THE LOCAL EVOLUTION OF HAWAIIAN FEATHER CAPES AND CLOAKS. BY TE RANGI HIROA (PETER H. BUCK). INTRODUCTION.

(Volume 53 1944 > Volume 53, No. 1 > The local evolution of Hawaiian feather capes and cloaks, by Te Rangi Hiroa, p 1 16 ) THE LOCAL EVOLUTION OF HAWAIIAN FEATHER CAPES AND CLOAKS. BY TE RANGI HIROA (PETER H. BUCK). INTRODUCTION. THE feather capes and cloaks of Hawaii are beautiful products of native craftsmanship. The capes best known are circular in form, and the bright red and yellow feathers of which they are principally made are so closely attached to the foundation material that the overlapping rows of feathers form a smooth surface that recembles velvet or plush. The circular shape is so highly specialized for a Polynesian community that the possibility of diffusion from some non-polynesian source has been suggested. The unique shape of the Hawaiian feather-helmet with its raised median crest, has suggested Spanish influence in spite of the fact that the theory of Spanish discovery before Cook has been exploded. The circular type of cape is the end-result of a series of inventions that progressed through a number of simpler techniques. For New Zealand garments, I was able by an analysis of weaving-technique to establish their local evolution from the simple rain-cape to the more elaborate types of dress cloaks. In Hawaii, the importance of rectangular capes of coarser and simpler technique has been overlooked probably because of their rarity. This article is an attempt to show the significance of the rectangular capes and the ti-leaf rain-cape in the chain of local evolution in technique that led to the more complex circular capes and cloaks. CIRCULAR FEATHER-CAPES. The circular capes are the most highly developed type as regards shape, technique, and colour designs. As indicated by the number preserved, they evidently became the fashionable form, and were supplanting earlier types. The capes have a concave neck-border, a longer convex lower border, and two straight side-edges which meet in front to complete the circular shape of the capes. They are deeper in the mid-line at the back than in front, the general back-depth being about 12 inches though it may reach the exceptional depth of 28 inches. A good deal of variation occurs in the length of the neck-border and the depth of its concave curve (pl. 1). The foundation to which the feathers are attached is netting made with same technique as fish-nets, but with a very fine mesh (pl. 2). The twine used is two-ply cord made from the inner bark of the olona plant (Touchardia latifolia) of which the fibre has been proved to be stronger than manila hemp. No technique was developed to make the meshes on a curve to correspond with the shape of the cape and therefore netting made in the usual technique of straight rows, had to be cut to the required shape. The entire cape was but rarely cut out of one piece of netting. In some capes, a median gore with two sidepieces was sufficient (pl. 2) but in others, pieces of different shapes and sizes were joined together like a patchwork quilt to obtain the shape desired. The edges of adjoining pieces were laid together and a separate cord was run alternately through the marginal meshes on either side to join the pieces together. The feathers most commonly used were red from the 'i'iwi (Vestiaria coccinea) and apapane (Himatione sanguinea) and yellow from the 'o'o (Moho nobilis) and mamo (Drepanis pacifica). Black feathers from the 'o'o were also used, and more rarely green from the 'o'u (Psithacirostra psittacea). The feathers from these birds were fairly small, and before they were attached to the netting-foundation, they were carefully picked over and arranged in bunches according to size. One bunch examined had as many as eighteen small red feathers. In making up the yellow feathers, some shorter red feathers were placed over their quills to form what was termed a pa'u (skirt). The quills of each bunch were tied together with

a strip of olona bast (fig. 1, a). Women were allowed to sort the bunches but the making of the garments was restricted to men. The exact details of tying on the feathers has not been satisfactorily described. Brigham's work (2-4) lacks detail, and the information cannot now be obtained from the Hawaiians. J. F. G. Stokes, when Curator of Polynesian Ethnology at the Bishop Museum, did work out the methods of tying but unfortunately his work was never published. I worked out the tying-technique independently by examining a cave fragment under the low-power microscope. Once the fundamental technique became clear, it was possible to detect variations in the Museum capes and cloaks without damaging them. The feathers were tied to the netting in overlapping rows that commenced at the lower border with this part of the netting toward the worker. This commencement-method with the free or tip-ends of the feather-bunches directed toward the lower edge, left the meshes above or beyond clear for successive rows. The feather-bunches overlapped laterally and the rows were spaced close enough for the free ends to overlap the row below it and thus completely cover the surface of the netting-foundation. The tying-thread was a two-ply twist of olona fibre. In fourteen garments examined, the thread was continuous in passing from left to right to successive bunches until it ran out when a fresh length was used. In some capes, the direction of the thread changed evidently to fill in some special part. The knot used throughout was an overhand knot, the thread passing around the feather quills and the parts of the net-meshes upon which the feather-bunch was laid. The first knot was always tied around the near part of the quill where it joined the barbs. In some capes, one knot with an additional half hitch (fig. 1, b) was considered sufficient, but in other capes an additional knot (fig. 1, c, d) or two knots (fig. 1, h and j) were made around the far part of the quill before the tying-thread passed on obliquely to the next bunch. Other variations consisted of the thread returning to the first knot after two knots (fig. 1, f) or three knots (fig. 1, g and i) to be tied over it before passing horizontally to the next bunch. Owing to the curve of the feather-rows, the part of the meshes to which the bunches were tied varied as shown in fig. 1, b-j. The fineness of the work may be appreciated from the fact that the meshes are from 5 to 6 mm. in depth and width, and the feather bunches are one mesh apart between bunches and between rows. The featherquills are from 6 to 7 mm. long and three spaced overhand knots may be tied over them as well as a fourth overlapping knot. - 3 The colour-designs are mainly in red and yellow, one colour forming the background, and the other the geometrical motifs. In the capes, red slightly exceeds yellow in forming the background, but in all the red background capes yellow forms the lower-border of varying depth. Black and green appear less often in the motifs. The motifs are worked on the body and along the front and the neck-borders. The body-motifs are complete geometrical figures such as crescents, triangles, and lozenges, which may be repeated. Curved bands and zigzag bands occur less frequently, and small circles and rectangles are rare. The front-motifs are usually half-crescents or triangles arranged in pairs on either side, and these combine to form full crescents or lozenges when the front-edges are brought together. Usually a narrow vertical band, spaced into small rectangles of red and yellow, and sometimes black, separates the front-motifs from the frontedges. The neck-border is treated similarly to the front-edges. The circular capes may be divided into two groups according to whether the feathers used are fine or coarse.

a. The fine circular capes are covered entirely with fine red and yellow feathers with the occasional addition of black and green feathers, which are also small (pl. 1 and 2). b. The coarse circular capes have the neck- and front-borders covered with fine red and yellow feathers but the rest of the surface is covered with coarser feathers from the domestic fowl, manof-war hawk, or tropic bird. The netting-foundation is finely meshed to allow the use of fine feathers along the neck- and front-borders. In records of eighty-nine capes, seven had the body covered with coarse feathers (pl. 3). A good deal of variation is present as regards details of shape. Some have a long straight neck-border and seem to form a transitional stage in shape between the advanced circular type and the older rectangular type. Some occur with fine feathers (pl. 4), and others with a main covering of coarse feathers (pl. 5). CIRCULAR CLOAKS. The circular cloaks are an end-product of the technique of circular capes and no rectangular cloaks have been recorded. The cloaks vary considerably in size and shape, some being short and wide and others deep and narrow. Of the ten cloaks in the Bishop Museum, the longest is 66 inches and the widest 168 inches. The background is usually red, but a few are yellow. The yellow is supplied principally by the 'o'o, but mamo feathers occur to a limited extent in some cloaks. A remarkable exception is the famous Kamehameha cloak in the Bishop Museum, which is composed entirely of mamo feathers. Motifs, similar to those in the capes, were worked on the cloaks, but owing to the larger surface, the motifs were usually larger and more in number, not only on the body but along the front-edges. A common motif was the large crescent with a median upper point (pl. 6). In a few cloaks, coarse feathers of the tropic bird or domestic fowl have been used on the body much in the same way as in the coarse circular capes. Of forty-eight cloaks with the details known to me, one was decorated with tropic bird's breast-feathers and two with cock's feathers (pl. 7). RECTANGULAR FEATHER-CAPES. The rectangular type of cape is unknown to the Hawaiians of to-day not only because of their rarity but also probably because none - 4 have been retained in the Territory of Hawaii. The Bishop Museum photograph-catalogue has records of five in a good state of preservation distributed in museums as follows: two in Vienna, and one each in Petrograd, British Museum, and Norwich Castle Museum. Most of them were collected during Cook's Third Expedition. The number, though small, is sufficient to establish the existence of a distinct rectangular type of cape made with a simpler technique than that of the circular type. Of the two Vienna capes, one is incomplete as regards the finish of the neck-band but the other is perfect and it is selected as the type specimen for description. The Vienna type-specimen (Museum number 180) was formed of one piece of netting with a coarse mesh. Though the term rectangular is used to distinguish these capes from the circular type, the lowerborder is always longer than the upper. In the type specimen, the upper-border is 32.25 inches in width, the lower border 42.5 inches, and the middle depth is 15.75 inches. The netting is attached along the upper-border to a neck-cord of olona in square braid which has free ends for tying around the neck. The side-edges of the netting are also attached to olona braid which hang free at the lower edge of the netting and end in loops (pl. 8). The coarse netting-foundation (pl. 9) forms a marked contrast to the fine meshed netting used in the circular capes (pl. 2).

The feathers used are black from cock's plumes and the man-of-war bird (Aquila fregata), brown from the neck-feathers of roosters, and white from the body-feathers of the white tropic-bird (Phaethon lepturus). Black body feathers from the domestic fowl were also used, and there is some doubt as to whether some of the white feathers were also from the domestic fowl. These feathers are larger and coarser than those used on the circular capes, and they are attached to the coarse netting throughout except for a specialized narrow band along the neck-border. The method of tying the feathers to the netting has been described by Ling Roth (5, p. 45) for the Norwich Castle cape, which is 29.5 inches wide at the upper-border, 41 inches at the lower-border, and 14 inches in depth. The feathers are tied in overlapping rows from below upward to each row of meshknots. From Ling Roth's drawing reproduced in figure 2 a, it is evident that a single feather was laid with its quill over the mesh-knot and the tip-end directed downward. The tying-thread was passed around the quill and the two arms of the mesh below the mesh-knot and knotted. The thread then passed above the mesh-knot, the end of the quill was doubled down, and a second knot was tied around the doubled quill and the two arms of the mesh above the mesh-knot. The thread was cut, there being a separate piece of thread for each feather. Ling Roth does not state what kind of knot was used and I did not have time to study the details when I saw the Vienna cape. However, judging from the technique of the circular capes, the knot was most probably an overhand knot (fig. 2, b, c). The effect of tying below and above the mesh-knots gave the meshes a hexagonal appearance as shown in figure 2, a. A photograph of the back of the Vienna cape (pl. 9) shows that the meshes have a hexagonal appearance in the part to which feathers are tied, and there can be little doubt that the feathers in the Vienna and Norwich Castle capes were tied with a similar technique. This technique probably indicates the common technique for the rectangular type of cape. The feathers were arranged to form a simple colour-design: a horizontal band in black along the lower border, vertical panels or - 5 rectangles in black, brown, and sometimes white along the side edges, and the rest of the body usually filled in in white. The Norwich Castle cape differed in having two halfcrescents in black at each side-edge with their bases outwards so that when the side-edges met the halfcrescents also met to form full crescents after the style of the circular capes. A few inches from the upper-border of the foundation-netting, the attachment of feathers ceased with a horizontal black band, leaving an upper strip of netting free of feathers. In the Vienna type cape, the upper strip of netting was covered by a strip of tapa sewn to the neck-cord above and to the side cords at the side edges. To the front-surface of the tapa, pieces of bird-skin carrying their natural red and yellow feathers were gummed down in four horizontal rows; the upper row of red, the second of yellow, and the two following of red. Thus the tapa was completely covered with feathers, and as its lower free-edge overlapped the upper-edge of the horizontal band of black feathers attached to the netting, the whole front-surface showed a continuous covering of feathers. In the Petrograd cape, a band of fine plaiting took the place of tapa and though the pieces of bird-skin remained gummed to it, the feathers have been completely destroyed by insects. The Norwich Castle cape also has a plaited band in place of tapa, but both feathers and skin have disappeared leaving the plaiting bare. The British Museum rectangular cape has a white body with lower and upper bands of black and side-panels of brown. A neck-band covered with red and yellow feathers is present, but I have no information as to details. When I first saw the Vienna cape, I thought it was a kilt for wearing around the waist, but such a use was contradicated by the fact that the Hawaiian men did not wear kilts and women were not allowed to wear feather garments. Subsequent search of the literature revealed that Brigham (4, pp. 25, 26) had referred to these rectangular garments as war-capes. More convincing is a picture by Webber of a scene

in Waimea, Kauai, in which a man is shown wearing a rectangular cape with a white background (6, atlas, pl. 35). Brigham in referring to the Petrograd cape (4, p. 9) describes it as, One of the capes worn over the shoulder for convenience in battle for wielding club or hurling spear. It is not of a kind to mark chiefly rank. As the Hawaiians are credited with wearing the circular capes and cloaks in battle, the term war-cape is not distinctive for the rectangular type. Brigham's statement that the rectangular cape did not mark chiefly rank must be modified to mean that they did not mark the highest rank; for in the account of Cook's third voyage (6, pp. 136, 137), the following statement occurs after a description of the cloaks decorated with red and yellow feathers: The inferior chiefs have also a short cloak, resembling the former, made of the long tail-feathers of the cock, the tropic, and man-of-war birds, with a broad border of small red and yellow feathers, and a collar of the same. Others again are made of feathers entirely white with variegated borders. The first part of the description evidently refers to coarse circular capes (pl. 3). As the second Vienna cape and the rectangular capes in Petrograd, British Museum, and Norwich Castle Museum are mostly white with variegated borders, the last part of the description evidently applies to the coarse rectangular capes. The statement is - 6 clear that the two types of capes with coarse feathers were worn byinferior chiefs. The rectangular capes also have a coarse and a fine variety. The fine rectangular cape has a foundation of fine-meshed netting for the attachment of finer feathers. A specimen in the Bishop Museum (9070), though small, illustrates the variety. The netting-foundation with a fine mesh of 6 mm., is in one piece with an upper width of 490 mm., a lower width of 630 mm., and a middle depth of 240 mm. It has a square neck-braid, 7 mm. thick, with one end free for a tying-cord, and the other end missing. The sideedges are also bounded by lengths of square braid. Unfortunately most of the feathers were destroyed by insects in the cave where the cape was hidden, but the quill parts are still held in position by a continuous thread of olona with two overhand-knots in the technique shown in figure 1 d. All the feathers used appear to have been red. Two other netting-foundations of rectangular form with a fine mesh are in the Florence Museum, Italy. RAIN-CAPES. Rain-capes made of ti leaves attached to a netting-foundation were roughly and quickly made. Their rough appearance evidently did not appeal to collectors and were it not for a single specimen in the Bishop Museum, their technique would have been lost. The Bishop Museum specimen (pl. 10, 11) is made of a rectangular piece of ordinary fishing-net with a 0.6 inch or 15 mm. mesh. The upper border has been damaged but the lower border is 52.5 inches wide and the depth in the middle line is 26.5 inches. Split leaves of the ti (Cordyline terminalis) are attached on one surface of the netting by passing the stalk of the half-leaf around two arms of a mesh below the mesh-knot and tying the free end around its standing part with an overhand-knot (fig. 3). A slip-knot was thus formed which could be tightened by pulling the leaf. The half-leaves are tied singly along the lower border below every second mesh-knot in the same row which spaced them about 1.2 inches apart. The leaves of the first row hung down below the lower edge of the netting for 15.5 inches. The next row is similarly tied along the mesh-knots two rows above, thus spacing the leaf rows 1.2 inches apart. Successive rows are attached until the entire surface of the netting is covered up to the neck-border,

forming a very effective thatch. A cord is passed through the neck marginal meshes, knotted to the meshes at each end, and the cord ends left free for tying around the neck. ORIGIN OF FIRST CAPE. The branch of the Polynesian people who settled in Hawaii, found the climatic conditions very similar to those existing in their previous home in central Polynesia. The tapa clothing to which they were accustomed proved equally suitable in their new home, and the paper mulberry plants (Broussonettia papyrifera) that they had brought with them thrived abundantly to supply them with the necessary raw material. Thus nature did not force the immigrants to devise a new form of clothing, and tapo-cloth continued to provide the Hawaiians with clothing until the coming of the white man displaced tapa with the woven textiles of foreign looms. But though the manufacture of tapa adhered to the old technique of soaking the paper-mulberry bast in water and felting the strips together with wooden beaters on wooden anvils, the artistic sense of the people found local expression in the discovery of new vegetable-dyes, in carving patterns on their beaters - 7 to impress different watermarks on the cloth, and in carving patterns on bamboo-strips to stamp various coloured designs on the completed cloth. The men wore a long narrow strip (malo) that passed between the legs and around the waist. The women wore a short skirt (pa'u) that extended from the waist to the knees. On occasions, both sexes draped a sheet of tapa around the shoulders as a mantle (kihei.) Thus so far as clothing is concerned, the needs of the people were fully met by bark-cloth, and some other initial stimulus must be sought for the origin of the feather-capes and cloaks. Though nature was beneficent, there was one particular item in her bounty that caused man to think. This item was rain. Though rain in Hawaii is now often referred to as liquid sunshine, the early Hawaiians probably found it as inconvenient at times as other peoples did. The Hawaiians had solved the problem of protection against rain by thatching the roofs of their houses in overlapping layers from the eaves upward. The technique reached back into pre-hawaiian history, but in Hawaii the thatchingmaterial was changed to pili grass, which grew locally. Permanent dwelling-houses were strongly made as regards the wooden framework, and the thatch of pili grass was carefully applied. In the upland cultivations, however, temporary shelters were made that did not require such careful work as the permanent habitations. The material for thatch most readily available was the leaves of the ti-plant which everywhere grew abundantly. Thus temporary shelters used by hunters and cultivators were thatched with ti-leaves. All people on land could thus take shelter under a thatch whether permanent or temporary. There was one class of people, however, who could not take shelter under a thatched roof, and they were the fishermen. The large voyaging-canoes were fitted with deck-houses for shelter from sun and rain, but the small fishing-canoes could not very well be furnished with a thatched roof without making them unwieldy. The problem that faced the fisherman was the provision of a thatch-protection that would be portable and not take up too much space in his small craft. The use of ti-leaves in overlapping layers was known, but how could such a technique be adapted to his particular requirements? In his occupation, the fisherman used nets which he made himself. The idea may have occurred to a fisherman that a piece of fishing net would provide a friable framework to which ti-leaves could be tied to form a portable thatch. All that was required was a piece of netting large enough to cover the shoulders in the sitting position assumed in canoes. The experiment was tried and proved a success for the ti-leaf rain-cape came into existence. I have given the credit of invention to fishermen because their need was greatest, and the netting-foundation was part of their equipment; but it is possible that the idea may have occured first to a fowler or an inland cultivator. TECHNICAL DEVELOPMENT.

A technical process having become established, it required no great mental effort for some craftsman to realize the decorative possibilities of substituting feathers for ti-leaf. In carrying out the new thought, a rectangular piece of netting with a coarse mesh followed the pattern of the ti-leaf rain-cape. In selecting the feather substitutes, fairly large feathers were necessary to cover the mesh spaces adequately. The domestic fowl was readily available, and black and white body-feathers were utilized. The black tailplumes and brown neck-feathers of roosters offered variety, and a further range of selection was provided by the white and the speckled body-feathers of tropic-birds - 8 and the greenish-black feathers of the frigate-bird. The feather-quills, however, were too short and stiff for tying to the meshes in the same way as the ti-leaves, and so a new attachment-technique had to be devised. This consisted of tying each individual feather over each mesh-knot with separate pieces of olona-thread in two overhand knots (fig. 2). The thatch-technique of the rain-cape in overlapping rows from below upward was followed both for effective coverage of the surface and for clearance in attaching successive rows of feathers. The neck-cord was improved by using a square eight-ply braid of olona-fibre. A further addition was made by attaching square cords to the side-edges of the netting in a similar manner to the neck-cord. Thus we have the evolution of the coarse rectangular feather-cape. It may be assumed that the early capes of this type were covered entirely with the larger feathers. A type of feather-cape having been established, the brighter red and yellow feathers of certain forestbirds directed attention to the possibilities of colour-improvement on the somewhat drab colours of black, white, and brown used on the coarse rectangular capes. A technical difficulty presented itself in that the feathers of the forest-birds were too small to cover the mesh-spaces of the coarse netting in use. The problem was met in two ways. The simpler method consisted of the partial decoration of the coarse rectangular capes by gumming pieces of bird-skin with the red or the yellow feathers in situ to a separate foundation-band of tapa or of closely-plaited material, and fastening the band to the neck-border of the cape as an addition. This method is illustrated by the rectangular cape in plates 8 and 9. Though the technique of using pieces of bird-skin may seem rather out of the way, it must be remembered that the 'i'iwi was skinned in the forest and there was no necessity for plucking the feathers since they could not be tied to the foundationmaterial. Thus we have an advance in technique to a coarse rectangular cape in white, black, and brown to which a band of red and yellow feathers was added by a separate process. The second solution to the problem of the more extensive use of the smaller bright feathers was the fairly obvious one of reducing the size of the mesh in the netting-foundation. This step not only allowed the small feathers to be attached by tying but also allowed the whole surface of the cape to be covered with the smaller and brighter feathers. In is natural to assume that the first capes made by the new modification would be rectangular in shape and the rectangular fine-meshed specimens in the Bishop Museum and the Florence Museum illustrate this further stage in the elaboration of technique. In addition to shape, the neck- and side-cords in square braid adhere to the technique of the coarser rectangular capes. The change in the size of mesh, however, necessitated other minor changes. Even with the smaller mesh, single feathers made too thin a covering and the feathers were accordingly tied together in bunches with a strip of olona-fibre. The feather-bunches were tied to the netting with the overhand knot as used in the coarser capes but as the meshes were so close together, it proved easier to use a continuous tying-thread instead of cutting the thread after tying each bunch (fig. 1, b-j). The use of fine-meshed netting as a foundation was followed by modifications in shape as evidenced by a number of transitional forms that have been preserved (pls. 4, 5). Some capes retained the long, straight neck-border but the lower border was further increased in length. The increased lower-border length was usually accompanied - 9 by increased depth in the middle line which produced a marked convex curve of the lower border. Other capes show a compensating concave curve at the neck-border,

slight in some and more marked in others. A deepening of the neck-curve with a corresponding increase in the lower-border curve and length resulted in the typical circular cape (pl. 1). The feather rows followed the curve of the lower border but the tying-cord remained continuous and the overhand-knot technique was retained with variations (fig. 1, b-j). The square neck-cord was also retained, but the cords that bounded the side-edges of the rectangular capes were dropped in some capes. A new technique, however, was necessary to obtain the marked upper and lower curves because the circular shape could not be produced by a netting-technique which runs in straight rows. The circular pattern had to be cut out of the prepared material, either out of one large piece or what was more economical and more common, by joining a number of smaller pieces together. The circular cape was thus a tailored garment, and it was the end-product of a series of local improvements in technique that reflects credit on the mentality and skill of the Hawaiian craftsmen. THE VALUE OF FEATHERS. The use of feathers for decorative purposes prevailed throughout Polynesia. Black, white, and brown feathers were readily obtainable from larger birds that were killed for food. The feathers were worked into ornaments, but outside of the completed ornament they had comparatively little value. Red feathers, however, had a very high value because red was considered the chiefly colour, and as such, they were deemed the appropriate decoration not only for chiefs but also for the religious symbols that represented the gods. In the Society islands, the value of red feathers reached its peak because the feathers themselves were regarded as the symbols of the deity. The value of red feathers was increased because they were difficult to obtain. In western Polynesia, the Samoans and Tongans traded with the Fijians for the red feathers of the Fijian parakeet. In central Polynesia, voyages were made to outlying islands to obtain supplies of red feathers. In far eastern Mangareva, where there were no red birds, the most valued articles to be obtained from visiting voyagers were the kura or red feathers brought from other islands. In Hawaii, the common colours in black, white, and brown were obtained from the domestic fowl which had been introduced by the early Hawaiian settlers. The additional feathers of the frigate and the tropicbirds were obtained from rookeries on neighboring islets without much trouble. It took comparatively few birds to supply sufficient feathers to cover a rectangular cape. When the advance in technique embraced the use of red feathers, the question of adequate supply created a new problem. Both the 'i'iwi and apapane were small birds with small feathers and a much greater quantity of feathers was required to cover a cape. The demand, however, was so great that the feathers assumed a distinct commercial value of the nature of currency. The prospects were sufficiently attractive to provide a full time job for professional feather-hunters who were termed poe kawili. The fowlers studied the habits and habitats of the birds. They caught them with nets and with bird-lime smeared on perches in the trees where the birds fed. As both the 'i'iwi and apapane were extensively covered with red feathers, they were killed and skinned. The fowlers ate the birds and the dried skins were packed usually in bundles of twenty for the local - 10 trade. The rectangular netting-foundation of small mesh in the Bishop Museum shows that it was originally covered entirely with red feathers. It is thus probable that the earliest form of fine-meshed capes were entirely red, which gave rise to the term 'ahu 'ula (red garment) being used as the general name for the feather capes. Though the red capes in their day must have been greatly admired, the artistic sense of the people did not remain satisfied with one colour. Another colour was sought for the production of patterns as had been done with the different colours of the coarser feathers. The yellow feathers of the 'o'o and the mamo appealed not only because of their brilliance but because they blended more smoothly with the red feathers in size and consequently in the technique of attachment. The commercial field of the feather-hunters was extended to include the 'o'o and the mamo. The 'o'o was the more plentiful and was

found in all the islands but the mamo had a more limited distribution. The 'o'o was a black bird with a body-tuft of yellow feathers near each wing and a few near the tail. The mamo was also black with a few. yellow feathers above and below the black tail-feathers and on the thighs. Except for colour, number, and position, the wing tufts of the 'o'o are somewhat similar to the single, white tuft under the neck of the New Zealand parson-bird (Prosthemadera novae-zealandiae). It is interesting to note that the Hawaiian name 'o'o with the glottal k restored would be koko; and the New Zealand bird, though generally known as tui, has the older classical name of koko. Some of the 'o'o birds were killed and skinned as evidenced by pieces of skin bearing yellow feathers on the Vienna rectangular cape and on a feather-girdle in the Bishop Museum. The black feathers of the 'o'o were also used in the colour-designs of some of the circular capes. The fact that both the 'o'o and the mamo were fewer in number than the 'i'iwi and that each bird supplied but a limited number of feathers, created another problem for the feather hunters. It was wasteful to kill the bird that grew the golden feathers and so the fowlers adjusted their operations for obtaining yellow feathers to coincide with the moulting season. Apart from the few 'o'o that were killed for specific purposes, the birds caught with bird lime were plucked of their yellow feathers without harm and set free to maintain a future supply. As the duller black feathers of the mamo were not used, it is probable that they were never killed wittingly. After plucking, the yellow feathers were tied together with olona-fibre into bunches for transport home. The mamo feathers were more highly prized than those of the 'o'o because of their deeper golden colour and their greater rarity. Yellow feathers became so highly valued that yellow displaced red as the colour of royalty, but the capes, even when yellow predominated, continued to be called 'ahu 'ula (red garments). The commercial value of yellow feathers is indicated by the statement of an early European visitor that five of the yellow feathers are bought for a hog. By means of red and yellow feathers, the feather-hunters were able to pay their rent to their overlords and procure food, clothing, and the other necessities of life. As it is certain that the finer feathers were preferred to the coarser ones, the occurrence of the coarser feathers in the coarse circular and the coarse rectangular capes raises the question of why they were used. It is possible that a sufficient number of the red and the yellow feathers were not procurable in the districts in which the coarser capes were made, but it is more probable that the chiefs for whom the capes were made were not in the social and economic position - 11 to acquire a sufficient quantity of the finer feathers. Hence the coarser capes became associated with the inferior chiefs and their continued manufacture was dictated by social causes. COLOUR-DESIGNS. In the Hawaiian capes, the colour-designs were influenced to a large extent by the shape and technique of the netting-foundation. In the rectangular capes, the netting-rows were in straight horizontal rows. The feather-attachment commenced at the straight lower border and consequently the feathers were attached in a straight row to similar parts of the net-meshes. Changes of colour in black, white, and brown were effected in horizontal straight lines, and the resulting geometrical figures were long panels and shorter rectangular figures arranged as the artist planned. In the circular capes, the first row of feathers to be attached had to follow the curve of the lower border. A number of successive rows, usually in yellow, followed and maintained the curve of the first row. Thus when a change of colour to red took place, the first red row was also curved. A number of capes have curved lines composed of a few rows of the contrasting colour, and it is probable from its simplicity that this was an early motif. An advance on the curved-line motif was made by gradually increasing the thickness of the motif in the middle and so producing the crescent motif (pl. 1). Once the crescent motif was evolved, it could be reproduced in smaller size in pairs with their inner ends meeting

in the middle line. The two upper edges of the crescent-pair formed two concave edges with a raised point in the middle and this combined with the long convex lower curve of the pure crescent produced a new motif very extensively used in both capes and cloaks (pl. 6). In some cloaks, the upper middle point of the crescent has been prolonged to such an extent that some museums refer to it as a battle axe. Other geometrical motifs were readily produced on the fine mesh beckground. SOCIAL INFLUENCE. The Hawaiian feather-worker was both craftsman and artist. As a craftsman he perfected the technique of the netting-foundation and the attachment of feathers. As an artist he improved on the shape of the cape and created the style of colour-arrangement. Artists worthy of the name have some inherent creative urge that need not always be attributed to a desire for material gain. The first rectangular feather-cape was an experiment and, when judged successful, it was presented by the craftsman to his chief. No commoner would have been allowed to wear a form of decoration that was superior to anything possessed by the chiefs. The craftsman received some material reward from his chief and probably praise and admiration from his fellows. At the very beginning, the relationship established between the feather-worker and the chief was that between the producer and the buyer. The chiefs were the patrons of the arts and crafts. They demanded the best of everything, and so they supplied the external stimulus that encouraged the craftsmen to greater effort in elaboration and invention. The first rectangular capes were made for chiefs, but with the improvement in technique, the chiefly fashion changed to circular capes in red and yellow. However, there were various grades of chiefs in Hawaiian society, and their economic resources varied with their grade and influence. Thus though the chiefs of a higher - 12 grade could have the circular capes made for them, the chiefs of a lower grade had to rest content with the coarser capes adorned with the more easily procurable feathers of the fowl, tropic-bird, and man-of-war hawk. Some chiefs were able to order the circular type of cape, but their resourses were insufficient to provide enough red and yellow feathers to cover the cape entirely with the finer feathers. On the other hand, the wealthier chiefs were able to buy more yellow feathers and hence the preponderance of yellow over red in any cape is probably an indication of family wealth and rank. The peak was reached by the high chiefs or so-called kings who ruled over large districts or islands; for with their greater resources, they were able to provide sufficient material to have the shoulder-cape expanded into a cloak that reached to the heels. Thus feather-capes and cloaks became the official regalia of the Hawaiian aristocracy and the type of garment indicated the grade of the chief. FINIS. Under the influence of western culture, a change in native values took place in Hawaii as elsewhere. The bark-cloth malo gave way to tailor made foreign garments, and the pa'u was supplanted by the ugly Mother Hubbard that was afterwards improved into the holoku in silks and satins with frills and flounces. The royal family of Hawaii abandoned the feather-cloak as the regalia of rank in favour of military and naval uniforms bedecked with braid and gold. The feather-capes and cloaks were given away or sold for new necessities. The 'o'o and the mamo became extinct through the destruction of their forest feeding-grounds and through introduced bird-diseases. The 'i'iwi survived in lessened numbers to provide an occasional splash of red in the sombre woods that had been spared by the woodman's axe. Silver and gold displaced red and yellow feathers as currency. The feather-hunters ceased to ply their trade, the feather-cape makers abandoned their craft, and the manufacture of feather-capes and cloaks became a forgotten art. In spite of material changes, however, a strong sentiment regarding the past still lived in the minds of the Hawaiians. Various patriotic societies were formed, and they sought to resurrect elements of their native

culture in ritual and dress. The capes and cloaks were reestablished as the tokens of a vanished splendour. In the new world of changed material goods, foreign loom-woven cloth took the place of the olona netting-foundation, the feathers of the peacock, guinea fowl, and other introduced birds supplanted the beautiful feathers of the now unprocurable native birds, and the overhand knots that nimble fingers had laboriously made with olona thread were replaced by a continuous stitch in cotton thread made much more easily and quickly with a Singer sewing machine. Nought remained of the ancient technique except the shape of the mantle and the use of feathers. A further discard took place when feathers were dropped for strips of red and yellow paper sewn on in overlapping frills in very neat and realistic fashion. Finally the circular capes were cut out of red or yellow cloth and colour alone represented the vanished feathers. Though these modern forms may seem but tawdry representations of the ancient capes, they serve their purpose as symbols of a past glory and, when worn by dignified Hawaiians, they create the native atmosphere that lingering sentiment and pride of race have sought to retain. - 13 LITERATURE CITED. 1. Te Rangi Hiroa, The Evolution of Maori Clothing: Polynesian Soc. Mem., vol. 7, New Plymouth, 1926. 2. Brigham, W. T., Hawaiian feather work: B. P. Bishop Mus., Mem., vol. 1, no. 1, Honolulu, 1899. 3. Brigham, W. T., Additional notes on Hawaiian feather work: B. P. Bishop Mus., Mem., vol. 1, no. 5, Honolulu, 1903. 4. Brigham, W. T., Additional notes on Hawaiian feather work: B. P. Bishop Mus., Mem., vol. 7, no. 1, Honolulu, 1918. 5. Roth, H. Ling, The Maori Mantle, Halifax, 1923. 6. King, Captain James, A voyage to the Pacific Ocean 1776-1780, vol. 3, and Atlas, London, 1784. PLATES. Plate 1. Fine circular cape, Bishop Museum (9670). Presented by King Kamehameha 4 in 1861 to Lady Jane Franklin, wife of the noted Arctic explorer. Background of yellow 'o'o feathers; lower half of large crestent, inner half crescents, and alternate neck squares of red apapane; upper half of large crescent, outer half crescents, and alternate squares of black 'o'o. Plate 2. Fine circular cape, Bishop Museum (957). Back view showing fine netting. From Queen Emma collection. Plate 3. Coarse circular cape, Australian Museum, Sydney. Collected on Cook's third expedition. Triangles of yellow 'o'o and red 'i'iwi feathers; body probably of frigate-bird feathers. Plate 4. Transitional shape, fine feathers, Vienna Museum (179). From Cook's third expedition, purchased in London in 1806 by order of Emperor Francis 2 of Austria. Red 'i'iwi and yellow 'o'o feathers. Plate 5. Transitional shape, coarse feathers, Madrid Museum. Fine feathers along neck and front borders lost; body probably of black cock's plumes. Plate 6. Circular cloak, fine feathers, Bishop Museum (C.9558). Background of red 'i'iwi feathers; lower border and motifs of yellow 'o'o. Plate 7. Circular cloak, coarse feathers, Dominion Museum, Wellington, N.Z. From Cook collection, presented by Lord St. Oswald. Front triangles and neck-bands of red 'i'iwi and yellow 'o'o feathers; body apparently of black cock's plumes.

Plate 8. Rectangular cape, coarse feathers, Vienna Museum (180). From Cook collection. Upper tapa band with pieces of 'i'iwi and 'o'o skin gummed on; coarse feathers attached to netfoundation. Plate 9. Back view of Vienna cape (pl. 8). Coarse netting-foundation, showing natural shape of meshes behind upper tapa-band and hexagonal-shaped meshes below through tying on of feathers. Plate 10. Ti-leaf rain cape, Bishop Museum (311). Front showing ti-leaf overlapping thatch. Plate 11. Back of rain-cape (pl. 10). Coarse net-foundation. - 14 FIGURES. - 15 Figure 1. Tying-technique of fine feather-bunches, a, spiral tying of quills of feather-bunches, b, single overhand knot, half hitch, and horizontal continuation, c, feather-bunch laid over two mesh-knots and first lower overhand-knot made, d, two overhand-knots with oblique continuation, e, two overhand-knots with cord about to return to lower knot, f, two overhandknots, with third knot over lower knot and horizontal continuation, g, three overhand-knots, with fourth knot over lower knot and horizontal continuation, h, three overhand-knots with oblique continuation, i, three overhand-knots with fourth knot over lower knot and horizontal continuation, j, variation to h in position on meshes. Figure 2. Tying-technique of single coarse feathers, a (after Ling Roth), single feathers tied over each mesh-knot, giving mesh a hexagonal appearance, b, feather-quill laid over a mesh-knot, c, the tying-cord makes an overhand-knot (1) around the quill and the two arms of the mesh below the mesh-knot, the end of the quill (2) is doubled back, and the tying-cord makes a second overhand-knot (3) around the doubled quill and the two arms of the mesh above the mesh-knot. Figure 3. Tying-technique of ii-leaves, showing commencement along lower border of netting. The stalk end (1) of the half ti-leaf (2) is passed around the back of the two arms of a mesh below the mesh-knot and the free end is tied around its standing part in an overhand-knot (3), thus forming a slip-knot. The slip-knot is then drawn taut against the mesh-knot. (To save space in the figure, the ti-leaves are shown tied below each mesh-knot instead of alternate mesh-knots.)

- 16 Fig. 1.

Fig. 2. - i Fig. 3.

- ii PLATE 1. - iii PLATE 2.

- iv PLATE 3. - v PLATE 4.

- vi PLATE 5. - vii PLATE 6.

- viii PLATE 7. - ix PLATE 8. PLATE 9.

- x - xi PLATE 10.

- xii PLATE 11.