21 feet deep and about 2 feet broad. One groove I observed about six feet deep and two feet 

 broad with small secondary or inner grooves fluting its surface. While examining this rock 

 a hea^y shower of rain began to fall, and as my time was cxhausted I was obliged to leave 

 before I could mak e more precise observations or any measurements, and, in truth, before I 

 had recovered from my first sensation of wonder. It appeared to me that the rock must have 

 been split on being elevated from a lower level. On returning I observed many smaller rocks 

 near the beach with channeled sides. On the top of one of these there was a long ] deep 

 trough with small grooves converging into its upper end, like the ribs of a fan. The rest 

 of the surface was covered with slight depressions. 



I believe this is the first time that grooved rocks have been observed so close upon the 

 Equator. Their absence has been considered an argument in favour of the glacial theory of 

 the bmlder formation. None of the channels or grooves, however, which I observed resem- 

 ble the parallel inclined or approximately horizontal furrows which are caused by the mo- 

 tion of glaciers in descending the rocky trough of a valley. But they appear to correspond 

 strikingly, save in being vertical, with the giant cauldrons, passing into long deep grooves, 

 which are described by Agassiz as being produced in the Alps and Jura by streams of water 

 falling over the sides of chasms in advancing glaciers, and acting as a locomotive erosive 

 force upon the subjacent rocks. My hurried and restricted observations hardly [warrant a 

 conjecture as to the probable origin of the Pulo TJhin grooves. The idea that occurred to 

 me on the spot was, that the several rocks, before they were shattered and separated by the 

 force which placed them in their present positions, and in some different local distribulion 

 of land and sea from that which now prevails , had formed the site of a cascade of no great 

 force which had gradually worn the sides of the rocks into channels. A succession of falls 

 would account for the relative positions of the rocks with respect to each other and for the spoon 

 shaped hollows on the surfaces of some of them. It appeared to me that ordinary meteoric 

 erosion and decomposition were totally inadequate to explain the shapes and size of the 

 grooves. In many places they are overgrown with mosses, and in some, if not in all, they 

 are prolonged beneath the ground, and thus protected by the soil of the hill, which must 

 have covered them for a considerable period, since large trees are rooted in it. The aspect 

 of the rocks is not such as rapidly disintegrating granite wears, but, on the contrary, re- 

 sembles that of an ancient building. I could find no tracé of any fissures coinciding with 

 the direction of the furrows. Yet there can be little doubt that, to whatever agency they 

 may be referred , the grooves were first opened along lines where the cohesion of the granite 

 was comparatively weak. The regularity with which the projecting columns of the rock first 

 noticed are scooped round at two places across the direction of the grooves, seems to prove 

 that the granite has an internal arrangement similar to that so frequently observed in this 

 rock, and which causes it to be shattered into blocks more or less cubical. In one of the 

 lower rocks which the Chinese are quarrying we found two parallel vertical veins traversing 

 the en lire rock so as to include between them a plate about an inch in thickness. One side 

 of this plate sparkled with metallic grains of a golden hue (iron pyrites). The other was 

 covered with a rusty stain resulting probably from the fissure on that side ha ving been per- 

 meable by the air and the consequent decomposition of the grains. 





