None such are found toward Sungei-Lamo and the rest of the shore northward of Marlborough Point, where, on the contrary, you perceive the effects of continual depredations by the ocean.  The old forest trees are there yearly undermined and, falling, obstruct the traveller; whilst about Pulo the arau-trees are continually springing up faster than they can be cut down or otherwise destroyed.  Nature will not readily be forced from her course.  The last time I visited that part there was a beautiful rising grove of these trees, establishing a possession in their proper soil.  The country, as well immediately here about as to a considerable distance inland, is an entire bed of sand without any mixture of clay or mould, which I know to have been in vain sought for many miles up the neighbouring rivers.  To the northward of Padang there is a plain which has evidently been, in former times, a bay.  Traces of a shelving beach are there distinguishable at the distance of one hundred and fifty yards from the present boundary of the sea.

But upon what hypothesis can it be accounted for that the sea should commit depredations on the northern coast, of which there are the most evident tokens as high up at least as Ipu, and probably to Indrapura, where the shelter of the neighbouring islands may put a stop to them, and that it should restore the land to the southward in the manner I have described?  I am aware that according to the general motion of the tides from east to west this coast ought to receive a continual accession proportioned to the loss which others, exposed to the direction of this motion, must and do sustain; and it is likely that it does gain upon the whole.  But the nature of my work obliges me to be more attentive to effects than causes, and to record facts though they should clash with systems the most just in theory, and most respectable in point of authority.

Islands near the west coast.

The chain of islands which lie parallel with the west coast of Sumatra may probably have once formed a part of the main and been separated from it, either by some violent effort of nature, or the gradual attrition of the sea.  I should scarcely introduce the mention of this apparently vague surmise but that a circumstance presents itself on the coast which affords some stronger colour of proof than can be usually obtained in such instances.  In many places, and particularly about Pally, we observe detached pieces of land standing singly, as islands, at the distance of one or two hundred yards from the shore, which were headlands of points running out into the sea within the remembrance of the inhabitants.  The tops continue covered with trees or shrubs; but the sides are bare, abrupt, and perpendicular.  The progress of insulation here is obvious and incontrovertible, and why may not larger islands, at a greater distance, have been formed in the revolution of ages by the same accidents?  The probability is heightened by the direction of the islands Nias, Batu, Mantawei, Pagi, Mego, etc., the similarity of the rock, soil, and productions, and the regularity of soundings between them and the main, whilst without them the depth is unfathomable.