328 LIVINGSTONE'S LAST JOURNALS. [Chap. XII. 



their union are formed streams of from thirty to eighty 

 or 100 yards broad, and always deep enough to require 

 either canoes or bridges. These I propose to call the 

 secondary sources, and as in the case of the Nile they are 

 drawn off by three lines of drainage, they become the head 

 waters (the caput Nili) of the river of Egypt. 



Thanks to that all-embracing Providence, which has 

 watched over and enabled me to discover what I have 

 done. There is still much to do, and if health and pro- 

 tection be granted I shall make a complete thing of it. 



[Then he adds in a note a little further on : — ] 



But few of the sponges on the watershed ever dry ; 

 elsewhere many do ; the cracks in their surface are from 

 15 to 18 inches deep, with lips from 2 to 3 inches apart. 

 Crabs and other animals in clearing out their runs reveal 

 what I verified by actually digging wells at Kizinga and 

 in Kabuire, and also observed in the ditches 15 feet deep 

 dug by the natives round many of their stockades, that the 

 sponge rests on a stratum of fine white washed sand. These 

 cracks afford a good idea of the effect of the rains : the 

 partial thunder-showers of October, November, December, 

 and even January, produce no effect on them ; it is only 

 when the sun begins to return from his greatest southern 

 declination that the cracks close their large lips. The 

 whole sponge is borne up, and covers an enormous mass of 

 water, oozing forth in March and April forming the inun- 

 dations. These floods in the Congo, Zambesi, and Nile 

 require different times to reach the sea. The bulk of the 

 Zambesi is further augmented by the greater rains finding 

 many pools in the beds of its feeders filled in February, as 

 soon as the sun comes north. 



Mem. — In apparent contradiction of the foregoing, so far 

 as touches the sources of the Zambesi, Syde bin Habib 

 informed me a few days ago that he visited the sources 



