122 SISSANO. 
11. mi-mbiaa Kowamerara, Sigab. 17. bibtts Put, Kumenim. 
?2. bttai Namarodu, Lalinau, Punam, | 18. bok6 Hamatana. 
Bissapu, Laur, Kamban- | 19. pua San Cristoval, Maramasiki. 
geriu, Kondo, Yalui, Lam- Ulawa. 
bell, Lamassa. 20. vaé Avelus, Lauan, Mongai, Sali, 
13. 4-mbtiai Gazeile Peninsula. 21. potu Vrinagol. 
14. bua-kau Pokau. 22. pou Paup, Yakomul. 
15. bua-tau Motu. 23. pupts Paup, Yakomul. 
16. baaché Burruwe. 24. fum Afue. 
INDONESIAN. 
25. bua Ceram. 31. fua Buru. 
26. buah Ambon, Bahasa. 32. hua Bahasa. 
27. buwa Ceram. 33. huah Ambon, Bahasa. 
28. pua Maldives, Rotti, Banda. 34. huwa Ambon, Bahasa. 
29. puah Timor, Ambon, Bahasa. 35. hual Bahasa. 
30. puwak Ceylon. 36. wua Tombulu, Tonsea. 
Through this series continuous through Indonesia and into the north- 
ern islands of Melanesia there runs a straight thread of easy identi- 
fications of a stem of which the most frequent typeis bua. ‘The vari- 
ants of the initial labial lie within the range of mutation with which 
we have made ourselves familiar—from mute through spirant to 
semivowel in 36 wua and to aspiration in 32 hua. In a few cases we 
find an accretion to the standard stem; in 35 hual a terminal liquid, in 
30 puwak a terminal palatal mute, but here we must note that all the 
languages of Ceylon lie outside the linguistic province with which we 
are dealing; in 16 biia-ché we meet a palatal of uncertain import, in 13 
a-mbtai we find a vocalic accretion. In 14 and 15 bua-kau of Pokau 
and bua-tau of Motu it is clear that we have dialectic differences of 
the same word. ‘The former element is manifestly the most completely 
developed form of the present stem bua. Ray notes that Motu t 
regularly becomes k in Pokau; therefore we accept tau as primal in 
this pair; this negatives the immediate guess that in this compaction 
we recognize the common stem kau tree, and in both these languages 
(p. 31, No. 35) we find that kau is represented by the form au. The 
vowel skeleton of the stem is represented by the pair u-a and the 
former undergoes no change in the variant forms which remain dis- 
syllabic. The latter vowel undergoes mutation from a to e in 5-7 
and 20, to 0 in 8 and 9g. 
Because of the intricate interlacing of monosyllabic with dissyllabic 
forms of the stem in closely related languages, we feel justified in 
classing with the bua stem a stem of the type bu; but without attempt- 
ing to settle whether this is a more primitive type or the product of 
degradation by abrasion, save that it is proper to repeat the note 
that the abrasion of a final vowel with which we are most familiar is 
directed to the production of a closed syllable and does not seem appli- 
cable to any instance in which abrasion exposes another vowel. Atten- 
tion may be directed upon the occurrence of stem buin Uap of Micro- 
nesia, a storehouse of much that seems to represent a very primitive 
