XCVI. DISCUSSION. 
tained, the sum total of all the ratios will be equal to the 
number of years of observation. The principal question 
the engineer has to consider is the fluctuation of rainfall 
in order that he may design his works to equalize the years 
of bounty and scarcity, and this diagram presents at once 
to the eye the salient facts of the case. Let us take the 
wettest and driest years, and it will be found that the 
mean falls very close to the average rainfall or unity, thus 
(1860) (1849) 
1°713 + 0°444 
| 
= 1°078 or 7°8) above the average ; 
* 
~~ 
b 
here the wettest is 71°» above the average, and the driest 
is 56% below the average. Take the average of 2 wettest 
and 2 driest consecutive years 
(1860) (1861) (1908) (1907) 
(1°713 + 1°207) + (0°659 + 0°648) _ 1°0 
4 
average. Here the mean of the 2 wettest is 46% above the 
average and the mean of the 2 driest is 357 below the 
average. Take the average of 3 wettest and 3 driest con- 
secutive years 
(1890) (1891) (1892) (1905) (1906) (1907) 
(1°684 + 1°144 + 1°430) + (0°724 + 0°659 + 0'648)um 1048 
6 
or 48> above the average; here the average of the 3 wettest 
is 42° above the average, and the average of the 3 driest 
is 32> below the average. A study of the averages of 6 
wettest years 1840-5 and of 7 driest years 1901-7 shews a 
ratio of 1°035 or 34°» above the average; whilst the average 
of the 6 wettest years is 31) above the average and of the 
7 driest is 23% below the average. This discloses the fact 
that the shorter the period under review the greater the 
fluctuation in extremes of rainfall. Now if 29 inches are 
required to flow off the catchment to fill the Cataract dam, 
and 40% is allowed as the run off, we shall require a rainfall 
of 724 inches, which is 50% above the average. But if 25% 
57 or 5°77 above the 
