INDONESIAN ANNOTATIONS ON THE VOCABULARY. 109 
40 wasua and 41 jassu, and reaches its limit in the initial palatal 
mute k in 35-37. In 40 wasua the assumption of a final vowel is 
not matched in any other use of this stem. The stem asu is not 
traceable in our Melanesian material, except in so far as we might seek 
to establish a devolution series in the form asu-ahu-au; but this direc- 
tion of variety is quite unlikely, as I shall show in its proper place. 
We must not neglect to point out that in Viti oso to bark we have 
the designation of one of the dog’s faithful activities which closely 
resembles this stem, although the Melanesian gap is as yet unbridged. 
Stem afuna.—The characteristic form of this stem consists of the 
vowel succession a-u with or without an initial palatal mute, but 
invariably parted by one of the labials. This preface may be k as in 
30, 33, 18, 19, 20, 21; g asin 22: or the nude stem asin 31 and 32. ‘The 
labial which we establish as critical of this stem may be f as in 30 and 
31; p as in 32-33, 18-21; or b asin 22. It is quite possible that this 
labial septum is represented not only by the spirant and the mute, 
but by the semi-vowel adjacent to the labial series, w. With this 
employment of w as a resultant product of the labial inefficiency we 
have already become quite familiar, and the resemblances of semi- 
vocalic forms with purely consonantal forms are clearly shown, 8 
kawuna with 30 kafuna, 14 gawun with 22 gabun, 26 awuna and 27 wona 
with 31 affna. The stem in Indonesia and largely throughout Mela- 
nesia is trisyllabic, the final syllable being for the most part -na, the 
consonant has undergone mutation n-l in 32 apula and 21 kapul, a muta- 
tion supported by an abundance of established instances; the final 
vowel a mutates to e in 19 kapiné, and has undergone abrasion in 
14 gawun, 20 kopon, 21 kapul, and 22 gabun. 
Stem auna.—Here we enter upon the intricacies of the origin of this 
series of dog words. In the consideration of the stem afuna we have 
carried the series from the employment of the strong labial mute through 
the spirant down to the semi-vowel. Regarded solely as an academic 
problem in phonetics, we may quite as readily proceed to the next 
step, the weakening of the labial already debilitated to w, and find 
the last stage in auna, which preserves the characteristic vowel pair 
a-u and the specific consonant of the final syllable. On this hypo- 
thesis we should from 8 kawuna derive 34 and 6 kauna and with a 
recognized vowel change 7 kaunéd and under abrasion from kauna to 
kaun (not found) we should reach 9 koung; similarly from 14 gawun 
developed by abrasion from gawuna (not found) we should derive 15 
gaune and therewith include as a degradation form 16 g’one and from 
gawun direct 12 gaun, 13 gaon, and, with mutation of the initial con- 
sonant, 17 ngaun. In like manner 26 awuna through an as yet undis- 
covered awun should direct us to 25 aun. Still regarding the n as 
critical, we are left with a few forms which seem associable with this 
stem, yet which are somewhat anomalous: 24 haun may be under- 
