42 THE SAKAI AND SEMANG DIALECTS. 
Sakais or Semangs have in all cases correctly described them. 
He therefore tests these statements by the locality, physical pe- 
culiarities, etc., of the tribes in question. He points out that 
Semangs do not exist in the southern part of the peninsula and 
quotes R. Martin who gives as their country northern Perak, Ke- 
dah, Rahman, Rangan, and Kelantan, a description with which 
Stevens agrees. He further notes that the Semang use or have 
used the bow, and that there is no record of the Sakais having 
done so. He concludes that the Semangs in his first group are 
correctly described but that de Morgan’s ‘Séman’ and the ‘Perak 
Semangs, and ‘Kenning Semangs’ mentioned in fifth volume of 
the J. S. B. R. A. S. may very possibly have been Sakais or 
at all events mixed races. The Sakai who form his second 
group fall linguistically into two sub-classes the divisions be- 
tween which seem to be confirmed geographically by Clifford’s 
line from Blanja on the Perak River to the Bidor Mountains 
and thence to Kuala Angin in Kelantan to the north of which 
line Clifford found his T'em-be to the south his Sen-oi. He 
concludes therefore that the Semang and Sakai form two differ- 
ent branches of one language and that the Sakai branch shows 
two sub-branches. 
The second part is headed ‘comparison of the Sakai and 
Semang languages’ and opens with a list of books consulted by 
the author in his study of the latter. Then follows a list of 
those Mon-Khmer words and roots which are found to be simi- 
lar to words and roots in Sakai and Semang. The author’s 
comments on this areas follows :—‘* The above agreements seems 
‘to me to be amply sufficient both in number and kind to nega- 
“tive the suggestion of ‘A mere external borrowing.’ As to the 
‘ their number out of the 1249 forms contained in the vocabulary 
‘there are about 240 such agreements. That is in itself a notable 
‘result but it gains in meaning when two things are borne in 
“mind :—First that most undoubtedly a part at least of the 
‘materials for the Sakai and Semang languages are recorded 
‘with a wrong or uncertain meaning thus rendering it difficult 
‘or even impossible to find their correct equivalents in Mon- 
‘‘Khmer, and secondly that another part,—more specially that 
“collected by de Morgan and Stevens, is of such a nature — 
‘(names of implements and individual parts of them, of individ- 
Jour. Straits Branch 
