F. P. Mennell — South African Petrogmpliy. 357 



In the present chaotic state of petrological nomenclature it is 

 necessary to explain the significance of the terms employed in 

 describing the rocks. No system of classification can be regarded 

 as wholly satisfactory, but it certainly seems hopeless to rely on 

 such features as the presence or absence of particular minerals. 

 Such a course brings together types which differ widely in chemical 

 composition, and separates others which are chemically and geneti- 

 cally closely related. In the brief notes which follow, the grouping 

 depends solely on mode of occurrence and chemical composition, 

 a method of classification which appears to be steadily gaining 

 ground owing to its simplicity. The significance of the nomen- 

 clature, which accords strictly with the classification adopted, will 

 be readily perceived from the following table, names in general use 

 being employed as far as possible and simply endowed with greater 

 precision. In individual cases the name of the most characteristic 

 mineral is prefixed for purposes of distinction. 



Plutonic 



\ = Rock consolidated 

 at great depths) . 

 Granite 

 Syenite 

 Diorite 

 Gabbro 



With these preliminaries we may pass on to the rocks themselves. 

 In South Africa as elsewhere, acid lavas and basic plutonic rocks 

 are rare, while basic lavas and acid plutonic masses are developed 

 on an enormous scale. The only rhyolite which has come under 

 my notice occurs near the Express Mine in the Umniati district 

 of Mashonaland. It shows sparingly distributed phenocrysts of 

 hornblende and orthoclase felspar with rarer quartz, set in a glassy 

 groundmass crowded with globulites, margarites, and longulites. 

 Partial devitrification is evidenced by a cloudiness between crossed 

 nicols, but there are no signs of spherulitic structures. Ilmenite and 

 brilliantly polarizing sphene occur as accessories, but neither is 

 abundant. 



Sub-acid lavas appear, like the rhyolites, to be poorly represented, 

 and I am not able to refer any rocks to this division with anything 

 like certainty. There are also a number of lavas probably belonging 

 to the andesites, but in the absence of analyses discrimination from 

 the basalts is a matter of considerable difficulty. A ' melaphyre ' ^ 

 from the Kimberley Mine at Kimberley belongs, however, to this 

 class. It is a fine-grained aggregate of lath-shaped felspars with 

 some interstitial chlorite, etc., which may partly represent originally 

 vitreous material. Some of the felspar is untwinned, and is 

 probably orthoclase, while the twinned crystals seem to be andesine. 

 There is no porphyritic constituent, and the amygdaloidal cavities, 

 which are lined with chlorite and filled with chalcedony or with 

 calcite, are the most interesting feature of the rock. Another and 

 rather more basic andesite occurs interstratified with sandstone 



1 This is obviously a different rock to that described by Stelzner. 



