256 Transactions of tJie Boyal Society of South Africa. 
than it. Another possible explanation of the structure of Geitsi Gubib is 
that we have a ring-shaped fault forming a tube down which the upper 
part of the cone of fragmental rock has dropped either alone or carrying 
with it a casing of the Fish Eiver beds. On this view one would expect to 
see more brecciation of the walls at the contact and a less regular arrange- 
ment of the contents of the pipe, but there is no evidence of faulting 
beyond local slickensides such as are found in many sedimentary rocks 
which have been shghtly disturbed. 
The description of Coon Butte in Arizona, written by G. K. Gilbert * 
seemed to offer an analogous example of a great explosion-crater un- 
accompanied by the appearance of ordinary volcanic rocks, and it is of 
more recent date. About a crater 1,300 yards wide in flat sedimentary beds 
there is a ring of fragmentary material, 200 feet high at most, composed of 
pieces of the sedimentary rocks, in this case limestone and sandstone. 
The depth of the crater floor below the ridge of fragments is 600 feet. 
The size of the fragments at Coon Butte is very much larger than any- 
thing seen at Geitsi Gubib, the larger blocks measuring in some cases 
100 feet across. There has been no hardening of the breccia on the wall 
at Coon Butte, but it evidently presents some points of resemblance to 
Geitsi Gubib. 
More recent information ! got through extensive prospecting opera- 
tions certainly shows that the alternative hypothesis considered and 
discarded by Gilbert may be correct, though it involves great difii- 
culties. It is that a meteorite travelling very fast hit the earth there 
and smashed in the rocks, causing a sort of explosion violent enough to 
throw pieces of rock 50 tons in weight a mile from the point of impact. 
The chief evidence supporting the meteoric origin is briefly : (1) the 
finding of several tons' weight of meteoric nickel-iron fragments round the- 
crater and in the rim ; (2) the presence of nickeliferous magnetite and 
schreibersite fragments in the smashed rock at various depths down to 
600 feet below the floor of the crater ; (3) the supposed presence of an 
unbroken sheet of sedimentary rock under the crater at the depth of 
700 feet or so ; and (4) the absence of any volcanic rocks in the crater 
and in the ejected material. The third point is conclusive if correct, 
* Fourteenth Ann. Rep. U.S. Geol. Survey, i., p. 187, and Presidential Address, Geol. 
Soc. of Washington, 1895. 
t D. M. Barringer, Proc. Acad. Nat. Sc. Philadelphia, Ivii., pp. 861-880 ; B. C. Tilghman, 
ibid. pp. 881-914; G. P. Merrill, Smithsonian Misc. Coll., 1., pp. 461-489, and Proc. 
Australasian Ass. for Adv. Sci., 1909, vol. xii,, p. 320. A later account of the work by 
Mr. Barringer was published separately at Philadelphia in 1909, but I have not seen it. 
Through the kindness of Mr. G. W. Lamplugh, who read the book in the Geol. Society's 
Library in London, I hear that in Mr. Barringer's opinion the " sandstone has been 
found to be in place, to be unaltered, and to be occupying a horizontal position." The 
proof that the sandstone met with is not a great block would appear to be still wanting. 
