THE WATER SUPPLY OF SYDNEY. XCV. 
character as to hold all the accumulation and to give it up 
gradually to the rivers. The intensity of rainfall on the 
former class of catchment would have an important bearing, 
and it is quite possible that an evenly distributed rainfall 
much below the average for a series of years, might be more 
advantageous to the reservoirs than a series of bounteous 
years with a rainfall far above the average, but which fell 
at a high rate of intensity. I think it will be found that 
it is not the accumulation of rainfall as disclosed by the 
residual mass curve which is the factor engineers have to 
look for, but the run off and above all the fluctuations of 
the run off. Perhaps the residual mass curve may be of 
service for indicating the state of saturation of the ground 
and the periods when a greater or less run off may be 
anticipated, but I think the plotting of the rainfall itself 
would shew the same thing in a less deceptive manner. I 
think the plan adopted by Sir Alexander Binnie in a paper 
read before the Institute of Civil Hngineers in 1892 is a 
better one for studying the effect of rainfall. The method 
adopted is to reduce the rainfalls for each year to a per- 
centage or ratio of the average rainfall and then plot them 
from one common datum, the average rainfall being taken | 
as unity. The method consists in dividing the rainfall for 
each year by the average rainfall thus— 
49°06 
19°35 1°015 
shewing the fall for that year to be 147: above the average, 
or again— 39°14 _ pve 
48°35 nae 
that is the fall for the year was a little more than 72% of 
the average, or fell short thereof by 28%. 
I have prepared a diagram (Plate 49) of the Sydney rain- 
fall on this principle. It will be seen that all negative 
quantities are avoided, and whenall the years in the series 
are so treated, if the average has been accurately ascer- 
