HORN EXPRDITION — PHYSICAL OKOfiKAPTrY. 15 



• 



.ind duiinf;- tlie long intervals of dry woatlipr, surfacp runiiiiii;- water is ahsoiit from 

 tlie greater Iciigtli of the cliaiinel, its appearance for slioi't distances only at cei'tain 

 parts being due purely to local causes. The aljsence or paucity of springs is to be 

 attributed to tlie absence of a sufKciently thick layer of soil on the mountain slopes 

 to act as a reservoir l)y absorbing the rain water and giving it out gradually at 

 lower levels in the form of springs. Tt is also due to the aljsence of joints in tlie 

 Pre-Cambriaii and their scai'city even in the Ordovician rocks. The rain watei-s 

 are, therefore, not absorlied, but form torrents, and rush down the Itare mountain 

 slopes into the \alleys. When the watei's have reached the beds of the water- 

 courses, their rapid absorption by the poi-ous strata and the excessi\'e amount of 

 evaporation that is always taking place cause the almost total disappcai-ance of 

 surface water from the river channels. 



(<■) W'aferlioles. — At certain .seasons of the year, should tlie fall of rain l)e 

 sufficiently great, the supply of water in the cliannels of the rivers exceeds the 

 amount that can be alisorbed by theii' sandy beds, and this pioduces a flow of water 

 down tlie cliannel. This is often spoken of locally as a, flood, as the waters as a 

 rule are not confined to the channel l)y the low ri\erdi;inks, but spread out over 

 the wide alluvial flats which border tlie main channel. 



When rain ceases to fall the flow of water diminishes almost innnediately in 

 volume. This it continues to do until running water disappears from the surface 

 over an ever-increasing length of its course. There are, however, for a, mouth or 

 two after a heavy rainfall portions of the channel over which running water may 

 still be seen ; but these gradually decrease in length until the channel assumes 

 what may be termecl its normal state. \\\ this state tlu^ chaniKd is occujiied by 

 long stretches of white sand dcndid of sui'face water, separated at i-are intervals by- 

 short lengths of the channel, where water may be seen flowing gently <iver a- rocky 

 bed. Tn the apparently waterless stretches, however, water may be obtained l>y 

 sinking to deptiis \arying with the nature of the bottom and tlie laj)se of time 

 since the last heavy fall of rain. The appearance of surface water af r.ire intervals 

 in the bed of the river is due to the presence of "bars " of rock, which cross the 

 channel at these places. 



As the water in its downward progress cmnot filter through these I'ocks as it 

 does through the sand and gravel at other parts, it has to rise tr) the surface fo 

 pass over them. The change fi'om the norm.al state of flie ch.iniiels to their next 

 condition is a- more gradual one. As the siip])ly of water becomes less .iiid less, 

 the amount and r.ate of flow over tin' rocky liars gradually diiiniiish, until finally 

 the supply becomes too small to cause the water to rise over flicm at all. 



