408 W. M. DAVIS OBSERVATIONS IN SOUTH AFRICA 



on it, is shown in plate 52, figure 2. The largest boulder that we saw 

 here was a block of Barberton slate 6 feet through. All the significant 

 facts noted here — the rounding and striation of the older rock floor, the 

 unstratified structure of the Dwyka, the varied composition of the in- 

 cluded fragments, large and small, their subangular form and fre- 

 quently scratched surfaces — are absolutely typical of strong glacial action. 



A special interest attaches to the direction of ice-movement as recorded 

 in the grooving and striation of the Barberton slates. The alignment, 

 roughly north and south, is manifest at a glance. After careful inspection 

 of several localities we came unanimously to the conclusion that the mo- 

 tion of the ice-sheet had been southward. The true bearing of the 

 striations was south 25 degrees east. As will be seen later, this agrees in 

 a general way with the direction of ice-motion determined at a number 

 of other locailties, and thus indicates one of the most extraordinary pecu- 

 liarities of the South African Permian glaciation — the ice-sheet in lati- 

 tude 27 moved from the region of the equator toward the region of the 

 pole. 



We followed up one of the larger ravines southward along the west 

 base of ISTgotshe mountain and ascended through the whole thickness of 

 the Dwyka formation — the greater part of 1,000 feet, as I estimated it — 

 to the lower members of the overlying Ecca. The division between the 

 two formations was drawn where the shales or sandstones begin to be free 

 from included stones. The Dwyka is shown to consist of several sub- 

 divisions of stronger and weaker members by the occurrence of well de- 

 fined benches, scarps, and slopes that contour about horizontally around 

 the hillsides. Some of the weaker members were stratified, as they were 

 at Laingsburg, but their sequence seemed to include a greater number of 

 alternations of tillite and shales here than there, especially near the upper 

 part of the formation. The overlying Ecca followed in perfect conform- 

 ity, as far as we could see. It contains a good seam of coal several hun- 

 dred feet above the Dwyka ; it is capped at the top of the mountain with a 

 sheet of dolerite, as is so often the case in the strongly dissected belt of 

 country where the eastward descent is made from the plateau of the High 

 Veld to the coastal lowlands. 



This fourth day on the Dwyka was most instructive from the contrast 

 presented by the unconformable succession of the tillite on the glaciated 

 Barberton surface shown here and the conformable succession of the 

 Dwyka members on the Witteberg shales shown near Laingsburg. The 

 contrast is the same as that which is presented between the glacial 

 features of the Laurentian region of Canada and of the northern United 

 States, or between those of Scandinavia and northern Germany, and this 



