GEOGRAPHY OF THE MIGRATIONS. 175 
It is difficult to excavate below the present culture of the Malayan 
Archipelago with much hope of finding out what may have been the 
extent of Proto-Polynesian occupation of the islands before the coming 
of the invader. A careful examination of the speech and of the custom 
life of a few scattered peoples within this area, such as the Aeta of the 
Philippines, the Punam of Borneo, the Barée-Toradja of Celebes, 
leads in the direction of the opinion that during the period of undis- 
turbed Proto-Polynesian race occupation there were areas of peoples 
yet more primitive. But it seems fair to assume that the Proto- 
Polynesian occupation of Indonesia was practically uniform and 
general. Accordingly the lines on Chart XVII differ in quality from 
those which we have drawn on the Melanesian chart. Here we have 
a consolidation of all the lines connecting the loca of affiliation recorded 
in the chapter next preceding. Yet they do not present the character 
of lines of folk movement as do those in Melanesia; we do not sense 
upon them the feeling of destination. The motion in this region, the 
lines of conquest and occupation of Indonesia, neither of these per- 
tains to its Proto-Polynesian element; it is the proper function of the 
student of the Malayan peoples. Until the moment of escape from 
a lost field of battle the Polynesian here is sessile; mobility is the 
character of the Malayan. Yet certain of the exterior lines of affilia- 
tion, while they do not impress us as so instinct with motion as do 
the Melanesian traces, do suggest lines which might be traveled. At 
the extreme west we find the ultimate outposts of the Polynesian race, 
not great in extent, yet sufficing to establish a line at least of Indo- 
nesian migration, one determined by Silong on the west of the Malakka 
Peninsula, the other better determined by Nicobar and the significant 
status of the Mentawei Islands west of Sumatra. From Sumatra we 
have three lines of affiliation which are very strongly marked. One 
leads through Borneo in the direction of the Philippines, the other 
into Java and the southern chain of islands, the third toward Mada- 
gascar. Almost all our affiliations of the Malagasy, of whatever 
linguistic stock, point with great uniformity to Sumatra-Java and 
very little to other parts of the archipelago. We regard this migra- 
tion to Madagascar as comparatively late, dating from a period when 
the people of Sumatra-Java had assumed whatever Polynesian ele- 
ment characterizes their respective languages. We are not in any 
sort justified in the theory, which was held as recently as.the time of 
De Quatrefages, of Polynesian occupation of that remote island of the 
African continental mass. 
Krom Java we trace eastward two lines. One may smoothly be 
drawn along the scarcely interrupted chain of southern islands as far 
as Timor Laut and Aru, where it approximates that Viti stream of 
migration which the study of the Melanesian material has led me to 
propose and earnestly to support. From Madura we lay out a trace 
