INDONESIAN ANNOTATIONS ON THE VOCABULARY. C17 
currency. On the charts we can follow out these lines with interest. 
We have already noted that tama in its father sense (Subanu, 144) 
is wholly Proto-Samoan; we associate with this fact the interesting 
circumstance that the only instances of the lingually prefaced tama 
in Indonesia derive from the Klemantan of Borneo and the Toradja 
of Celebes, and that of the latter great island an outlier is Salayer, 
an island name carried over leagues of sea by the Samoans and pre- 
served to the present day in the honorific salutation of Tutuila as 
the Island of Salaia. 
It seems, therefore, wholly preferable to regard ama as the primitive 
stem, primordial so far as our explorations can be prosecuted into 
this beginning of a world speech, and that the present variety shows 
us the operation of a great law of particularization of sense determin- 
ing specification in form. 
45. kaluk wooden pillow. 
32. kalang ulu Malay, Lubu. 39. olonan Tagal. 
33. karang ulu Java, Buru. 40. ulunan Magindano. 
34. halang ulu Batak. 41. gulunan Bagobo. 
35. Olaoan Bontoc Igorot. 42. galuling Dyak. 
36. goloan Subanu. 43. karaluni  Tettum. 
37. olon Tagal. 44. kluni Tettum. 
38. hélén Dyak. 45. iklunin Galoli. 
The three forms which head this Indonesian list explain themselves 
very simply. The Malay kalang signifies support and ulu is common 
to Indonesia and Polynesia in the sense of head. ‘The series 35-41, 
all save 38 of Philippine provenience, shows a little accented develop- 
ment from ulu head. In the Indonesian affiliates of this stem (Subanu, 
149) we find the presence of a final n in Wayapo, Massaratty, Teor, 
and Tobo; of the similar m in Kayeli. This may indicate that in 37 
olon the Tagal makes a distinction between the head support with the 
final n of an older form and the head itself as ulo equivalent of olo in 
the more modern form. ‘This may also be true of the Dyak 38 hélén. 
On this reading of the material we should identify olon as head and set 
apart -an rather than -nan as formative element in 39-41, all forms 
from the Philippines. This opinion as to the formative element then 
receives support from 35-36 in which -an is affixed to an olo represen- 
tative. The assumption of a palatal preface g in 36 and 41, both Phil- 
ippine and of an early type, is of great interest when we put these 
occurrences into geographical relation with the similarly prefaced 
forms in Melanesia 1-10, all being on that shore of New Guinea and 
the adjacent Bismarck Archipelago which is most ready of canoe access 
from the southern Philippines. The forms 42-45 exhibit an elusive 
similarity with the ulunan type; 42 and 43 clearly form a pair, 44 and 45 
another, and 43-44 show the transition; but when we orient the ulun 
stem upon the best-developed phase of this type, 
karaluni 
ulun 
