98 SISSANO. 
34. woate Malanta. 54. OS Dyaul. 
35. uos » Efaté. 55. hosa Palabong. 
36. uohe Efaté. 56. wosa Matantuduk. 
37. She Jenbi. 57. wosa Namarodu. 
38. Ose Emirau, Eluaue, Emsau, | 58. uasa Pala. 
Kung. 59. wasa Bisapu. 
39. ode Kabadi. 60. wasa Pire. 
40. 0e Waremo. 61. Was Lihir. 
41. faso Jamna. 62. wés Mahur, Massait, Mali. 
42. VOSO King, Kait. 63. Vole Nakanai. 
43. voho Epi. 64. pore Barriai, Kobe, Wuvulu. 
44. hds6 Labtr. 65. ore Manam. 
45. g6sd Lamussong. 66. wai Jio. 
45. Oso Kawiéng. 67. ai Yakomiul. 
47. 686 Komalabt. 68. aiis Paup. 
48. 6s6 Komalt. 69. ais Ali. 
49. 6s6 Kokola. 70. aus Anggél. 
50. 6s0 Panakondo. 71. VO Kabakaul, Mol6t. 
51. 6s0 Murapa. 72. vea Tanna. 
52. a-6s6 Belik. 73. na-pe Kilenge. 
53. OS Lemusmus. 
The stem is fohe and in most of the varied forms here assembled we 
engage with the mutations of the two consonants. In the first forty 
items we run the whole course of lingual variability even to extinc- 
tion, and nothing calls for particular comment except 37—40, in which 
we find, in remotest New Guinea, the form ohe, which recurs in 
southeastern Polynesia. The mutations of the h in the second syl- 
lable run a characteristic path with but few anomalies; 24 poke is 
admissible only as a kappation upon a pote base suggested by 26-27. 
The vowel skeleton is chiefly o—-e with variation to o-i in 14, 23, 29, 
32. The former vowel is found as i in 3 fis, which appears a satisfac- 
tory identification, as e in 7, 20-22. In 22 hev we find a metathesis 
upon a vehi which does not appear in the record, but which may prop- 
erly be interpolated from 20 vesi. ‘The items 41-62 seem to belong 
to a stem in 0-0, which is strangely parallel with the fohe stem. In my 
estimate of the tendency of these languages to preserve their strength 
in the vowel structure I incline to distrust the likelihood of so general 
a variation from fohe to a putative foho; yet the two forms-as naming 
the same physical object interlace all through Melanesia, and it is 
possible that they have a common source, and in the examination 
of this stem we are led to the acceptance of forms having a—o vowel 
structure, o—a and a—a, through which we are led to 62 wes. ‘The items 
63-65 establish yet another stem.* It might be derivable from fohe, 
except for the fact that the h-r mutation nowhere appears in a position 
to be confirmed. In items 66-70 we have another stem from the New 
Guinea region which we can readily establish within its own limits, 
but which it is extremely unlikely has any association with the fohe 
stem. The items 71-73 seem scarcely associable with one another 
and not at all with the stem principally under this examination. 
