70 THE SUBANU. 
taran, from the stem gatad; and tobora from the stem tobod. I have 
already directed attention upon the fact that rude speakers give par- 
ticular force to the beginnings and endings of words; this weakening 
mutation is in itself confirmatory. In the three following instances we 
shall observe the d-r mutation applied to the initial d when it is buried 
under a formative prefix: maralag, from stem dalag; corala, from stem 
dala; marope, from stem dope. 
In the sole instance of batasan from stem batad, we find a mutation 
that stops midway: instead of upon the semivowel, the weakening mute 
rests at the sibilant. We lack data for the determination whether the 
variants boid and buis represent the same halfway mutation or whether 
this is an example of an s—d mutation not elsewhere discovered. 
If this inversion of mutation movement remains in doubt in the 
matter of S-d, we may regard it as definitely settled in the case of r-d in 
two instances: the variants danao and lanao and dongog and rongog. 
The proof is external to Subanu, external in fact to all Indonesian speech, 
yet we are justified in advancing it from its proper later place in order 
to settle this matter of phonetics. The parent of rongog is certainly 
the Polynesian longo (rongo) ; thus it is clear that this is really a case of 
r—d mutation. ‘The other word, lanao, is not quite so clear; it appears 
associable with two Polynesian words, /ano sweet water and Janu a lake, 
these two being probably a divaricated stem. If this be indeed the 
source of Janao, the r-d mutation, already once established, receives 
confirmation. Inthe Bontoc Igorot we find tj@num, danum, denom, all 
in the sense of potable water. ‘This series, so fortunately preserved, 
gives us both Jano and lanu derivatives in the signification of sweet 
water, and goes far toward establishing the original unity of the divari- 
cated vocables. 
The uncertainty which I have manifested in the discussion of 
lanao rests not only upon the diversity of sense but rather more upon 
the lack of acquaintance with the phenomenon of ao employed in 
diphthong value as representative of an earlier source vowel. ‘The 
collation of the Bontoc Igorot upon which I engaged after the foregoing 
note had been written has given a satisfactory suite of instances in 
which this dipthong appears as the mutation product of the vowel 0. 
\| | 
Subanu. Bontoc. | Subanu, Bontoc. | Subanu. Bontoc. 
SSS 
boligan faolengan | goloan olacan | tao takao 
gayo kayao } linao alinoao ! toon taaowin 
This does not exhaust the source of this dipthong. From a we 
have balin-gawa—kaaowa, from e goyamet—komaot, from u pusu—baosig. 
This discovery applies particularly to the etymology of lanao. 
We see here abundant evidence that lano and Janao are homogenetic; we 
have an item of confirmation of a possible lanu—lanao association. 
