142 dr. f. Oswald on the [June 1 914, 



angular boulders o£ scoriaceous nepheline-basalt 1 (see PL XXI), 

 30 yards in length from south-west to north-east and 20 yards 

 wide, rising rather abruptly to a height of 30 or 40 feet from the 

 general surface of the lava-plateau. 



The lava on the surface of this plateau is highly scoriaceous, with 

 steam-cavities measuring up to 4 inches in length, and it descends 

 very steeply, but apparently with its original slope, into the valleys 

 of Kachuku on the south and of Ubware on the north. In both 

 these cases I found the lava-stream to be highly brecciated with 

 large fragments of its own substance, indicating its superficial 

 character ; and, near the head of the Kachuku Valley, I noticed 

 that the lava-stream had not merely caught up and rolled along 

 cooled fragments of its own substance, but contained one of the 

 calcified tree-stems from the uppermost clays of the Miocene beds 

 over which the lava had flowed. 



The nepheline-basalt seems to be particularly susceptible to 

 atmospheric denudation, and, owing to its homogeneous nature, it 

 weathers readily into conical hills. The outlier of Homa still pre- 

 serves the plateau-like character, although the detached northern 

 peak is conical. The dissection of Ruri is much farther advanced, 

 and it consists entirely of a group of cones, while the outliers on 

 the extreme eastern margin of the basaltic area consist of separate 

 symmetrical cones : for instance, Chamanga, Uchimbo, and Asego, 

 rising with remarkable abruptness (slopes of 32°) from the level 

 alluvial plain. 



The products of denudation of the nepheline-basalt consist of 

 black cotton-soil or regur, a black alkaline clay (very sticky 

 and plastic when wet, but friable and cracking readily when dry). 

 Everywhere in this area the regur swathes the lower slopes of 

 the basalt hills in a thick mantle, and fills up the valleys. As 

 Mr. H. B. Maufe has remarked, 2 the most suitable conditions 

 for its formation are to be found on 



4 level or gently rolling ground where the rocky substratum is impervious and 

 the under-drainage consequently bad.' 



Now, the clays of the Miocene beds form a completely impervious 

 substratum, and here I found the greatest thickness of regur: 

 namely, up to 20 feet. Everywhere it contained land-shells ( Tro- 

 pidophora sp.) still showing colour-bands, with occasional angular 

 fragments of basalt ; and in the lower half the little gre} r , rough, 



1 This rock is also very similar to the Rungwena basalt, but contains no 

 porphyritic nepheline ; the augite shows the same characteristics, and ser- 

 pentinized olivines are also present, with a very few flakes of biotite, in a 

 somewhat similar ground-mass of looser texture. The nepheline of the 

 ground-mass has, however, become more individualized, although not showing 

 crystal-outlines ; allotriomorphic nepheline enclosing small augites and mag- 

 netites forms the major part of the matrix ; the remainder consists of a 

 yellowish base, probably serpentinized glass. A single hexagonal crystal of 

 leucite is visible in the slide. 



2 ' Report relating to the Geology of the East Africa Protectorate ' Colonial 

 Reports, Miscellaneous, No. 45, 1908, p. 55. 



