48 SISSANO. 
existence has not been firmly established in Melanesia (The Poly- 
nesian Wanderings, 302); the most that we may venture is to propose 
this identification as not without some ground. ‘The t-k which has 
had such a compelling sweep over Polynesia is less marked in the 
Melanesian tongues, yet its existence has been satisfactorily estab- 
lished. So far, then, as relates to this unit of variety in items 52-57, 
we need have no doubt, although in other particulars these are the 
most obscure of all the fatu derivatives. The extinction of t, of 
equally compelling sweep in Polynesia as is its kappation, is less 
frequent in Melanesia, but it has been proved in a sufficient number 
of instances; this enables us to include items 45—51 in the series without 
hesitation. The final vowel is frequently abraded after the common 
custom of many of these languages in order to produce the closed 
stem which seems more cordial to Melanesian enunciation. We find 
the mutation u—o in 11 and 33, u-e in 32, for both of which there is 
abundant confirmation. At the end of the list (52-59) I have set a 
small group of forms which are in varying degrees doubtful. In 57 
pak we shall have little hesitation in hanging the form by kappation 
on 34 pat, concerning whose affiliation with fatu there can be no doubt. 
In the group of consimilars 54-56 I am sure that we are justified in 
finding a valid derivation; omitting the added syllable in vaku-na we 
shall find in kappation a good derivation from vatu and this once 
established 55 and 56 swing into line. The Kiriwina forms 52 daku and 
53 gaku are very difficult. If53 be established as a fatu derivative it is 
apparent that 58 and 59 follow as easy corollaries. Although in the 
‘Polynesian Wanderings’ I have noted a few and very doubtful 
instances in which the mutation f-d, f-g slightly suggests itself, I can 
not accept it. If we were to assume for these forms a mutant batu 
stem, we might be led aside to consider a progression of mutes, a 
principle well established in other language families; but in these 
isolating languages such a progression is not to be considered for a 
moment, for mutation here is narrowly restricted to the possibilities 
of each speech organ and does not pass to any other, save for the very 
modern movement of interserial mutation in the nasals and the mutes 
which has not yet been established beyond the palatals and the linguals. 
18a. bei flying fox. 
REFERENCE: Deutsch-Neuguinea, 201: 19. 
POLYNESIAN. 
peka Futuna, Niué, Rarotonga, Ma- | pe‘a Samoa. 
ngaia. opeapea Hawaii. 
pekapeka Maori. beka Sikaiana, Viti, Tonga. 
MELANESIAN. 
1. béka Kowamerara, Namatanai, 5. bianga Barriai. 
Namarodu, Lalinau, Bi- 6. bék Nokon. 
' sapu, Punam. 7. mimbék ‘Tatau. 
2. mimbéka Kowamerara. 8. ambiak = Lihir. 
3. béga Vitu. 9. bea Wuvulu, Aua. 
4. bege Belik. Io. béi Sér, Sissano. 
