THE DWYKA FORMATION 401 



certainty less than its original area, for its outcrops everywhere indicate 

 a loss by erosion. 



The glacial origin of the Dwyka formation is as unquestionable as is 

 that of the drift sheets of northeastern America or of northwestern 

 Europe; but, singular to relate, the Permian ice-sheet by which the Dwyka 

 was formed moved in general southward, from the region of the equator 

 toward the region of the pole. 



Accounts of the progress of geological investigation in South Africa 

 with particular reference to the recognition of the Dwyka formation as of 

 glacial origin are given by Corstorphine, Hatch and Corstorphine (10-18, 

 197-210), Eogers, and Mellor. It is probable that the glacial origin 

 of the Dwyka would have earlier found a more general acceptance if 

 various other origins had not been suggested for it by several South 

 African observers. 



My object here is to set forth certain observations that I made on the 

 Dwyka with its associated formations and certain conclusions to which 

 these observations led. 



The capital letters in figure 1 are the initials of the localities that our 

 parties visited; the small letters are initials of especially significant local- 

 ities described by Eogers, Mellor, Anderson, and others, which we did not 

 see and to which reference is made farther on. In speaking of the un- 

 stratified parts of the Dwyka, I shall avoid the term "congolmerate" and 

 adopt Penck's term, tillite, meaning thereby a rock formed by the con- 

 solidation of glacial till. 



The Dwyka near Matjesfontein. — Our first sight of the Dwyka was on 

 the excursion led by Mr Eogers, as already described. We reached Mat- 

 jesfontein, in the Karroo district, on the morning of August 20, and 

 there found ourselves in the Karroo, the dry interior country of east-and- 

 west ridges and valleys that lie between the great interior highland and 

 the Cape Colony ranges. Here we had a walk of about 5 miles northward 

 from the neat railway village across broad ridges and valleys of Dwyka 

 and Ecca. A few miles to the south rose a long, even crested anticlinal 

 mountain of Witteberg sandstones. Farther away in the north rose the 

 escarpment of the plateau country of horizontal Ecca and Beaufort for- 

 mations. The whole district hereabout is treeless ; the water-courses were 

 nearly all dry at the time of our visit. The surface of the ground was 

 abundantly exposed between the scattered plants, and the exposures of 

 bare rock were plentiful where a thin, stony soil had not accumulated. 

 The following facts were noted. 



The matrix or groundmass of the Dwyka tillite is here dark, fine- 

 grained, unstratified and trap-like in appearance; it contains numerous 



