60 BULLETIN 835, U. S. DEPARTMENT OF AGRICULTURE. 



zontally an average distance of 12 inches before it can move 

 downward. 



The detailed measurements will not be given, but after 60 days the 

 water in the gauge glass on the outside of the flume showed water 

 up to a point within 11 inches of the surface of the water in the tank; 

 that is, after 60 days that part of the tube below the point desig- 

 nated " Gage E " in gauge glass was completely saturated. After 

 the sixtieth day, the rate at which the water rose in the gauge 

 glass was very slow, and upon the "seventieth the experiment was 

 terminated. 



This experiment, as did the previous ones cited, gave free water 

 as a result of capillary action. 



Three additional experiments were run with the same tube, but 

 containing soils of a different type. In each case the same result was 

 obtained, except that they were terminated sooner and for that reason 

 the water did not rise so high in the glass. 



Finally, it may be stated that in every flume, covered and open, 

 that, was inclined downward at an angle from 15° to 45° free water 

 was developed when the experiment was run for a sufficient time. 

 In only 3 or 4 instances out of the 20 or more flumes so inclined 

 were the experiments terminated before free water was dripping 

 from the lower ends of the flumes. 



Several tests were made of the amount of water taken up from 

 the tanks and delivered again at the lower end of the flumes as free 

 water. One of these tests will be given. 



The flume selected is No. 95, containing the lava soil from Idaho. 

 This flume was covered, inclined downward from the horizontal at 

 an angle of 30°, and was 15 feet in length. The records show that 

 the flume commenced dripping water at the lower end on February 

 25, 1917. Commencing with March 1, the quantity of water lost 

 from the tank by the wick was 18 liters. During this same period 

 there was caught at the lower end of the flume 8.78 liters, or approxi- 

 mately 50 per cent of the quantity taken from the tank. The water 

 was caught in a can as it dripped from the flume. 



It has been suggested that a true siphon might have been formed 

 as a result of " soil puddling " or other natural mechanical means. 

 It did not occur in many cases and it is doubtful if it occurred at all. 

 It is found, for instance, that with the use of clean, coarse build- 

 ing sand, devoid of clay or other fine material, the same result is 

 obtained. However, to test this point further, a system of ventila- 

 tion within the wick was installed. 



Ventilating pads were made out of ordinary window-screen wire. 

 From six to eight thicknesses of wire were rolled into a very small 

 diameter and then flattened out. This made a pad of wire about 



