382 Memoirs Bcniicc P. Iiishof' Miiscitiii 



see p. 283.) According- to Handy (personal correspondence) there was a kind 

 of coarse mat, called Iiiipaii. which was used by poor people. 



No pandanus mats have been made in tlie Marquesas for several years, and 

 the informatiim obtained in regard to the methods of manufacture was insut- 

 ficient and sometimes contradictorv. 



The (Irv or half dead leaves appear tn have been used. The base, niidrili and thnrny 

 edges were cut away with a pearl shell scraper or knife of the sort used in scraping bark for 

 tnpa. leaving two long straight stri]5s about two inches wide. .According to one informant, these 

 strips were then soaked in sea water for a few days, washed in fresh water, and dried in the 

 sun. The extreme softness of some of the mats would seem to indicate that the dried strips 

 were beaten or otherwise manipulated before weaving, but no information could be obtained 

 on this point. Before weaving the broad prepared strips were split with a small stick like a 

 needle into narrow sections which varied in width according to the degree of fineness of the 

 contemplated mat. These strips were called hcnii ( ^2) . Marquesan mats appear to have been 

 quite coarse as a rule, the width of the strips in all the specimens seen averaging three- 

 eighths of an inch. The weaving was done entirelv by women and the manufacture of some 

 mats for special purposes was attended by ceremonial observances. According to one in- 

 formant the work was done upon a sort of lap-board placed across the knees. Some mats 

 were of single, others of double thickness, but the same methods were employed in making 

 both, the double mats differing only in having each element made from two strips of the 

 leaf laid one upon the other. The weave was a simple checkerwork, the warp and weft 

 being indistinguishable and the strips running at an angle of forty-five degrees with the edges 

 of the mat. Each element, as it reached the edge, was folded over and woven back into the 

 mat at right angles to its former direction. A weft strip thus became, when turned and 

 carried back, part of the warp. In one specimen, apparently a section cut from a large 

 mat, the ends are finished by cutting ofif the ends of the warp strips and bending back and re- 

 weaving these of the weft strips in the manner just described. The sides of this mat are 

 finished by the first method. In a second specimen the edges are finished in a more com- 

 plicated manner which gives an ornamental elTect. (See PL lxvi, B.) The exact technique 

 can not be ascertained without injuring the specimen. 



Marquesan mats were rarely if every decorated with woven patterns, but 

 one specimen in the Bishop Museum (PI. Lx\i, O is embroidered in simple over- 

 lay with thin strips made from the upper surface of the ])andanus leal and dyed 

 lirick red. The decoration consists of six narrow stri])S, each two elements wide, 

 two of which are placed at either end of the mat as a border while the other 

 four are arranged longitudinally at equal intervals, the outermost strips l)eing 

 about four inches within the sides of the mat. According to Handy (personal 

 corres]>ondence) the cuttings of women's hair were sometimes saved and woven 

 in tufts around the edges of mats. 



In view of the scanty material available for study it is quite possible that 

 mats of finer grade than those described were manufactured in ancient times. 

 but the general inferiority of Marquesan mats is meiitioned by various early visit- 

 ors. Petit Thouars (48, p. 348) says that the mats were coarser and less (irna- 

 mental than those of the people living further west, and the \vife nf nne of the 

 first European missionaries sent to the group from Hawaii mentions, in a letter 



[122] 



