised flange with middle part notched to form a zigzag bar, carved part of handle 70 mm. long and greatest width 26 mm.; handle beyond last flange (4) wrapped with fine braid 2 mm. wide, for distance of 40 mm. on diameter of 17 mm.; braid in two sets, one dyed dark red (stippled), other undyed; wrapping arranged in lozenge pattern so that two sides of lozenges in different colors; total length of handle including tang, 342 mm. (British Mus., L.M.S. 218). d, 1, double head divided vertically by narrow groove; eyes and mouth roughly formed by incised curves; 2, curved flange without ornamentation, and round hole between it and heads; 3, plain raised part with large hole, and also hole between it and flange 2; 4, flange with distal edge notched; 5, additional smaller flange with vertical incisions on distal edge; length of carved part 85 mm.; greatest width of heads, 45 mm. due to greater outward inclination of chins; handle beyond carving wrapped with red and yellow sennit in lozenge pattern (Cambridge Univ. Mus., Z.6101). e, 1, double head with eyes and mouth formed by curved incised lines and separated by single vertical line; 2, series of four bilateral angular projections with triangular depressions to the inner side pierced at bottom with round holes; 4, distal raised flange with zigzag raised bar; length of carved part 89 mm.; greatest width of heads, 41 mm.; width of distal flange (4) 32 mm.; handle beyond carving wrapped in red and yellow sennit in lozenge pattern (Cambridge Univ. Mus., Z.6102).
Rarotongan fans (pl. 4, C) are readily identified by the handles which are carved with two heads, back to back, and the eye and mouth forms peculiar to the carving on that island.
Specimens are to be seen in the British Museum (6851; L.M.S. 218), Cambridge University Museum (Z. 6101; Z. 6102), American Museum of Natural History (S. 5125), and the Peabody Museum, Salem (2144; Z. 21214). All have the double head pattern, but carving details adjacent to the heads vary, as shown in figure 23.
Between the carved end and the commencement of the plaiting, the handle was wrapped with fine sennit. The part near the plaiting was for binding the ends of the lowest leaf wefts, but the lashing turns were continued on the handle for decorative purposes. On one fan (fig. 23, b), a fine coir cord was used in transverse turns, but a decorative effect was produced by a cord dyed red which made alternate turns with a cord of the undyed material. A more decorative effect, however, was produced by working a pattern of multiple lozenges, as shown in figure 23, c. The part of the handle so worked is 40 mm. long and 17 mm. in diameter.
Figure 24.—Cook Islands fan (British Mus., L.M.S.). a, outline; length of plaited part, 395 mm.; base width, 430 mm.; length of external handle, 130 mm. b, front, showing short butt knobs (1, 1), raised flange (2) divided in middle line, rounded body (3) with large perforation, bilateral curved flanges with lower end (4) incised with chevrons and upper end (5) notched, terminal flange (6) with perforation below it, and plain lashing (7) of fine cord. c, side view with same numbering as b; upper end (5) of curved flange has a geometrical pattern resembling carving of Atiu and Mauke.
Figure 25.—Mangaian fan technique: a, the two-ply lashing thread, attached (1) near distal end of handle tang (2), passes down on front of tang and makes two transverse turns (3) around butt ends of first pair of leaflet wefts (4, 4) which are placed against sides of tang. Thread descends on front and makes two transverse turns (5) around butt ends of next pair of wefts (6, 6) and continues until all weft pairs are attached. Leaflet wefts are about 0.37 inch wide and are attached to tang about 0.5 inch between pairs. b, commencing at bottom, lowest pair of wefts (4, 4) are twisted over pair above (6, 6) on their respective sides and lashing cord (1) ascends on front of tang to point midway between butt lashings (7, 8) and makes two transverse turns (3) around tang to further fix lowest pair (4, 4). Next pair (6, 6) is then twisted over pair above and lashing thread, again ascending on front of tang to midway between butt lashings (8, 9), makes two transverse turns (5) around tang to further fix pair (6, 6). Second turn is placed below first and crosses it on left as it ascends to next position. Twisting and lashing is done consecutively with each pair in turn. c, handle is turned, and all wefts from below are crossed in simple check. Wefts have leaflet midrib on upper edge. Commencing to plait on one side, lowest weft (1) is given a half-twist to place midrib to outer edge and plaited in check through wefts above it so that it lies parallel with tang (2). Next weft (3) is treated in a similar way and so consecutively until required width of one side is reached. Portion plaited forms lower half of lozenge. Turning of last weft forms outer angle of lozenge-shaped fan, and weft itself forms outer edge of upper half of lozenge, and crossing wefts are doubled over it and ends passed under two or three crossing wefts on opposite side. Though first wefts run parallel with tang in first part of course, they incline inward in later part. Other side of fan is plaited and end portion is completed by wefts from either side crossing each other. d, edges of fan are bound with strips of coconut leaflets (1) doubled over them and fixed with a two-ply fine coir cord which passes around binding and through wefts in a continuous series of half-hitches or overhand knots (2). In Mangaian fan, a characteristic sennit pattern (3) in zigzag form with three courses was made with the binding near handle. Back of tang (4) is shown divided into panels by transverse narrow grooves through which lashing thread passes so that it is not visible on surface as in the Wesleyan University fan. Each panel is carved with two K-motifs with their arms toward middle line. Upper part of handle is wrapped in white tapa (5) and tied with thin coir cord. e, on front of Horniman Museum fan, four strands of fine sennit braid come through in the middle line from a point opposite distal end of tang and descend to disappear under bark-cloth wrapping of upper half of handle. A single strand of fine human hair braid (2) is worked back and forth across four sennit braids in check. On back, the four braids descend in a pair on each side of tang and also disappear under cloth covering of handle.
The Rarotongan fans differ slightly from the modern Aitutaki fans in that the midrib wefts do not push down into angular projections near the handle and the base is smoother than the saw edge characteristic of Aitutaki. The Rarotongan fans also have the basal angles brought to a point instead of cut off like the Aitutaki ones.
A fan in the British Museum is of the Rarotongan shape and technique, but the handle is carved differently and the three edges are reinforced with a continuous chain knot in a fine cord. The fan is probably from Atiu or Mauke (pl. 4, D; fig. 24).
Fans of the better type are no longer made in Mangaia, but a specimen in the Horniman Museum (3.280), London, served as a key to an interesting problem. The fan differed from those of Rarotonga and Aitutaki in shape and technique, but its locality was proved beyond doubt by the characteristic K-motif carved on the wooden handle (pl. 5, A). Though this fan was well made, the technique was so different that I wondered whether it was a freak that did not truly conform to the orthodox technique of Mangaia. Fortunately, in 1939 I found a similar fan in the Wesleyan University Museum (372), Middletown, Connecticut which, though uncarved, was made with the same technique (pl. 5, D, E). The fan was old and formed part of a missionary collection that was acquired by the University in 1870. Through the courtesy page 62of the museum authorities, the fan was lent to me for study. The details of technique are shown in figure 25. A similar fan was located later in the Peabody Museum (E.5355), Salem, Massachusetts (labeled Feejee), and another was figured by Edge-Partington (24, II-19-1) from St. Augustine's College Museum, Canterbury, England (attributed to Tahiti). The carved handle of the Horniman Museum fan led to the identification of the three fans without carving; and the four fans establish the technique as orthodox for Mangaia. The St. Augustine's fan is the largest of the four, being 32.25 inches long (including short handle) and 20.5 inches at the widest part.
The Mangaian fan, according to Gill, symbolized peace. He writes (28, p. 301):
Men daily carried about with them, in symbol of peace, an outrageously large fan, now obsolete. This fan was sufficiently large to protect the upper part of the body from sun or rain. It was found necessary to forbid its use in church, as the person of the owner was nearly hidden behind it.
The missionary ban against its use in church probably caused the large fan to be abandoned and the technique to be forgotten.
1. They are lozenge-shaped instead of triangular and have two angular notches at the distal pointed end. 2. The wefts, after attachment to the handle tang, are twisted over the wefts above them commencing from below (handle). This technique is observed in the coconut-leaf fans of Manihiki (75, p. 132). 3. The wefts are crossed over the front of the tang but not over the back. The front of the tang and the spiral turns of the lashing thread are completely covered by the weft crossings but the back of the tang and the transverse crossings of the lashing thread are exposed throughout its whole length. In the fans from the other islands, the tang is completely covered, except for the short distal end that projects on the back. 4. The wefts after crossing on the front are plaited in check without the midrib part being split from the soft leafy part. Hence, the Mangaian fans do not have the ribbed appearance of the other fans. 5. All the edges of the fan are bound with strips of coconut leaflets attached by a continuous thread.
I have seen no good fans from Atiu, but the problem as to whether they followed the triangular type of Aitutaki and Rarotonga or the lozenge-shaped form of Mangaia is settled by the following remarks from Cook (20, p. 187), 1777: "After walking a little way amongst them, we found a person who seemed a Chief, sitting on the ground cross-legged, cooling himself with a sort of triangular fan, made from leaf of the cocoa-palm, with a polished handle, of black wood, fixed to one corner."
It is evident that Mangaia was the only island in the Cook group that had the lozenge-shaped fan with its peculiar technique. The other islands had the triangular form and identical technique. For distribution and comparison of fans, see pages 426-428.
Temporary eyeshades (taumata), which were used mostly by fishermen, were quickly made from a strip of coconut leaf with a technique similar to that of Manihiki and Rakahanga (75, pp. 124-125) (fig. 26). The permanent eyeshade with a wooden framework and cover of fine sennit netting, once made in Tahiti, was not known in the Cook Islands.
Figure 26.—Rakahanga eye shade (Bishop Mus., C3071): a, strip of coconut-leaf midrib bearing seven leaflets (1-7) with commencing check plait; lower leaflets (4, 5) have set width of band. b, band continued by bending in wefts 4, 5 and succeeding pairs as they reach edges; weft 6 marks length of band. c, weft 6 is doubled back to change direction and other wefts treated as shown on left; weft 6 again marks end of plaiting and other six wefts plaited in three-ply braid (8); an extra leaflet (9) is tied to first weft near midrib and used with braid tail (8) to tie at back of head. (After Te Rangi Hiroa, 75, fig. 41.)