Proceedings, &c, for 1885. 169 



minutes, and hence, though in a small watershed they might cause 

 comparatively heavy floods, yet in a large area their effect might be 

 quite inappreciable, for it is known that the very heavy rains, such 

 as 1 inch per hour, are extremely local. Such rains, also, are 

 short in duration, and the quantity which falls increases very 

 slowly with the time. 



The law of increase is approximately represented by the formula 

 F = k T J where F = fall, t = time, the areas and times being 

 supposed small. 



F 2 



Hence the rate of fall, R =— = LT~z 



Now let us consider a watershed of constant narrow width, and 

 let us assume that the water flows at a constant velocity down the 

 channel. 



Then the length drained will vary as the time 

 i.e., L cc T 

 but D = RxA = ExLxC 



i.e., J) oc R.L (2) 



Hence by (1) D oc — — (a) 



Corresponding with the formula given by L. D'a Jackson. Ed. 

 1883. 



But in practice none of the premises are correct. 

 Hence we should amend the formula to the form 



T> = h. 9 L 



T s ±p 



where p is some variable depending on soil, &c, but constant for 

 any given watershed. 



In most watersheds the slope of the ground is such as will tend 

 to equalise the velocity over the whole length, so that we are led 

 to the conclusion that the best simple formula is of the form 



a parabolic curve, though not a common parabola. 



From a careful study of the Bendigo Creek, the following data 

 resulted. 



D = 4100 cubic feet per sec. 



K = 10,000 acres 



L = 7Jm. 

 The Coiiban gave 



D = 10,000 cubic feet per sec. 



K = 100 square miles 

 And another small area gave 



D = 4 cubic feet per sec. 



K = 4 acres 



L = 7 chains 



