chap. i GEOGRAPHICAL DISTRIBUTION
As a matter of fact, the relations which exist between these different types are exceedingly complex, and cannot be satisfactorily discussed without referring to the other families of speech in Southeastern Asia with which the dialects of the Peninsula are connected, or by which they have been modified. eut before entering into the consideration of these difficult problems, it is desirable to explain the geographical distribution of the dialects, and to indicate at the same time the subdivisions into which they fell; for these subdivisions, though based on linguistic data, do in fact agree to a considerable extent with the territorial arrangement and geographical relations Af the several tribes.
Roughly speaking, then, the dialects fall into groups 'vhich correspond, though not accurately, with the anthropological varieties of the aboriginal races. In the north of the Peninsula are the Semang dialects (called on the eastern side of the main mountain range by the name of Pangan); in the centre the Sakai; and in the south the very mixed and broken-down dialects which are here grouped as Jakun.1 Semang.aThe Semang dialects, including those the Pangan tribes, are spoken in an irregular tract extending from at least as far north as lat. 6A 30'2 to ^bout lat. 50 5' on the western side of the Perak river, and about lat. 40 45' on the eastern side of the main range in the States of Kelantan and Trengganu; and lying between longs. iooA 40' (though a century ago
1 A reference to the map here given outside these lines, while within them
illustrate the relative position of some of the recorded dialects may have
ese groups. The boundary lines become extinct.
0n map mereiy indicate 2 Except where otherwise stated,
Aughly their limits as evidenced by all latitudes are north. AH longi-
actual data of dialects recorded. tudes are east of the Greenwich
ls quite possible that others exist meridian.ls quite possible that others exist meridian.