278 On the Rain and Drainage - Waters at Rothamsted. 
the near approachment of the drainage from the 20-inch and 
6()-inch gauges during this time, but it fails to explain the 
fact that the 40-inch gauge has during the same period furnished 
so much more drainage than either of the others. As these 
differences do not seem to be adequately explained by the con- 
siderations above referred to, we must either assume some defect 
in the 20-inch gauge, leading to a leakage outwards, and result- 
ing in a deficient drainage, or some defect in the 40-inch and 
60-inch gauges, leading to an increased, and in the case of the 
40-inch gauge to an excessive drainage. In regard to these 
points it should be stated that the only really known leakage 
was in the 20-inch gauge, already referred to as occurring in 
February 1879 ; but this leakage was from the outside inwards, 
whereas it must be a leakage from the inside outwards that 
would help to explain the discrepancies in question. The 
present relatively excessive drainage of the 40-inch gauge is 
not confined to wet seasons, though it is then greatest, but it 
appears also in dry years, and even during the spring and 
summer months. Thus the average drainage from March to 
August during the last six years was respectively 5'07, 5"68, 
and 5' 18 inches for the three drain-gauges. 
On a consideration of all the facts it would seem that the 
drainage from the 40-inch gauge has been relatively somewhat 
excessive during the last six years ; the excess is however less 
marked at present than it was two or three years ago. The 
amount of drainage from the 60-inch gauge has also apparently 
been in some excess, but in a less degree, during the same period. 
An excessive drainage can clearly only occur by leakage of 
water from the outside through some defect in the walls. It is 
evident that from the position of the 40-inch gauge there may 
be at times a greater pressure of water on two of its sides than 
there is within the gauge, due to the sloping edges of the walls 
surrounding the gauges delivering an excessive amount of rain 
on to the two thin bands of soil which separate the gauges (see 
Fig. 2), and the thickening of the walls, which took place towards 
the end of 1874, would increase this side pressure. It must 
be frankly confessed, however, that it is difficult to believe that 
leakage would be more marked after the thickening and cement- 
ing of the external walls than it was before. 
Notwithstanding the difference in the relation of the gauges 
at different periods of the experiment, it is pretty clear that at all 
times evaporation has been somewhat greater in a dry season 
from the soil of the 60-inch gauge than from the soil of the 
20-inch gauge, and that consequently capillary attraction has 
proved capable of bringing water to the surface from a depth 
exceeding 20 inches. The excess of evaporation on the 60-inch 
