§2 SISSANO. 
In the study of this suite of forms we are under the necessity of 
examining’ at the outset the effect of metathesis, a phonetic prin- 
ciple peculiarly frequent in this stem. This principle needs close 
study before we can assume to comprehend the phonetic motive of 
its application, but its mechanics are simple. In an earlier work 
(Easter Island, page 26) I have made the beginning of an examination 
of the theme, and at present I am not justified in advancing beyond 
the results there set forth. ‘The most that has been done is to devise a 
system for the record, and eventually for the comparative study, of 
vocables in which metathesis is discoverable. In that examination 
sufficient material was available to establish the fact that metathesis 
can be consonantal or vocalic. The device for establishing the types 
of metathesis is of extreme simplicity in the languages of open struc- 
ture and may be made to serve for the closed syllables also. In this 
scheme A is employed to designate the vowel of the first syllable. 
whatever it may be, E the vowel of the second, I of the third, and so 
on; B is employed to designate the initial consonant of the first syllable, 
c of the second, and so on. 
In the first group of metatheses which we encounter in this suite 
we find the change applied to the stem lanu. This will be repre- 
sented in the record scheme by the type BACE. ‘The metathetic form 
is of the simplest and numerically most frequent type represented by 
the type CABE, in which the vowels of the stem remain fixed, the 
initial consonants of the syllables interchange places. ‘Thus we find 10 
nalu, by a slight and familiar liquid mutation 9 naru and by a scarcely 
less frequent mutation 11 nanu. Friederici notes the possibility, it 
seems to him highly probable, that these forms are derivative from 
ngalu, the common Polynesian word for wave and particularly breaker; 
to me it seems far more reasonable to regard these as metatheses 
upon lanu. The second group consists of variants upon the closed 
stem danum. In my earlier examination of the principle it was not 
necessary to take into consideration the forms of closed syllables, 
for that work was concerned solely with the open Polynesian languages. 
The closed syllable may readily be represented by the employment 
of a typographically varied symbol of the same order as the initial 
consonant; thus meto is of the type BACE, and mento would become 
BADCE and meton would become BACEc. ‘The closed stem danum is 
of the type BACEc; by metathesis to the type BAcEc we find 3 démtn; 
from stem 2 dénim, subjected to a simple vocalic mutation, we find 4 
danim; and from a stem ranum, which we may properly interpolate 
on the strength of the occurrence of 6 ranu, we find 8 rémin. Pro- 
visionally we observe that metathesis in the closed syllable is accom- 
plished through the interchange of the former with the latter conso- 
nant. ‘This is provisional, adopted only for mechanical convenience. 
