18 



rock is one rudely rounded mass and apparently composed of concentric layers. Towards one 

 tide a wide chasm penetmlcs it teimDjj at the eastern an^Ie a columnar portion ii- 

 rided by a lioritonial fissure into (wo rounded blocks of which the upper is MitJ at the 

 top connected 1 1 li the mass. The neck of junction coii si a is of thu rcmnarils of numerous 

 layers wedded iq as it were, one half wilh their euiiYci sides to the spectator or embra- 

 cing the rock above and the other half Willi their concave siilcs toward!* [lie spectator or 

 capping the Lloek beneath , thus presenting (lie appearafics which ft number or somewhat 

 stiff sheet* of paper or oilier flexible substance do if I hey pre compressed lightly in ihc 

 middle and their em to made to diTergc on bo!h sides. 



Beyond Pith Tarn a point occur* from whkh a large furrowed rock advance* into iho 

 in. Tlii? I inspected with some minuteness, The channels, c*en on the same fac«, nin 

 in different directions so that they sometime* ergs* ; but p mere frequently r niter tuecliu;; > 

 one only is continued- all eaies thry evidently occupy the line* of division or imperfect 

 cohesion, On tfio side facing the NW. the direction of the larger ones u nearly NE. 

 and NW, On the S. side f hey arc deep and face the S. These latter it appeared 1o me 

 had been gradually cicavaEed hy the allcrualinrr nut ion of the sea and the atmosphere. A 

 little further to the E. the gutters of ihc rocks faced P3E. by Pff. nearly; aE one place 

 the rock u nearly worn through and the breach js huU fdlcd with hirpc angular frag- 

 ments j the remnants of the layer or ledge which had Orjjji natty occupied the eavily, lis 

 tides Pire parallel ond mark IW0 planes of division. The. furrows bey mid tint were gene* 

 rally in one or other of two directions according to the slope p that is either facing; the 

 KE. by ft + or W?iW. nearly. The general surface at some place? is uneven, which oc- 

 casions varying dopes 7 and it appeared clear that the direction* of the furrows ut these 

 places were fully captained lay the directions in which the slopes would cause I he rain 

 currents to run down them. In many ease*, the latter did not exactly coincide with the 

 former j because no fissures or lines of division did. But of the two systems of parallel 

 divisional lines by which the surface was intersected , I hot whieli most nearly nrrrM wilh 

 the slope had {jivcu their di tree [ions fo Hie flutters Where the face of the rock was ■lightly 

 hollowed the gutters converged. The ridges between ihc dlTisiipnnl lines are sometimes 

 crossed by gutters but only where this would he rendered necessary for r|jo descent or the 

 waler. 



It was not until 1 had examined both the eastern and western portion* of the Island that 

 I was able to revisit the point where my acquaintance with it bad commenced. It proved 

 that I had seen only one small section on the S r side, and that the SE- and E, laces 

 possessed the principal rocky These stand in great profusion along ihc heacb or ri*e from 

 the water in front, while the hill behind appears like some ancient it castled steep" with 

 remnants of flanking wall* midway up, and broken La I Hero cuts frowning from the summit. 

 Many of ihe P''e* ore on a grander scale than any that are found elsewhere; they have a pea- 

 lex air of andnjnityi and rising as they do in diversified furrns, — nera in solid cubical masses , 

 there traversed by deep chasm* and bristling with shnrp pinnacles, at one place standing 

 out in full relief in their prey mossy coating, and at another catered with a trellis work of 

 roots, trees ascendinjj from llicir summit mfo midair r and the entire rock buried under 



