254 Transactions of the Royal Society of South Africa. 
them contain a greater variety of minerals derived from igneous rocks, 
while at some of the places mentioned there are fissures filled with 
igneous rocks. 
The Kobe Eiver neck is the only one of these occurrences in w^hich 
minerals of probable igneous origin have not been found. But in each 
case the material from non-volcanic sources preponderate amongst the 
recognizable constituents of the rocks. 
The occurrence of barytes at Geitsi Gubib is another point of 
resemblance to Saltpetre Kop, though the mmeral seems to be more 
abundant at the latter place. 
There are frequently slickensided surfaces in the tuffs produced by the 
local slipping of parts of the rocks after they were consolidated. The 
hardening of the rocks by deposition of silica is most marked in the 
peripheral part of the pipe ; and this fact is certainly responsible for the 
ring shape of the mountain. The almost level floor of the central 
depression comes to an abrupt end where the stream-bed passes over the 
upper thick band of hard tuff which forms the higher of the two water- 
falls in the valley ; this strong bed can be followed easily up the mountain 
to the crest of the ridge on either side of the valley. The lower waterfall 
is caused by the presence of a second thick band of hardened tuff. These 
beds are fully exposed across the valley, and their unfractured condition 
proves that the valley owes its form to erosion alone and not to any 
radial crack through the wall of the neck. 
The hardening due to the deposition of silica extends to the Fish River 
beds round the neck, for these are frequently traversed by small quartz 
veins and are more thoroughly quartzitic in the immediate neighbourhood 
than they are further away. 
On the plan (p. 250) six small kopjes and three dykes are inserted. I 
only visited the two kopjes and the dyke south of the main mountain, the 
others were sketched in from the northern ridge of the mountain ; they 
are conspicuous objects, and are almost certainly of the same nature as 
the southern ones. 
The dyke G is a yellow ferruginous and calcareous rock, generally with 
fewer fragments in it than the breccias of the mountain contain. It is 
similar in general appearance to the tuff dykes of the Saltpetre Kop group. 
The kopje A is about 100 feet high and some 400 feet in diameter at the 
base ; the contact of the breccia with the Fish River beds is hidden under 
debris. Kopje F is rather smaller and not more than 50 feet high, and 
the contact is again concealed. The breccias of these satellite pipes are 
much coarser than any seen in the mountain, large blocks of the Fish 
River sandstone and shales are the most abundant fragments, but whitish 
quartzites and granites are frequent and there is much fine-grained grey 
rock with the peripheral zones bleached. There are apparently no rounded 
