40 DR. A. HOLMES ON THE PRE -CAMBRIAN [vol. lxxiv, 



(c) The Ribawe Mountains. 



No schists nor crystalline limestone have yet been detected in the 

 gorges and precipitous slopes of the Ribawe Mountains (see map, 

 PI. XI). The chief rock is a biotite-gneiss, dipping gently north- 

 wards and passing locally into a gre}^ porphyritic granite. There 

 are, however, bands of hornblende-gneiss, and later intrusions of 

 augite-granite, in the neighbourhood of certain curious caves which 

 may have once been occupied by crystalline limestone that has 

 since been dissolved away. 1 The other intrusions in this district 

 are pink and grey granulitic gneisses, while the } T oungest rocks are 

 coarse pegmatites (some of which carry tourmaline and topaz, and 

 others magnetite and ilmenite) associated with red biotite-granite 

 containing large phenocrysts of orthoclase. 



(d) The Nrassi Basin. 



North-east of the Ribawe Mountains the upper waters of the 

 Nrassi River and its tributaries flow through a basin-like area 

 surrounded almost everywhere by linear series and clusters of peaks 

 (see map, PI. XI). In one of the small gullies, which in the wet 

 season carry the drainage from the peaks of Ericola to the Nrassi, 

 bands of crystalline limestone are exposed, again interfoliated with 

 hornblende-gneiss that merges on each side into biotite-gneiss. 

 In the Mwima range the last-mentioned rock passes into an unfoli- 

 ated grey granite with little biotite. As in other districts where 

 limestone occurs, there are small intrusions of augite-granite, 

 which in this case is associated with adamellite and diorite. 

 Here, again, are granulitic pegmatites, sometimes carrying haema- 

 tite, and associated in the isolated line of peaks north of Ribawe 

 with a pink and grey speckled biotite-granite. This rock con- 

 stitutes the main mass of some of the Karji and Koldwi peaks, 

 and in it Mr. E. J. Wayland was fortunate in finding enclosed 



1 For illustration, see ' The Lateritic Deposits of Mozambique ' Geol. Mag. 

 dec. 6, vol. i (1914) pi. xxxvii, facing p. 529. In that paper I wrote : ' On 

 each side of the [Sawa] valley the mountains rose up precipitously, and the 

 foliation of the gneiss could be plainly seen dipping away at a gentle angle 

 to the north. On the eastern side the nearly vertical face of the gneiss was 

 eaten away here and there into, caves, the entrances to which ran along the 

 " dip." Those which we explored widened and narrowed in the most remark- 

 able way, defying explanation. In each case a steady drip of water trickled 

 from the roof and from narrow fissures which penetrated the rock parallel 

 to the foliation, thus demonstrating that a considerable drainage of water 

 percolated northwards down the " dip " through the apparently impermeable 

 gneiss. The roofs of the caves were lined with a thick highly-polished 

 deposit of limonite, which also covered the upper parts of the walls, gradually 

 becoming thinner as the floor was reached and giving place to bright pink 

 rhodochroisite and black pyrolusite.' 



Since writing that paper, it has occurred to me that if, as is frequently the 

 case, hornblende-gneiss and augite-granite indicate the proximity of crys- 

 talline limestone bands, these remarkable caves may represent pockets of 

 crystalline limestone that were not completely absorbed by the invading 

 magmas. 



