311 
of the Rocks of Shetland, 
which I shall, on another occasion, point out as existing in a 
vein of granite, that traverses a mass of gneiss ; whilst in this 
mass no mechanical appearances are elsewhere to be observed. 
It is easy, then, to conceive, that such a form of globular con- 
cretions would, to rocks thus originally constituted, if, at the same 
time, we summon to our assistance the force of aqueous elements, 
give every liability to abrasion. But the presence of such aqueous 
elements literally affords a very navigable ocean for conjecture. 
I therefore dismiss this protracted discussion, convinced of the 
multiplicity of circumstances which favour the notion, that the 
conglomerated appearance is an original peculiarity of structure ; 
but, at the same time, aware of the difficulty of forming an ac- 
curate judgment respecting any manifestations which might indi- 
cate, that there has been a subsequent modification from me- 
chanical causes. 
The Strata forming the Cliff Hills. 
I shall now proceed to describe the rocks situated to the north 
of the Epidotic Sienite of Fitful Flead. These consist for the 
most part of clay-slate, associated with thin strata of quartz 
and hornblende-slate. The clay-slate forming the long range 
of the Cliff Hills, runs from near Spiggie and the Loch of Lu- 
nabister in Dunrossness, through a long tract of country to 
Quarf, where there is an interruption of the ridge, and thence to 
KibisteFs Point, on the east of Dale’s Voe. The whole of the 
strata of clay-slate, quartz, and hornblende-slate, seem connect- 
ed unconformably with the epidotic sienite. The junction is 
much concealed by the Loch of Lunabister, and by covered 
ground ; but the inference that it thus takes place, is from the 
visible proximity of both rocks to this site, from the direction of 
the clay-slate, which, tracing it from the north, is S. 15° W. ; 
whilst the northern boundary of the epidotic sienite, observable 
from its most southerly point of Cross Holm in Quendal Bay to 
the west of Dunrossness, is traced in a line of N. ^5*^ W. Thus 
the junction must take place at an angle, or by the lateral 
edges of the strata being opposed to the northerly surface of the 
sienitic mass. 
The clay-slate, quartz, and hornblende-slate, are in their po- 
sition unconformable to each other. The hornblende-slate is 
