Vol. 63.] THE MOTION OF SUB-SURFACE WATER. 99 



likely to come into operation in the case of the sujb-surf ace flow ; — 

 in the one case — that of surface-flow — the predominant factors are 

 the declivity and relative durability of the strata passed over ; and 

 in the other, the controlling factors are the difference of pressure 

 along the lines of flow, the varying texture of the strata traversed, 

 and the disposition of contiguous impermeable strata : nor does the 

 theory find full confirmation in the field. If we consider the 

 hypothetical case of the percolation of water into a conical mass of 

 rock of uniform texture, porosity, and absorptivity, resting upon a 

 horizontal bed of impervious material, we perceive that, succeeding 

 a period of inflow the percolating water will arrange itself so that 

 its external form will be that of a somewhat lesser cone with a 

 rounded vertex, and with sides sloping less steeply than those of 

 the cone of rock. 



During a period of excessive percolation, the slope of the surface 

 of the water-cone will more nearly approximate to the external 

 configuration of the rock-cone ; but, during a period of deficient 

 percolation, the outflow will reduce the quantity of the stored 

 interstitial water, and the slope of the water-surface will tend 

 more and more to horizontality. Consequently, the areas enclosed 

 by the higher contour-lines will steadily diminish and perhaps 

 disappear, while the areas enclosed by the contours of lower level 

 will gradually broaden, or at least be maintained almost constant, 

 by the inflow of the water from the higher levels ; also, in this 

 ideal case, the contour-lines will be uniform curves, more or less 

 parallel to each other and to those of the external surface-contours. 

 If several such cones be grouped contiguously in a cluster, or 

 arranged in a chain so that they interpenetrate one another in 

 the region of their bases, then the interstitial water within the 

 region of interpenetration will be banked up and stand at a slightly- 

 higher level than it would stand in the corresponding area of the 

 individual cone, and thus, instead of forming parallel or similar 

 curves, the contour-lines will now be distorted in the region of inter- 

 penetration. From considerations such as these, it follows that the 

 periodic variation of level will be in general of greater magnitude 

 in regions of high water-level than in the areas of low water-level in 

 the same strata ; and this is borne out by field- observations, as the 

 accompanying Table (XII, p. 98) shows. 



This seasonal variation of level will be greatest between a 

 high water-level period, following a season or seasons of excessive 

 flood, and a low water-level period, following a period of con- 

 tinued drought ; but it is also of considerable magnitude, even 

 in yearly periods, in high-level districts, as Table XIII (p. luO) 

 shows. 



This record, in no wise exceptional for a high-level well, is 

 directly dependent upon the incidence of the available rainfall, 

 which is the annual rainfall for the district after deducting the 

 losses due to evaporation and run-off, losses which vary considerably 

 in the four quarters of the year, as Table XIV (p. 100) shows* 



