Geography and Ethnology. 267 
rushing over the country in innumerable creeks, and 
covering it in many places three and four feet deep. 
The inundation of the country by the river, at that 
season of the year, December and January, is a pheno- 
menon which arrests our attention. The Sabaki did 
not overflow her banks, and the Ozi was quietly sleep- 
ing far below hers. Why then should the Tana, 
which flows between the two, be so excessively full, 
and deluge the country to such an extent ? The 
Tana must have its sources in a country under very 
different conditions to the state of things which exists 
here upon the coast, and at the sources of the other 
rivers. The rainy season of the interior, we know, 
differs somewhat from that of the coast. Dr. Krapf 
says: "On the coast the second rainy season is in 
September and October ; in the interior its period is 
the months of November and December: the first or 
chief rainy season begins at Mombaz usually in 
April, in the interior in May or June" (p. 349). 
From this, superficial observation would lead to the 
conclusion that the overflow of the Tana is produced 
by the interior rains of November and December. 
Closer attention to the subject, however, will show 
that this can scarcely be so. First, the rains of Novem- 
ber and December are the small rains, and by no means 
sufficient to swell the rivers to such an extent as we 
found the Tana swollen. Secondly, if the rains had 
filled to overflowing the Tana, why should not the 
Ozi, or at any rate the Sabaki, be filled from the 
same cause } The idea of the Mto Tana's floods 
being caused by rain must be rejected. Dr. Krapf 
suggests that the Tana is supplied from the snows of 
Mount Kenia. He says: "Rumu wa Kikandi told 
