436 Memoirs Bcniicc P. Bisliof' Miiseimi 



ment \va^ thus held rigid, reducing' tlie danger of breakage. It was then rolled in 

 several thicknesses of ta])a. When the teeth in a wreath had become yellow or dirty they 

 were scoured with fine sand. 



A curious type of headdres.s (packalia) is shown on Plate i.xxxi. ./. Its 

 manufacture appears to have been limited to Hiva Oa. Tan Ata and Fatu Hiva. 

 It apparently was not made in Ua Pou (Handy) nor in Xuku Hiva, although 

 ornaments of this sort, obtained in trade, were sometimes worn in Xuku Hiva 

 during the early historic ])eriod. Some informants of Hiva Oa sav that it was 

 not worn in Hiva ( )a in ancient times, and ascribe it to Fatu Hiva. 



The packalia consists of a number of plaques of shell and tortoise shell which are at- 

 tached to a band of woven coconut fiber. The warp strands run longitudinally, the weft 

 strands passing back and forth across the band in a wrap weave. The direction of the wra])- 

 ping is changed at regular intervals, ]iroducing an ornamental banded effect. The edges of 

 the band are re-iii forced by cords somewhat heavier than tlie regular warp strands, and its 

 ends are finished in stiff loops. The form and weaving of the band is shown on Plate 

 L.xx.xi, C. The outer surface of the band was originally covered with small discs of mother 

 of pearl about the size of shirt buttons. A small toothed di.sc of tortoise shell was fastened 

 to the center of each pearl disc. Machine made buttons of china and shell were introduced 

 by the traders at an early time, and soon supplanted the old pearl discs on these headdresses. 

 The ends of the band were covered by pearl shell plates, overlaid with thin openwork strips 

 of tortoise shell. The form and ordinary design of these plates can be seen on Plate l.xx.xi, A. 

 The end plates were called kana. the small discs, kana inoino (Handy). 



The plaques, which form the princi])le part of the ornament, were fastened to the lower 

 edge of the band. There seems to have been no fi.xed rule as to the number of plaques, 

 but more than seven tortoise shell and eight shell plaques were rarely employed. They were 

 attached to the band by coconut fibers passed through small holes drilled at their tops, and were 

 fastened together near the lower ends in the same way. The shell plaques were made from 

 large conch or Triton shells, and were highly polished on all surfaces. They were made from 

 the lower end of the shell and preserved its natural curve. The tortoise shell plaques were 

 two or three times as wide as the shell pla(]ues, and were elaborately carved. The decoration 

 usually consisted of a large highly conventialized human figure in the center, with panels of 

 superposed smaller figures or designs on either side. A narrow band of designs ran across 

 the lower end. The spaces on each side of the head of the large central figure were pierced. 

 In a few specimens the central figure was replaced by two smaller figures, but the examples 

 shown on Plate lxxxi. A, B may be taken as typical. The most curious feature of this decora- 

 tion is that, when the ornament was on the head, the figures were all upside down. This 

 lends some support to the statement of one informant that the packalia was originally worn 

 with the plaques above instead of below the band, although all natives agree that in historic 

 times it was worn with the plaques inverted. 



The tortoise shell plaques were bent outward at the bottom, their curve conforming 

 roughly to that of the shell plaques. When the ornament was on the head the lower edge 

 flared outward over the eyes, like the visor of a cap. Many of the tortoise shell plaques are 

 perforated at the lower edge, suggesting that small ornaments were suspended from them. .\n 

 ornament of white beard i paz'ahina ) was usually attached to the center of the packca band 

 and projected above it. 



The materials used in the packalia were all obtained locally. Two varieties of tortoise 

 shell bearing turtles were distinguished by the natives, the shell of one species being thick and 

 easily worked, and that of the other thin and brittle. Xo method of cementing the plates 

 together seems to have been known, and the way in which the tortoise shell plaques were bent 

 has been forgotten. Tortoise shell was highly valued by the natlyes, and the white traders 

 met the demand by imijorting brown sheet celluloid. Many of the packalia in American col- 



[176] 



