12 SUEFACE- AND SUB-IRRIGATION 



water was used, capillary attraction was not sufficient ; on the 

 other hand it soaked through the clay and even moistened the 

 hard, dry walks on either side. The bed containing the water- 

 tight bottom worked perfectly from the first. This, therefore, 

 showed that it was not altogether the fault of the soil. 



To prove the above further, we took a section in the center 

 of the bed on the ground, five feet in length and eight feet in 

 width, — the width of the bed. It was separated from the rest 

 of the bed by board partitions, at the bottom of which notches 

 were cut to accommodate the three rows of tile. We removed 

 an inch or two of the soil from the bottom of this section, re- 

 placing it with cement. We also cemented the sides and 

 around the notches in the partitions, thus making the bed 

 practically water-tight. This section was watered with the 

 same flow of water as the rest of the bed, the only difference 

 being that when the water was emitted into the soil, it could 

 not soak through the bottom. The result in every instance 

 was that if sufficient water was applied for the cemented part, 

 the remainder of the bed received very little. 



On the other hand if enough water was poured in the tiles 

 for the two ends of the bed, the cemented portion became mud- 

 dy. It is evident that if the sub-irrigated beds contain a tight 

 bottom the saving of water is very great ; but with some soil 

 bottoms, a greater percentage of water is lost by soaking through 

 the ground than could possibly have gone off' through evapora- 

 tion in surface irrigation. 



EXPERIMENTS OUT OF DOORS 



Knowing the results of this experiment in the greenhouse, and 

 the good results reported from other«sources with this system out 

 of doors, we arranged our grounds for testing it. Realizing that 

 celery is a crop that demands water on high land, we arranged 

 to water it in rows in this method, the tiles to be placed at a 

 depth to be missed by the plow. 



Good results were secured from this system from the grow- 

 ing plants, but the saving of water was not nearly as great as 

 claimed by other experimenters. On examination we found that, 

 as in the greenhouse bed without the water-tight bottom, a 

 great quantity of water escaped deep in the soil and ran oft" in 



