68 Annals of the South African Musemii. 



But the sand-dunes, under the impulsion of impetuous and season- 

 ably recurring winds, have swayed to and fro for years untold, moving 

 forwards and backwards, thus disintegrating the sandstone of the 

 hills encountered in their onward or retrograde march — a dis- 

 integration greatly helped by saline matters. These dunes must 

 have been, and still are, a most powerful erosive factor, as is but too 

 plain when the rocks which they cover are laid bare. 



Now, it must be remembered that bouchers and nuclei are found 

 at the top of the partly uncovered slope, some are so wedged in the 

 deeply and broadly pitted sandstone that a hammer had to be used 

 to remove them. As it is impossible that these very heavy imple- 

 ments could have been rolled upwards from a lower level by the 

 sand that forms these billowy dunes, we cannot but conclude that 

 they were deposited in the situation in which we found them from a 

 once more elevated level, a higher level eroded by the action of the 

 shifting sands, and an erosion doubtless of great antiquity.* 



But the recent quartzite scrapers are found in the same situation ; 

 and why should they not have come also from the eroded higher 

 level ? 



In dealing with the Tsumi Eiver deposit, I have pointed out 

 that scrapers of exactly the same type as those occurring on the 

 Gape Flats, and associated there with beautiful Solutrian laurel- 

 leaf-like implements worked on both sides, were found in such juxta- 

 position with most finished bouchers of an x\cheulean type that it is 

 impossible to doubt the contemporaneity of the two. The same 

 thing occurs also on the Cape Flats. Small bouchers of Acheulean 

 type, made of the same recent quartzite as the Solutrian implements, 

 are occasionally met with, but, I must add, lying by themselves. 

 Fig. 18, PI. 93, represents one of these bouchers. 



It would not, therefore, be safe to deny the possibility of these 

 surface quartzite flakes having been coexistent with the Chellean- 

 Mousterian type found at Fishhook, except for the following 

 reason : — 



Further investigation has revealed, close by, a midden containing 

 the usual domestic appliances of flakes and chips, querns and pot- 

 sherds, and among them a javelin-head of the advanced type, but 



* Since this account was written bouchers of the same palaeolithic type as those 

 found at Fishhook have been discovered at the very top of the mountains flanking 

 the valley. At Glencairn they show the same effect of sand-blast, and were dis- 

 covered along the only possible opening that could lead to the " Kommetjee," on 

 the other side of that part of the peninsula. 



We have thus a repetition of the occupation of heights by the early, or primitive 

 makers, comparable to the Paarl, Simondium, and other sites. 



