1 74 GEOLoar. 



between the highlands and the plains near Annesley Bay. 

 One such cap is seen north-east of Sowera on the oppo- 

 site side of a deep ravine. Another lies east-south-east 

 of Sowera. A third is not seen from the highlands, but 

 was found at a much lower level in the hills further east 

 towards Arafild by Lieutenant Holditch, E.E., of the 

 Trigonometrical Survey. 



The whole of the Sowera plateau consists of sand- 

 stone, but the rock ends somewhat abruptly on the 

 Asawat plateau between the Haddas and Komayli ravines, 

 and also west of the former, near Halai, which village 

 stands on metamorphic rocks. It is curious that the 

 northern part of each plateau, which consists of meta- 

 morphics, is higher than the southern part. This shows 

 on how uneven a surface the sandstone was deposited. 

 The limit of that formation nearly corresponds with an 

 ■east and west line on both sides of the Haddas vaUey, 

 but there does not appear good evidence of faulting. 

 On the plateau of Hamazen, north of Halai, there is said 

 to be no sandstone. 



The sandstone plateaux are covered with thin jungle, 

 and do not appear to be sufficiently fertile for cultivation 

 in general, except where a richer soil is derived from 

 trap rocks in the neighbourhood, as is the case near 

 Adigrat. 



The extent of the sandstone to the westward was of 

 course not determined. Looking from the verge of the 

 scarps near Takonda, it appears that for some distance 

 metamorphics prevail, but that the sedimentary rocks 

 recur on some ranges beyond Dixa. 



