120 THE SUBANU. 
a primal stem su, which is the only common factor entering into the 
several vocables there collated. This Subanu soong I regard as con- 
firmation of that judgment. From this primal su various determinant 
vocables have been formed. With a wider range of Indonesian mate- 
rial than was then accessible to me, I may arrange the material from 
this family in a provisional series. The key is the mutation of the s, 
weakly to the lingual liquid, strongly to its mute. Assuming the sec- 
ondary vocable zsu, which we find as the most common stem in Poly- 
nesian, we now list the mutations. 
iru Ambon, Kolon. nggilung Minahassa. 
irung Java. ili Ambon. 
idung Malay. uru-na Malagasy, 
ileng Bontoc Igorot. kam-—uru Macassar. 
hiru Ambon. urong Dayak. 
niru Allor, Ceram, Minahassa. ninura Ambon. 
nirun Kei. nunu Ternate. 
ngirung Minahassa. ngunu  MHalmaheira. 
iri Ambon. usnut Gani. 
ilu Bima. 
The Ambon dialectic forms serve to link together widely variant 
types in a continuity which otherwise would not be discoverable. The 
recurrence of final ng (n) in so many of these variant forms tends to 
establish that final consonant in Subanu soong as pertaining to the 
primal stem, on which point refer to the note under item 24. The chief 
links in this Indonesian chain are found in Melanesia, and particularly 
in the important region of the north shore of Torres Strait. The four 
entries at the end of the list are presented to complete the record so far 
as it goes; quite clearly they pertain in some fashion to the series, but 
for the present they stand as somewhat anomalous. 
28. kapa to flap the wings; Visayan capacapa id. P. W. 295. 
kapa Tonga, Futuna, Niué, Uvea, Manga- | kapak Malay. 
reva, Mangaia, Maori, Nuguria. pacpac Tagalog, Bicol. 
‘apa Samoa. pak-sa Kawi. 
apa Tahiti. pak-si Basakrama. 
pa  Fotuna, Rotuma. papak Magindano, Baliyon. 
kapakapa Magindano, Visayan. | 
It is quite plain that we are concerned here with two stems, or in 
better likelihood a primal stem with determinant accretion. ‘The 
primal stem seems to be pak, the derivative kapak. In the general 
theory of the evolution of isolating vocables we should look to find the 
primal stem in the possession of the earliest phase of the speech. ‘The 
evidence here presented is not decisive. ‘The pak stem is found as far 
to the west as Java—truly in the ancient speech, since it is credited to 
the Kawi, and to the Basakrama, which is frequently conservative of 
archaic forms; eastward, in the region of the archetype of Malayan 
speech, it is found in the Philippines in Magindano, Tagalog, Bicol, 
and in the immediately associable Baliyon of the Borneo Dayaks. 
Yet in composition with kaw, “a projecting member,’’ the primal stem 
pak appears in Polynesia in these words for “wing” as ‘‘flap-limb,”’ 
