524 PROFESSOR J. W. GREGORY. 



Ball and Shaler report as " the common name for the Lubilache formation " (Ball 

 and Shaler, 1910, p. 688). These beds are probably represented in eastern Benguella 

 by the Bihe Sandstones. 



The Bihe Sandstones are a series of soft sandstones, which weather into beds of loose 

 white quartz sand ; they give rise to the wide plains of the Bulu-Vulu, and apparently 

 also of the Hungry Country to the east of Bihe and Fort Silva Porto. I only reached 

 the western margin of this area, but from the descriptions of the country to the east 

 of Bihe, it is probable that the wide extent of poor woodland in that part of Benguella 

 is due to the disintegration of the Bihe Sandstones.* 



The palseontological evidence, though meagre, is in favour of the Karroo and 

 probably Triassic age of the Lubilash beds, for Ball and Shaler (1910, p. 687) 

 definitely correlate them with the Sto'rmberg beds. Studt, however, correlates them 

 with the Waterberg Sandstones. South African opinion appears to be steadily 

 growing in favour of the view that the Waterberg System is Devonian and is the 

 Transvaal representative of the Table Mountain Sandstone. This opinion is not 

 universal, for Professor Schwarz has advanced (1912, pp. 187-188) arguments in 

 favour of the pre-Devonian age of the Waterberg System ; he has referred to the 

 similarity of the Waterberg Sandstones to the Torridonian of Scotland, though he 

 accompanies this suggestion by the warning that the two formations are too far 

 apart for the lithological correlation to be convincing. 



The Bihe Sandstone and other Lubilash beds present some important differences 

 from the Waterberg System of the Transvaal, which includes an upper series of well- 

 cemented red sandstones and a lower series of volcanic rocks. It seems to me there- 

 fore that the Bihe Sandstones may be provisionally correlated with the Lubilash beds, 

 and are younger than the Waterberg System ; and from the evidence in the Congo 

 they are probably of Triassic age. 



(2) The Oendolongo Series. — This series consists of firmly coherent sandstones, 

 quartzites and shales, and of rhyolites and rhyolite tuffs some of which have been 

 silicified into banded cherts. The rocks are neither foliated nor cleaved. The sand- 

 stones and quartzites are composed chiefly of quartz, but contain abundant grains 

 of felspars (PI. II, fig. l). They agree in character with the " Gres durs fels- 

 pathiques" of Cornet (1897, No. 1, p. 29) ; he identified these rocks in 1896 as his 

 Kundelungu System, and subsequently (1912, p. 9) included them in his slightly 

 dislocated division. The Oendolongo beds, apart from the presence of rhyolites and 

 tuffs, agree best with Cornet's description of the lower part of the Kundelungu 

 System ; this subdivision he has called the " Systeme de la Mpioka," and described it 

 (1897, No. 1, p. 29) as composed of red, micaceous, argillaceous beds which pass into 

 " psammite," and alternate with medium or fine-grained sandstones which are very 

 coherent, often felspathic, and red, gray, or dark in colour. They are slightly 



* This Hungry Country is said by Captain Boyd Cunninghame (1904, p. 155), who knows it well, to have a 

 good soil, and lie regards its poverty as unexplained. 



