J5 U M A T BL A. S7 
the tid«5 Irom eafl to weft, this coaft ovjght to receive a contmuai ^ccef- 
fion, proportioned to die iofs which others, expofed to the direction of 
this motion, muft and do fuftaln ; and it is likely that it does gain upop 
the whole. But the nature of my work obliges me to be more attentive 
to effects than caufes, and to record fa^ls, though thcy .fliould d,afii \vith 
fyftcms the nioft juft in theory, and moft refpedabie in point of au- 
thority. 
The chain of iflands which lie parailei witli the weft coaft of Swrnatra, Iflan^^s near 
^ ^ the weft coaJl, 
may probably have once formed a part of the main, and been fepcrated probably ones 
from it, either by fome violent effort of nature, or the gradual atrrition 
pf the Tea, I would fcarcely introduce the mention of this apparently 
yague furmifej but that a circun^ftance prefents itfclf on the coaft, which 
affords fome fl-iw^r colour of -proof than can be ufually obtained in fuch 
inftances. In many places, and particularly about tally and La^e^ we 
©bferve detached pieces of land ftanding fmgly, as iflands, at the dift- 
ance, rof one or two hundred yards from the flaore, which were head- 
lands of points running out into the fea, within the remembrance of the 
inhabitants. The tops Qontinue covered with trees or ihrqbs; but 
the fides are bare, abrupt and perpendicular* The progrefs of in- 
fulation here . is obvious and iocontrovertable, and why pay not larger 
iflands, at a greater diftance, have been formed, in the revolution of 
ages, by the fame accidents ? The probability is heightened by the diE;ec- 
tion of the iflands, Mantcmayc^ Mtgo^ Sec. the fmillarity of foil 
and proi3u€^lon<;^ and the regularity of foqndings between them and the 
main; whiift without them the aepm u.ir«a-.^«rt.^t»u,- 
Where the fcore is flat or ftielving, the coaft of Sumatra, as of all comlRo^ks/ 
other tropical iftands, is defended fr9m, the attacks of the fea^ by a reef 
or ledge of coral rock, on which the furfs exert their violence withoyt 
further effect than that of keeping its furface even, and reducii^ to pow- 
der thofe beautiful excrefccnces and ramifications which have been fo 
much the objed of the iiaturalift*s curiofity, and which fome ingenious 
men, who have analyfed them, contend to be the work of infers. The 
coral 
