POLYNESIAN AND MALAYAN. aol 
Visayan material I express myself more positively in support of the 
position which in my earlier work seemed less clear and that further 
support appears in the Igorot. 
48. malemos to drown; Visayan lomos id. 
lemohaki Tonga. maremo_ Rarotonga. 
ndromu Viti. palemo Hawaii. 
emu Rapanui. paremo ‘Tahiti, Maori. 
malemo Samoa, Futuna. peremo Mangareva. 
melemo ‘Tonga. parego Paumotu. 
The final s is found in the Visayan and in normal mutation in the 
Tongan lemohaki. We have in Polynesia three forms of the simple 
stem, two (lomu, lemo) in Nuclear Polynesia and one (emu) in Rapanui, 
as an interesting article of the proof of a settlement upon that remote 
island of a Proto-Samoan colony; even so recently as my recension of 
the dictionary of that speech this item escaped my attention inasmuch 
as I had not then the Visayan evidence. ‘The remaining forms in 
Polynesia are conditional with ma and with a variant pa; pa is found 
in languages of the Tongafiti settlement, ma is Proto-Samoan, except 
that Rarotonga is a Tongafiti community but has the conditional prefix 
of the earlier stock. In Paumotu the m-g mutation, a shift across the 
utmost nasal range, is not unknown in other instances. 
49. Mama to chew; Subanu mama id. P. W. 280. 
mama Samoa, Tonga, Futuna, Niue, Ra- sss 
panui, Marquesas, Manga- | mamah Malay. 
reva, Hawaii, Viti. mama Subanu. 
manga Nukuoro. mangaga Bontoc Igorot. 
maanga Uvea. 
Except for the mutation in Nukuoro and Uvea this identification is 
so complete as to be featureless. We note the almost complete absence 
of the stem from the Melanesian traverse, its only appearance being in 
Aneityum a—mai. Insecondary and derivative forms in Nuclear Poly- 
nesia we encounter the form maga, which passes before the superficial 
judgment as of the common type of verbal noun formed from stem ma 
by the usual suffix —ga, the secondary sense denoting either an act of 
chewing or the person who chews. If this were the true explanation 
of maga we should find ourselves under the necessity of arguing that 
in Nukuoro and Uvea the verbal noun, after it had been created pur- 
posely to express a distinction for which the language had felt a need, 
sacrificed that distinction and took the place of the primal verb from 
which it is derived. ‘This runs contrary to the grammatical course of 
the speech. The discovery in the Bontoc Igorot of mangaga establishes 
the existence in the earliest type of the Polynesian of a verb radical 
manga and authorizes us in classifying the Nukuoro and Uvea forms 
as Proto-Samoan. Elsewhere in Nuclear Polynesia, in regions to which 
the later Tongafiti swarm found readier access and where its domina- 
tion was better established, the abraded stem ma of that phase of the 
common tongue came into use in its duplicated form. ‘The particular 
