201 



rainfall curve of County Low an (Vict.) and that of the south- 

 eastern counties of South Australia, the latter being through- 

 out somewhat higher than the former (see fig. 10). 



The apparent lack of correlation between the rainfall 

 and the lake levels, as shown in the foregoing graph, was 

 very disappointing, since the general evidence of the close 

 relationship of these two factors is most clear. Attempts 

 were therefore made to devise a satisfactory way of recording 

 the rainfall so that its relation to the curve of the variation 

 of the lake levels might be graphically demonstrated. It 

 was realized that the most important point to be considered 

 was the fact that the influence of the rainfall on the great 

 reservoir of underground water in the South-East is a 

 cumulative one. 



An examination of the figures shows that when the rain- 

 fall affecting this area is from about 18 to 20 in. per annum, 

 the level of water in the lakes remains practically stationary. 

 From this it may be deduced that the amount of water that 

 is added to the underground supply, from an annual rainfall 

 of 18 to 20 in., is just about equal to the loss of water from 

 this great underground reservoir per annum. This loss is 

 due to some small extent to evaporation, to a very minor 

 extent to wells and pumps, and to a considerable extent to 

 the outlets along the southern coastline, where it constantly 

 gushes forth in great quantities as, for instance, at the 

 Beachport Springs, Dingley Dell X^reek, Ewen Ponds, the 

 Piccaninnie Blue Lake, etc. 



If we regard the underground reservoir of the south- 

 eastern district as a closed system, with an annual outflow 

 of a quantity which w 7 e may call "x," and if we further 

 accept *'x" as the average annual increment of water received 

 by this basin from an annual rainfall of 18 to 20 in., then we 

 are in a position to construct a new graph, taking into account 

 the cumulative effect of either a series of years wetter than the 

 average (which is 18 to 20 in. per annum), or a series of 

 years drier than the average. 



Graphs of this nature are shown in fig. 11, and were 

 constructed as follows : — Selecting an arbitrary point to 

 represent the position for the year 1895 (curve B, fig. 11), 

 the amount by which the 1896 Lowan rainfall is less than 

 19 in. is plotted below the level selected for 1895. Similarly 

 the amount by which the 1897 rainfall is less than 19 in. is 

 plotted as a further downward movement below the level 

 shown for 1896, and so on, so that for each year the difference 

 between the actual rainfall and 19 in. is plotted cumulatively. 

 In fig. 11, the bottom line A represents the actual curve of 

 the variation in the water level at Blue Lake. The line B 



