87 



Earossa catcliment area annually a loss o£ 8i inclies in ISSi, 

 and 5 inclies flowing in the river in the hills ; at Beetaloo, in 

 1885, Joss by percolation 12y'^o inches, and flowing in the river 

 3% inch; and at the Eiver Wakefiel, in 1886, loss by percolation 

 lOf inches, and flowing in the river in the hills where the river 

 gauge was fixed, yVo ot* an inch. 



It will be seen by these calculations from the river gaugings 

 that as the rainfall decreases, as in the years 1885 and 1886 in 

 this colony, the proportion of water percolating underground 

 largely increases. The same takes place in other countries. 

 Eor instance, at the gathering ground of the Liverpool "Water- 

 works it was observed that with a rainfall of 48 inches the 

 water lost by percolation was only 18 per cent., whilst during 

 a year with a rainfall of only 34 inches the loss from this cause 

 was 33 per cent, of the rainfall. I have not noticed the differ- 

 ences of the geological features of the catchment areas of our 

 rivers, as probably not much variation is due to this cause. 



3. The next conclusion I draw is that in South Australia 

 generally the waters passing underground are much larger in 

 quantity than those flowing on the surface. It will be seen 

 that the amount of loss by percolation in the mountain gather- 

 ing grounds of the rivers referred to is very much larger than the 

 portion flowing in the rivers. The comparative account is as 

 follows, viz. : — In 1881, a year with about the average rainfall 

 of the previous 20 years, the water lost by percolation in the 

 hills is one-aud-a half times the quantity flowing in the rivers 

 at that point. In 1885 the percolation is twenty times the flow 

 in the rivers, and in 1886 the loss by percolation is no less 

 than about sixty times the flow in the rivers. Some of these 

 waters reappear in the beds of our rivers in the hills in the 

 form of springs more or less permanent, but only in small 

 quantities compared to the water that has gone permanently 

 underground, and, in most cases, this spring water, after flow- 

 ing a short distance, is again lost by percolation. This is seen 

 in Mount Brown and Spring Creeks, Willochra district ; the 

 Saunders Creek and South Ehine Eiver ; the Elvers Torrens, 

 Wakefield, and Broughton ; in some of the creeks in Baroota 

 and other localities on the western slopes of the Elinders 

 Eanges ; also in many other districts that might be named. 



In addition to this loss in the hills we have also to consider 

 the further great loss of water in the beds of the rivers and 

 creeks after they debouch upon the plains, which, as is well 

 known, sometimes amounts to a total disappearance of the 

 stream into underground channels. 



In August, 1886, I observed the flood waters of the Eiver 

 "Wakefield coming down after a long continuance of dry 

 weather, and the progress in its channel across the plain near 



