Ricuarpson—Lines of Flow of Water in Saturated Soils. 313 
outside, and of exactly half the linear dimensions. This must be 
done with a sharp, narrow-bladed knife. 
A section of this solid by a plane through the axis is shown in 
the annexed diagram. ‘he faint lines would be contours at equal 
intervals of the stream-function and potential if the flow were 
parallel to the axis. The black lines are contours of the actual 
stream-function and potential, found by hand-sketching until each 
black chequer bore the same proportion of length by breadth to 
the same ratio in the pink chequer underneath. (See the paper 
referred to above.) Asa matter of fact, the proportion is nearly 
unity. On measuring seventeen pairs of chequers the quantity : 
width of black chequer across flow f length of faint chequer 
length of black chequer width of faint chequer 
was found to have a mean of -92 (and a standard deviation -10). 
This implies that if the drop in potential per contour is the same 
for faint and black, then the flow in every black tube is ‘92 of the 
flow in every faint one; and as there are twelve black tubes, the 
total flow is equal to ‘92 of that through twelve faint ones. Now, 
if the difference in level of the black equipotentials at the points 
where they crop out on the air-surface be taken as a unit of s, 
then by equation (5) the difference of potential is Kye per black 
contour. And therefore zs per faint contour, as the figure 
shows. Further, twelve faint tubes form a cylinder of 4°35 radius 
measured in the said units. 
.. ‘92 (flow in 12 faint tubes) = total flow from the cavity 
= '927(4°35)? . ae cubic units per second. 
Now there are 7 x 2? square units of water-surface in the cavity. 
So that the surface will fall at the rate of ‘92 x aoe ae 
units per second. But there are 8 units in the height of the block. 
Therefore it will fall 
92 x 3 981. K x 
height of block 
8 
= 16,200 x K x (height of block) per minute. 
That would be one way of measuring K. Errors will arise owing 
per second 
