CHAPTER XXIV. 177 



ration saved, plus the rain penetration. This increase amounted to 3.4 

 per cent., with a rainfall of over 1% inches. The average difference 

 between the sod cover and the cultivated soil was 10i/4 per cent, more 

 moisture in the latter. Now, as penetration and evaporation taken 

 together amounted to only one-third of this difference, it is very 

 certain that evaporation alone could only explain a very small portion 

 of the difference. In fact, it is of interest to note that evaporation may 

 not have reduced the percentage of moisture at all. This would be the 

 case if capillary action was sufficiently active to bring up all tne water 

 needed by the grasses, and there was plently of sub-water to be brought 

 up — conditions that seem to have existed during the test. If the 

 supply is good, a pipe does not show less waste because the faucet at 

 the end is open. 



At Rothamsted, the place previously spoken of, there are drain 

 gauges for observing the proportion of rainfall that percolates through 

 the soil. Determinations are made of the quantity passing at depths 

 of 20, 40 and 60 inches. The 20 and 60 inch gauges are in similar soils. 

 The 40-inch is in sub-soil of more gravelly nature, and yields from 2I/2 

 to 5 per cent, more water than either of the other two. Observations 

 conducted for twenty years show that an average of 47% per cent, of 

 all the rainfall percolated through 20 inches of soil, and 44.9 through 

 60 inches. The rainfall was measured on the 1-lOOOth of an acre 

 gauge. 



As the Rothamsted records are classic in the literature of rainfall 

 and percolation, a description of the gauges used may be found inter- 

 esting. 



The funnel portion of the one-thousandth of on acre gauge is con- 

 structed of wood lined with lead, the upper edge consisting of a vertical 

 rim of plate glass beveled outward. The rain is conducted by a tube 

 into a galvanized iron cylinder, and when this is full it overflows into 

 a second cylinder, and so on into a third and fourth and finally into 

 an iron tank. Each of the four cylinders holds rain corresponding to 

 % an inch of depth, and the tank an amount equal to 2 Inches. Each 

 cylinder has a gauge-tube attached graduated to read to two-thou- 

 sandths of an inch and by estimation to one-thousandth. Smaller 

 quantities are transferred to a smaller cylinder with a gauge-tube grad- 

 uated to one-thousandths. 



The three drain gauges were constructed by digging a deep trench 



along the front, gradually undermining at the required depth and put- 



12 ting in perforated cast-iron plates to support the mass of soil. The 



