Linton — TJie Marquesas Islands 331 



the lower part of a narrow toki koitina whose tang has been broken off. The mi- 

 plement appears to be complete however, and shows no signs of hafting. It is 

 possible that this implement marks a transition from toki koitiiia to gouges and 

 that the two classes of implements are related. 



KNIVES AND SCRAPERS 



Marquesan knives were usually made from bamboo or shell, and shells were 

 used as scrapers, but stone knives and scrapers also existed. In the absence of 

 iron tools, stone knives have been used as a makeshift until comparatively recent 

 times. Pere Simeon Delmar, head of the Catholic Mission in Taiohae, Nuku 

 Hiva, has witnessed the manufacture of stone knives. He says that the knives, 

 often six to eight inches in length, were made from flakes of a special variety 

 of phonolite, which were detached from a core by a single sharp blow with a 

 rounded stone held in the hand. The natural fracture gave a fairly sharp edge 

 to the flake and it was used without retouching or hafting. When the edge be- 

 came dull from use, the implement was thrown away. In M. Delmar's experi- 

 ence these stone knives were used only to open and scale fish. Many phonolite 

 flakes were found in old village sites, but their certain identification as knives is 

 impossible. (See PI. xlviii, G.) 



Two stone artifacts were collected which appear to have been scrapers. 

 One of these, found in an old village site, is a large semicircular spall of some 

 basaltic rock which has evidently been retouched along its lower edge. Its form 

 is not unlike that of some of the shell scrapers used for tapa making. (See PI. 

 XLVIII, £, I.) 



The second specimen was collected on the island of Ei Ao. It is a thin 

 spall of phonolite which has been worked into rectangular form and dressed 

 along the outer end and one side with an abrupt retouch. This implement ap- 

 pears to be complete, and is much too thin and carefully worked to be a reject 

 from the adz workshops. It was identified by natives as a scraper. (See PL 

 XI.VIII, E, 2.) 



RUBBING STONES 



The Marcjuesans used rubbing stones in the manufacture of various wooden 

 objects and in dressing the large stone slabs which are a common feature of their 

 masonry. The large rocks upon which adzes were ground and polished have al- 

 ready been described (p. 321) and mention has been made of the probable use 

 of whetstones, but it is vmcertain whether rubbing stones were used in the manu- 

 facture of other stone artifacts. The modern makers of popoi pounders deny 

 their use, and none of the ancient artifacts show a finish which could not have 

 been produced by careful pecking and a final polishing with sand. The large flat 



[71] 



