96 SISSANO. 
MELANESIAN. 
1. usa Sesake, Marina, Maewo, Va- | 23. kusch  Sissano. 
turanga, Tangoan Santo, | 24. katts Paup. 
Suau. 25. Vus Sissano, Arop. 
2. usana  Dobu, Kubiri. 26. ua Efaté, Epi, Ndeni, Koiari. 
3. usan Kiviri. 27. yuo Baki. 
4. us Santo, Malekula. 28. aud Bongu, Bogadjim. 
5. eus New Ireland. 29. O Ambrym. 
6. ous Nokon. 30. OT Graget. 
7, na sawa Kilenge. 31. wa Yoda, Binandele, Amara. 
8. uha Ngela, Bugotu. 32. bwa Efaté. 
9. uh Lakon. 33. Va Koiari, Kokila. 
10. uhe Arag, Omba. 34. nihua  Bierian. 
11. ehe Aneityum, Eromanga. 35. buk Sér. 
12. usha Aola. 36. reu Arag, Merlav. 
13. utha Guadalcanar. 37. abara_ Roro. 
14. ura Rubi, Manam, Vitu, 38. awara_ Barriai. 
15. urata Buka, Bougainville. 39. auaha_ Kobe. 
16. uruotta Bougainville. 40. wat Pak. 
17. urei Gog. 41. wet Sasar, Alo Teqel. 
18. uta Alite, Gower Island. 42. weta Lo. 
19. kusana Mukawa. 43. wend Volow. 
20. kuse Sariba. 44. wen Norbarbar, Vuras, Mosin. 
21. qesi Murua. 45. wena Mota. 
22. qes Nada. 
The Proto-Samoan stem has been established as uha, retained in 
Niué and Tonga in Polynesia, in 8 Ngela and Bugotu and_ readily 
identifiable in 9-11 in Melanesia. From this primal stem certain of 
these languages proceed by obliterating the inconvenient aspiration 
yielding the ua type. This has been the case most largely in Poly- 
nesia; in Melanesia we find 26 ua in the New Hebrides and in one of 
the so-called Papuan languages of Torres Straits; we can trace this 
form through 27-30; in 31 we find the first stage of hardening in the 
employment of the semivowel; in 32 it has advanced a stage farther 
and has become distinguished by one of the amorphous labials so 
frequent in Melanesia; thence in 33 has passed to a more sharply 
defined labial; in 34 we find ua or metathetic uha with a preface of 
undetermined sense. 
In certain other of these languages the inconvenient aspiration is 
disposed of by a process of strengthening, in which, so long as the 
mutations are limited to the lingual tract, we have no difficulty in 
following the movement. ‘The briefest mutation ish-s. ‘This appears 
metathetically in Rotuma was, is found in 1-6, possibly in 7, and 
again in 19-22 and 24-25. ‘The mutation to shis found indisputably 
in 12 and seems to appear in 23, but as to the latter we must note 
that the validity of the word needs determination and that the initial 
k may be found to remove 19-24 from affiliation with the uha stem; 
yet in the same connection we observe that vus, which Friederici 
records in the place of kusch, offers the sole instance of a labial pre- 
face, for in 31-33 we have established that the labial is not a preface 
but a modification of the initial u. The next stronger mutation th is 
found in 13 utha from Guadalcanar and a slightly advanced stage in 
