POLYNESIAN AND MALAYAN. 129 
The languages comprised in this table are these: 
ai Wango. tagai Mota. tagar Gog. 
gae Fagani. tigai Maewo. tga Motlav. 
bwai Wango. aga Pak. tigi Maewo. 
pwai-ke Ulawa. ’iga Leon. tig Maewo. 
hai-ke Saa. taga Merlav. teji Norbarbar. 
It will be seen that the preface members of such composites as are 
in this list are variants of the te type. 
The Subanu da and di are sonant varieties of the ta and #, of which 
we have evidence in Melanesia, and its daay is paralleled by tagaz. 
Bontoc Igorot yields us dd. 
43. liko around; Subanu molio curved. 
liko Futuna, Viti. balico Visayan. 
li‘o Samoa. molid Subanu. 
This is the least-used of a group of three consimilars, liko, niko, 
and piko, in which we may recognize as operative the effect of conso- 
nant modulant prefaces upon a primal stemzko. ‘The /zko form is found 
at these three stations in Nuclear Polynesia and doubly at this single 
Philippine station; relative to the speech families in which they occur, 
Nuclear Polynesia and the Philippines are recognized as archetypal and 
representative of the Proto-Samoan. In Subanu we have no difficulty 
in dissecting out the particle of condition ma, and the ba of the Visayan 
is a familiar variant of the same. 
44. longo to hear; Visayan dongog id. P. W. 398. 
longo Samoa, Nukuoro, Futuna, Uvea, ee 
Niué, Fakaofo, Rarotonga. langan Matu. 
rongo Viti, Maori, Mangaia, Mangareva, | rungu Java. 
Paumotu, Rapanui, Aniwa, | rohona Malagasy. 
Fotuna. rungak Uap. 
ongo ‘Tonga. hungu Chamorro. 
lono Hawaii, Nuguria. dongog Visayan. 
ono Marquesas. déngek Bontoc Igorot. 
oko Marquesas. dangar Malay. 
In Polynesia we lack derivative forms which might protect a final 
consonant if this had been a closed stem. ‘The final mute palatal in 
dongog and rungak I incline to regard as verb-formative suffix in the 
eastern Malayan; it is suggested again only in Omba ronghogosi, and 
there obscurely, for we have no means of determining if the g is terminal 
of the stem or initial to the latter composition member. In examining 
the Melanesian material we find suggestions of final m in Vaturanga 
and of final v in Vaturanga and Kabadi. In our slovenly American 
orthoepy it may not be wholly unnecessary to draw attention to the 
fact that the Malay recognize in /angar full consonant value for the 
final r; this seems to belong to the stem, at least in one stage of its 
development, for it recurs in Lambell, King, Duke of York, Baravon, 
Raluana, Mukawa, Tavara, Wedau, Awalama, Taupota, Oiun, and 
Raqa. It will be seen that these are stations at the two exits from 
Indonesia, five at the gateway through the Bismarck Archipelago, seven 
