12 Transactions of the South African Philosophical Society. 



attributed to the roughness of my measurements of depths (100- 

 foot intervals), and I accepted a loss of 1-70 inches per month as 

 the most probable loss from percolation only, this loss being that 

 calculated for an assumed mean depth of water of 6 feet at the end 

 of July. 



I had a second means of approximating this loss. I had at an 

 early period noticed that for two or three days after a rapid rise of 

 water in the dam there was a curiously rapid fall, and this I found 

 to be owing to local loss about the perimeter of the flooded area, and, 

 knowing the comparative depths of water in wells on three sides of 

 the dam, I constructed curves of escapement sections and used the 

 rapid losses following on a rise of water in calculating the coefficients 

 of escape for different depths. On the whole I found this method 

 to give me nearly the same results as those gained by the other 

 method ; the means giving a fraction over 18 inches loss per 

 annum. 



With respect to the measurements after the great catch in 1887, 

 I find that between March 7th, the day on which the water attained 

 its maximum height of 13 ■ 2 feet, and March 24th, when the next rain 

 fell, I reported to Government that " the loss of 0-42 feet in 17 days 

 at this season is below my tabular rate for lesser depths, which is as 

 it should be." This loss of 0*42 includes the rapid fall I have before 

 spoken of, and I have not now data at hand which would enable me 

 to state its value, but even the 0*42 is only at the rate of 8*89 inches 

 per 30 days against a mean of 10*10 inches for the two years tabu- 

 lated ; but the probability is that at least 1 inch dropped within 

 the first two days, which would bring the monthly rate down to 

 8 inches. 



At a later date, June 25th, I reported a loss of four inches in 35 

 days, or at the rate of 3*42 inches per 30 days, and this at a time 

 when I had fully 500 morgen of land under irrigation. Here, again, 

 the corrected limit would show less than the mean for the two years 

 tabulated, but my diaries in which detailed observations are noted 

 are still at Van Wyks Vley. 



I much regret that I have not access to the data upon which I 

 made my ultimate determination as to the loss by percolation. I 

 had hoped that I should be able to show that it cannot be much less 

 than 20 inches per annum, but I am afraid that I must ask you to 

 believe that the evaporation curves persistently refused to account 

 for the balance of loss that I now ascribe to the effect of percolation ; 

 and that, in regard to the continuance of that rate of loss for some 

 years to come, which I think probable. I thoroughly examined the 

 bed of the dam in 1891, and found so slight a film of silt that I am 



