S86 SISSANO. 
either through the ng—g mutation, or else in Ambrym, as in so many 
cases in the Pacific g is employed to represent ng. A liquid preface 
gives an interesting series both for isu and for usu; in 20 lisui we 
find it employed upon isu with a terminal addition, in 30 lusu we 
find it employed with usu; each of these secondary compactions 
undergoes further distinction with an element which is found as ba 
in 21 and 31, as pain 32, and in 33 is affected by the Melanesian dif- 
ficulty in enunciation of the labials. We find another preface syllable 
kiin 17 kinihu. A terminal syllable makes its appearance in 5 ishuda. 
Final consonants occur. Sissano suk suggests the inclusion of 37 soku 
in the derivation chain; final ng is found in 16 and 23, final n in 4, 
final m in 24. The Efaté 34 ngore introduces us to a small and 
interesting group. Iam not prepared to pass upon the affiliation of 
this group with the su stem; yet if we start with 27 ngusu, also 
from Efaté, we find in 26 osu warrant for the o and in 8 iru warrant 
for the r. It is, however, quite clear that 35 and 36 are associable 
with ngore and perhaps the same is true of 38. 
94. taméng woman. 
REFERENCES: Melanesische Wanderstrasse, 145. Deutsch-Neuguinea, 216. 
1. tamén ‘Tumleo. 3. tamine Kobe. 
2. taming Paup, Yakomul. 
Friederici includes these forms within the series of fine woman, 
but I am unable to establish the chain of affiliation through which 
such association might be settled. This stem recurs with slight pho- 
netic variety in 54 lon-tamin sister and in 119 wun-damin wife. 
97. tapo crocodile. 
REFERENCES: Melanesische Wanderstrasse, 112¢. 
tapt Sér. 
Friederici records these two forms after discussion of the widely 
disseminated bua crocodile, and the passage deserves citation in full. 
Zu erwahnen sind dann noch zwei Worte, die wieder zeigen, dass ein gewisser 
Zusammenhang zwischen den Melanesiern des Westteils der Nordkiiste von 
Kaiser-Wilhelmsland und der Gegend der Tabar-Inseln und von Limba und 
Langanie (Neu-Mecklenburg) besteht. Denn wir haben in: Sér: tapi; 
Sissano: tap6; Tumleo: alé6; Paup: alié; Yakomul: alii; Kowamerara 
(Tabar-Ins.): mi-l6wa; Tatua (ebenda): mi-képo; Limba und Langanie: 
lava. 
In the material here assembled I find myself unable to trace the 
interassociation which Friederici postulates. No more is it possible 
to adjust even so congruent a pair of vocables as Sér-Sissano with the 
general bua type. 
