OF THE RIFT VALLEY 
485 
geological antiquity, for all the craters through which it was 
discharged have been swept away ; but, here and there, on 
the plains, are low circular elevations gently sloping outwards 
in all directions. These rises are probably the vents through 
which the lava was discharged. 
The phonolite of the Kapiti Plains may be regarded as 
material squeezed from below the crust by the sinking of the 
floor of the Indian Ocean. 
The subsidence of the ocean was probably accompanied 
by the uplift of British East Africa into a vast low arch ; and 
as the country settled after these two associated movements 
the centre of the arch was insufficiently supported and sank 
between a series of parallel fractures. This subsidence was 
the first of the movements which formed the Rift Valley. 
That this movement may have begun as early as the 
Oligocene is indicated by a fossil sea-urchin (an Echinolompas) 
in the British Museum, which probably came from Lake Nyasa ; 
and, if so, this fossil indicates that the sea occupied the southern 
end of the Rift Valley in Oligocene times. The evidence of 
this fossil is somewhat doubtful, since its exact locality and 
discoverer are not known. Moreover, Messrs. E. 0. Teale 
and R. C. Wilson have shown that the Rift Valley faults 
which continue those of the Nyasa basin, south of the Zambesi, 
are there post-Eocene, and their collections give no evidence 
of any post-Eocene marine beds in the southern end of the 
Rift Valley. 
That the volcanic history of British East Africa began 
before the Miocene is indicated by the evidence of the augite 
andesite of Kikongo to the east of the Victoria Nyanza. Some 
lake beds in the same province have yielded the bones of a 
primitive elephant — a species of Dinotherium, which has been 
appropriately named D. Hobleyi. Its age is Miocene, and 
Dr. Oswald has show T n that the lake beds containing this 
fossil are earlier than the basalts and phonolites of Lumbwa, 
but younger than the augite andesite of Kikongo and the 
volcanic agglomerate of Metamala. 
If this augite andesite is the same age as the augitic 
lavas of the Rift Valley, then they also are pre-Miocene ; and 
the phonolites of the Kapiti Plains must be much older than 
