Gifford — Tongan Place Names 23 



graphic propinquity. Actually Haapai and Vavau have only 86 names ex- 

 clusively in common, while Haapai and Tongatabu have 193, and Tonga- 

 tabu and Vavau 212. The evidence seems to point to some disturbing 

 historical factor, perhaps intermittent hostility such as was rife between 

 Haapai and Vavau during Mariner's sojourn in Tonga. 



The situation revealed by the distribution of Tongan place names is 

 roughly shown by the accompanying diagram which is based on the 

 assumption that the principal diffusion has been from the large groups 

 to the small ones. Numerous minor diffusions have doubtless taken 

 place, many of them being reciprocated — that is, for example, not only 

 have names flowed outward from Tongatabu, but they have likewise flowed 

 inward to Tongatabu. 



Niuatoputapu, 

 Niuafoou. -" ' 



IbngatdbiP 



The diagram, and in fact this whole discussion, tacitly assumes a pri- 

 macy for Tongatabu, not only politically, but also in length of period of 

 human occupation. In regard to this last point I should like to state that 

 kitchen middens several feet in depth are to be found in a number of 

 places in the Tongatabu group. With the exception of a thin surface of 

 blackened soil with scattered shell, which I observed on Euakafa island in 

 the Vavau group, I failed to find either in Haapai or Vavau any evidences 

 of prolonged occupancy comparable with those in Tongatabu. Of course, 

 the Tongatabu shell heaps may be interpreted as indicating a denser popu- 

 lation, but it would be rash to entirely cast aside the theory of longer 

 occupancy. Quite likely the magnitude of the shell heaps is due to both 

 factors, longer occupancy and denser population. 



SAMO.\N PLACE NAMES IN TONGA 



Five hundred and sixteen place names mentioned by Kramer'^ have 

 been examined for Tongan parallels. One hundred and one of the Samoan 



19 Kramer, Augustin, Die Samoa-Iiiseln, Stuttgart, 1902, 1903. 



