Vol. 70.] MIOCENE OF THE YICTOEIA XYAXZA, ETC. 155 



At the base of the valley of the Agulu Muk, the river-gravels 

 overlie a rotten rock resembling a quartz-porphyry; and at the 

 south-eastern corner of Homa Bay the hills of Najanja (rising to 

 4151 feet, and trending from north-west to south-east) form an 

 important physical feature. They are built up of a purple quartz- 

 porphyrite, 1 probably part of the same mass. In places the 

 porphyrite is traversed by quartz-veins, and forms rounded hills 

 and ridges ; it weathers to a red-brown soil. 



Crossing this spur of quartz-porphyrite, I found that its eastern 

 slopes were bordered by a low selvage, 300 }'ards wide, of banded 

 gypsum, layers of pure gypsum (usually ^ inch, but sometimes as 

 much as an inch thick) occurring in frequent alternation with 

 yellow-brown argillaceous layers of similar thickness. The outcrop 

 has a general north-westerly and south-easterly strike, dips 10° 

 south-south-westwards, and shows a bare fretted surface. Many 

 loose blocks are lying on the surface, tilted in all directions. It is 

 probably of the same age as the presumably Pliocene beds, which I 

 found north of Homa Mountain (see below) ; but this area must 

 have been separated and shut off from the Nyanza, to form a lagoon 

 without an exit. 



My course lay north-eastwards across the alluvial plain, about 

 4 miles wide, of the Southern Awach or Aloychi River, from which 

 isolated cones of basalt, as, for instance, Uchimbo and Chamanga 

 (or Chanvanga) rise abruptly with slopes of 30° to 40°. A thick 

 envelope of regur covers the northern part of the plain. On 

 passing Chamanga, one notes that the ground rises slightly at the 

 village of Kurngendia to form a low rounded ridge, about 2 miles 

 wide. The rock does not crop out at the surface, but the purplish- 

 brown sandy soil is probably derived from a rock similar to the 

 quartz-porphyrite of Najanja. 



The northern slope is steeper, and the path to Kendu passes 

 between a low hill of basalt on the west and the basaltic cone of 

 Orian (Rabur) on the east. Pebbles of quartz, jasper, and hsema- 

 titic quartzite (like that of the Kisii Highlands) lie scattered on 

 this slope, and perhaps indicate the old lake-beach at about 

 4000 feet, to which I have already referred (p. 146). 



Another alluvial plain (also covered with black earth) extends 

 for about 4 miles to the Northern Awach ; while on the west it is 

 bordered by an outspur from Homa, ending in a low scarped bluff : 



1 The rock has a splintery fracture, and contains somewhat melted-down 

 phenocrysts (up to 4 mm.) of plagioclase (oligoclase-andesine, Ab 5 An 3 ) 

 which enclose occasional particles of green augite and little patches of a 

 pale-green serpentinous aggregate ; some periclme -twinning occurs. The 

 quartz-phenocrysts (up to 7 mm.) are corroded and lobed in the typical 

 manner, but also show a few hexagonal sections ; they moreover enclose cir- 

 cular patches of the serpentinous aggregate: a very narrow, radiate, quartz* >se 

 selvage borders the quartz -crystals, but is absent round the felspars. Minute 

 crystals of pale-green augite occur very sparingly. The ground-mass is 

 dusty, quartzose, sometimes microspherulitic, containing abundant microlites 

 and needles of oligoclase and occasionally small aggregates of quartz-granules. 

 The pale-green serpentinous aggregate also occurs disseminated in the ground- 

 mass, and probably represents original glassy matter. Hcematite-dust imparts 

 to the rock its reddish colour. 



