Fauna and Stratigraphy of the Stormberg Series. 489 



dence of aridity furnished by the Bushveld and Forest Sandstones; 

 and it is of interest here to examine the .evidence afforded by a 

 microscopic examination of the rocks from the various localities. 



The closest approximation to the true Cave Sandstone is possibly 

 afforded by the fine-grained sandstone from Komatipoort. This was 

 described by Kynaston as follows:- The rock "shows numerous 

 small quartz grains, of a generally uniform size, and on the whole 

 angular and sub-angular rather than rounded, embedded in an 

 exceedingly fine-grained matrix, in which individual grains are bar- 

 ely distinguishable. Grains of plagioclase felspar may occasionally 



be noted The microscopic characters agree very closely with 



those of the fine sandstones from the Springbok Flats, the propor- 

 tion of matrix to individual quartz grains being somewhat higher in 

 the latter." 



The Buiskop sandstone is of fairly line grain, and there are very 

 few grains which are coarser than the general texture of the rock. 

 A considerable amount of ferruginous cementing material is present 

 in the red varieties ; and the grains of quartz, although not splintery, 

 are polygonal in outline with rounded corners. The red variety is 

 very similar to the sandstones of the Red Beds. 



A specimen of sandstone from Nelson's Kop, in the west of the 

 Waterberg District near the Limpopo and Pongola Rivers, contains 

 two types of quartz grains - - a larger and a smaller. The larger 

 grains are less prevalent than the other; they are rounded without 

 any trace of angularity and they vary in size from a grain having 

 diameters of O.'3 mm. and -t 22 mm. to one with diameters of O6 

 mm. and 0-4 mm. The smaller grains, of a diameter of 0-15 mm. 

 and under, are not so obviously rounded, but none of them are 

 splintery, although thin sections are somewat polygonal. Most of 

 the grains are coated with a thin reddish-brown layer of iron oxide, 

 while there are scattered grains of the same material. Felspar is 

 much less common than in the Forest Sandstone. The whole sec- 

 tion is reminiscent of that figured by Molyneux as a typical section 

 of the Nyamandhlovu Sandstone. For the sample I am indebted 

 to Dr. du Toit, who also supplied me with a piece of Bushveld 

 Sandstone from Castle Kopjes, on the main road from Louis Tri- 

 chardt to Messina. 



The Castle Kopjes rock is light in colour and consists of equidi- 

 mensional grains of quartz with a very occasional grain of plagio- 

 clase. The grains are fairly well rounded, smaller than those of the 

 Nelson's Kop stone and similar in size to those of Buiskop. 



Reviewing the features of these Transvaal specimens, it would 



