20 Irrigation and Drainage 



wide departures from what would have been observed had it been 

 possible to have measured as accurately the water consumed by 

 the surrounding plants of the same kind which were growing at 

 the same time in the field under every way normal field condi- 

 tions. 



In all of these pot experiments, the effort has been to hold 

 the amount of moisture in the soil at a constant quantity equal 

 to that which was possessed by the field soil in the spring of 

 the year, when it was in good working condition ; and this 

 was done by weighing the cylinders periodically, usually as 

 often as once a week, and then adding water in sufficient quan- 

 tity to bring the weight of the cylinder back to the original 

 amount. The cylinders were, of course, water-tight, so that the 

 only loss was through evaporation from the surface of the soil in 

 the cylinders and from the plants themselves. No effort has been 

 made in these experiments to distinguish between the amount of 

 water which actually passed through the plant and was evaporated 

 from its surface, and that which escaped from the surface of the 

 soil in which the plants were growing, as to do this would 

 necessitate the covering of the soil in which the plants were grow- 

 ing so as to prevent evaporation from it. To do this effectively 

 would interfere with the normal aeration of the soil, and thus viti- 

 ate the results by producing abnormal conditions. During the 

 work of the first year, when the wooden casks were used, there 

 was probably some loss of water through the walls of the casks, 

 due to capillarity in the wood and evaporation from it ; but 

 the amount was probably small, because they were all well 

 painted. 



The first year's trials were with oats, barley and corn. With 

 the oats and barley the surface of the soil was not disturbed after 

 seeding, but in the case of the corn the ground was stirred after 

 each watering, so as to develop a soil- mulch after the manner 

 of field culture. In each case the work was done in dupli- 

 cate. In the table which follows are given the results of these 

 trials : 



