526 PROFESSOR J. W. GREGORY. 



unaltered red sandstones which occur in so many parts of the world between the base 

 of the Cambrian and the great Eparchean unconformity. The considerable re- 

 crystallisation of some of these quartzites, with the development of large flakes of 

 authigenous muscovite (PI. II, fig. 2) and lines of epidote, is consistent with the 

 pre-Cambrian age of these rocks. 



(3) The Lej)i Gfraywackes. — The Lepi beds are nearly related to the Oendolongo 

 Series, but they are more disturbed and altered. They include a series of gray- 

 wackes, some of which are coarse-grained and others very fine-grained. Some of the 

 fine-grained varieties have been silicified, and I recorded them in my field notes as 

 cherts. The beds include rhyolite tuff's similar to those in the Oendolongo Hills, 

 but I saw no rhyolites. East of Lepi the beds show an incipient foliation, and the 

 felspathic material along these developing foliation planes has been altered into lines 

 of epidote. The Lepi beds appear to correspond with the lower part of the Oendo- 

 longo Series, and contain the same rhyolite tuffs, but they were deposited beyond the 

 range of the lavas. 



In North-Western Katanga the Kundelungu beds are underlain conformably by 

 the Lubudi beds, which are cleaved and include an extensive series of cherts and 

 oolitic limestones ; and though I did not find any limestones in my hurried traverse 

 across the Lepi Series, it agrees in so many respects with the Lubudi beds that these 

 groups may be provisionally correlated. I had no opportunity of determining 

 whether the Lepi graywackes are conformable to the Oendolongo Sandstones, for I 

 nowhere saw them in contact ; but the Lepi beds are the more steeply inclined and 

 more disturbed. They correspond in petrographic characters to the rocks of the 

 Transvaal System, the age of which is uncertain. It may be Lower Palaeozoic, an 

 opinion which has been strengthened by Hermann's discovery of a badly preserved 

 Orthoceras in the Otavi dolomites of Damaraland (Hermann, 1908, pp. 265, 266), 

 which have been correlated with the dolomites of the Transvaal System. The 

 evidence for this correlation is, however, not convincing ; and Hermann (ibid., 

 p. 270) regards the Otavi dolomite as the equivalent of the Bokkeveld beds, and if 

 so they are much later than any part of the Transvaal System. The correlation of 

 these fossiliferous dolomites of German South-West Africa with the Transvaal 

 dolomites may suffer the fate of several previous suggestions as to the identity of 

 marginal marine beds in South Africa with rocks on the high plateaus of the interior. 



(4) The Huambo Quartzites. — The Lepi beds are bounded to the east by a wide 

 tract of granite, which further east beyond Huambo is covered by granular non- 

 crystalline quartzites. These rocks must have been coarse sandstones originally ; 

 their cement has recrystallised into small grains, so that the larger grains are fretted 

 together and show strain effects. They show no definite foliation. Specimens were 

 collected on the Lundula River, also east of Candumbo, and north of Gordon's Store 

 near Saccanjimba (No. 187). Similar quartzites occur further north near Cinjamba 

 and the Canji River. 1 did not see them clearly associated with the gneisses, and 



