64 SISSANO. 
but we may defer its consideration until we recur to the stem in the 
discussion of Indonesian relations. 
Before leaving the Polynesian identifications we must consider a 
note made by Pratt in the Samoan dictionary. In the first edition 
of that work (1862) the learned author sets against ‘ulo the comment 
“introduced.’’ ‘The second edition (1876) is lacking to my library; 
but in the third (1892) the note had been amplified to the specific 
statement ‘‘a Tongan word.’’ I called to the attention of the Rev. 
J. E. Newell the fact that, if introduced at all, the word was quite as 
much loan material in Tonga as in Samoa and that its nearest source 
was to be found in Viti kuro. Agreeing with me in general, he was so 
conservative in his editing of the fourth edition (1911) of Pratt that 
he did not alter the note. In Tonga there is no manufacture of pottery 
nor working of clay; the pots were an object of oversea trade with 
Fiji and the name came with the pots, the only change being the shift 
from r to 1 in conformity with a phonetic principle operative in Tonga 
and Samoa. 
Tregear and Percy Smith make an equivalent note upon ilo in 
Niué. The word and the object may have been carried from Tonga 
to Samoa and to Niué, for the latter had no commerce with Fiji and 
the Samoan intercourse with Fijians in the early legend period I hope 
in time to prove had naught to do with the archipelago which at 
present bears the designation of Fiji or Viti. Notwithstanding this 
lack of intercourse, the ascription of the word to Tonga is less satis- 
factory than its reference to Viti as the point of original distribution. 
In the Melanesian series we have some very satisfactory identifications 
in devolution order. For technical reasons there is a lack of instances 
of the stem in Melanesia proper; in fact all that I have recovered are 
1,70; 10,010. 
In this connection we should observe also that Codrington, our 
best source of Melanesian vocables in comparison, has not noted this 
word in his tables; therefore it is not impossible that as more vocabu- 
laries of Melanesia become available we shall be able to fill up these 
lacune. ‘Tangoan kuro is absolutely established as of this stem, for 
it is exactly the Viti form. So much of Aneityum nipji-uru as may 
be related to this stem involves the change of the latter stem vowel 
from o tou. ‘This may be regarded as supported by the confirmation 
of 16-17. ‘The establishment of 11 kére requires the confirmation of 
the variation of the latter vowel from o to e and the variation of the 
former stem vowel from u too. Neither of these is at all unusual in 
the mutation system developed in these Melanesian studies. The 
u-o mutation is abundantly upheld by 10 ol and 16 olun; the o—e muta- 
tion finds support in 12-14. ‘The citation 19 buro which Friederici 
draws from Efaté I have been unable to verify, for it does not appear 
in Macdonald’s dictionary of some speech of that island. If accu- 
