168 



The average results of the above Witteberg soils may well be added 

 for comparison with those already given. They are as follows: 



Lime '051 



Potash -058 



Phosphoric oxide '065 



These soils, as before remarked, may be expected to resemble those 

 from the similar sandstones of the Table Mountain series. There is, how- 

 ever, more clayey material in the Witteberg rocks, and hence if the sub- 

 ject were more closely investigated we should probably discover that the 

 soils derived from them do not exhibit quite the absolute poverty of the 

 Table Mountain sandstone soils. It is true that the above four analyses, 

 representing the Witteberg series, are along the line of this view, but they 

 are altogether too few in number to add any confirmation thereto. 



The soils derived from the Dwyka series, which overlie the Witteberg 

 beds, lead us from the Cape into the Karroo system, where we meet with 

 an all-round richer type of soil than the average of the older system. Of 

 the soils which represent the Dwyka beds of shales and conglomerates we 

 have unfortunately examined only two that can be taken as in any way 

 typical, and both of these were collected within the Worcester Division. 

 Their analyses resulted as follows : 



XII. DWYKA SOILS. 



Phosphoric 

 Division. Farm. Fine earth. Water. Linie. Potash, oxide. 



Worcester. De Hoek. 841 3'80 1/640 157 '065 



Aan de Doom River. 86'6 T08 '386 119 "052 



We cannot, of course, generalise from the results of two analyses; all 

 that can here be said is that these two differ from those of the Table 

 Mountain and Witteberg sandstones, on the one hand, by shewing adequate 

 proportions of lime and potash, and from the Bokkeveld soils, on the other, 

 in their relatively low phosphoric oxide. The matrix of the rocks con- 

 stituting the conglomerates is composed of quite a variety of minerals 

 amongst which felspar, augite and calcite are prominent and ^would prob- 

 ably disintegrate into just such soils as are typified in the above two 

 analyses.* 



Passing over the intervening rocks of the Karroo system, in respect 

 of which it would be difficult, for several reasons, to submit any analyses 

 of typical soils at the present stage, we have next to glance briefly at the 

 upper series of this system, namely, the Stormberg series, which we may 

 group with the Burghersdorp or uppermost beds of the Beaufort series just 

 below them. These consist of sandstones, shales and clays, interspersed 

 with calcareous rooks, and producing soils of a red or purplish colour. f I 

 do not think that we can assign any of these soils, that have hitherto 

 been analysed, as definitely to certain geological series, as in the case of 

 the older rocks, partly because many of the samples were not officially 



* The Dwyka series consists of three members ; (1) the upper shales, (2) the boulder 

 beds, and (3) the lower shale 3 !: the resemblance to Stow's "olive shales" noted on an earlier 

 page (162) refers to the phale beds, which may be expected to yield soils differing markedly 

 from those resulting from the boulder beds with their calcareous matrix 



f See page 89. 



