Linton — The Marquesas Islands 275 



in which they occur are vmquestionably those of ordinary dwelHngs. No pits 

 were observed in the dwelhng pacpac of Hiva Oa and the natives had no knowl- 

 edge of such a feature. The single example seen on Fatu Hiva was in the valley of 

 Omo'a in a pacpae built for a chief's house ; it was explained as a place for the 

 temporary storage of food. The occurrence and use of these pits is of interest in 

 connection with the skull pits found in many mc'ae throughout the group. 



All the INIarquesan valleys afford an abundance of loose stone, a condition 

 which, coupled with the constant necessity of building platforms and terraces, led 

 to a development of the art of dry stone masonry truly remarkable in a people 

 otherwise as primitive as the Marquesans. It would be diiificult for Europeans to 

 improve on the workmanship shown in many of the ordinarv pacpac. The 

 builders seem to have taken pride in using the largest material available, bowld- 

 ers weighing half a ton being quite common in these purely utilitarian strvictures. 



POSITION OF THE HOUSE 



As a rule there was only one house on a pacpac, but a few pacpac were ob- 

 served on which there appeared to have been two, and possibly even three houses 

 arranged longitudinally on the raised rear portion. None of the sites of this sort 

 looked to be very old, and it seems probable that this was a recent development. 

 Such a house group consisted of one large structure, the dwelling, with a small 

 detached building at one or either end. The earth of the unpaved floors 

 of the small buildings contained much ash and charcoal, proving them to have 

 been cook houses. Normally a single house covered the entire rear raised section 

 of the pacpac, the floor of which was divided longitudinally into two almost 

 equal sections. The front section, which formed the house floor, was paved with 

 flat water-worn bowlders, usually small and of uniform size. The rear section, 

 which formed the foundation for the bed, was unpa^•ed, and its surface stood six 

 to eight inches below the level of the floor. 



FRAMEWORK OF THE HOUSE 



The main timbers of the house were commonly made from breadfruit 

 wood, but inio (Thcspcsia populnea) was also employed, especially when the posts 

 were carved into tiki form. The timbers were carefully hewn and dressed and 

 sometimes were elaborately carved. The essential elements of the framework were 

 the end posts and the ridgepole, the front posts and the stringer. The great 

 length of some of the ancient houses, upward of a hundred feet, necessitated the 

 use of additional posts in regard to which Gracia says (28, p. 123), "If it is 

 thought on account of the length [of the house] that the first two posts are not suf- 

 ficient they plant a third just in the middle of the hut, or even a row." None of these 

 very large structures have survived, and it is impossible to give additional details. 



[15] 



