THE PHYSIOLOGICAL ROLE OF WATP^R IN PLANTS. 5 



drains from a given surface of land and comparing it with the water 

 which falls upon the same surface, as shown by the rain gauge. Knowing 

 the amount of drainage and the water content of the soil, the amount of 

 evaporation may be calculated. In order that this method may be reli- 

 able it is necessary that no other water than that falling directly upon 

 the surfiice be measured, also that all tlie water of filtration be collected 

 by the drains. There is one source of error — j)oorly drained soils absorb 

 less water than well drained. The water will penetrate only wliere the 

 condition of the interstices allows it to descend and displace the air 

 filling them. Now, it has been demonstrated that drains increase the 

 number and diameter of interstitial spaces. The drains ofler a means 

 of escape for the air which the water displaces and afterwards in a 

 measure suck out the water of the soil by virtue of the vacuum which 

 tends constantly to form at their upper ends. 



Experiments were made byRisler on soil cultivated in wheat, clover, 

 alfalfa, and potatoes. The gauging of the drain which collected all the 

 water was made daily at noon. The number of seconds required to 

 fill a vessel of 4 or 5 liters was determined and it was assumed that 

 the amount of water collected per second remained constant during the 

 preceding 24 hours. This is only exact when the drains collect a dimin- 

 ishing amount of water for a number of days. Whenever there are 

 heavy rains a number of measurements should be made daily. Eisler's 

 tables show that the average evaporation amounted to 70 per cent in 

 1867 and 1868. Maurice de Geneve found it to be 61 per cent and Gas- 

 parin 88 per cent. These figures are comparable, and show that evapo- 

 ration amounts to about 70 per cent of the rainfall. These experiments 

 should be carefully repeated with different crops upon soils of different 

 chemical constitution, with differing inclination and exiwsure, as well 

 as upon bare soil. The effect of heavy rains and long-continued gentle 

 rainfall should also be compared. It is probable that very diflerent 

 results might be obtained which might have an important bearing on 

 determining the quantity of water to be subsequently distributed by irri- 

 gation. Just here, at the outset, there is a serious lack of precise data. 



Experiments of this character are worthy of especial consideration 

 by agricultural experiment stations able to conduct them on an exten- 

 sive and varied scale. The results will be of still greater value if 

 advantage is taken of recent investigations relative to the influence 

 of fertilizers upon the movement of soil water. Numerous investi- 

 gators, especially Lawes, Gilbert, and Warington,' Frankland, Ber- 

 thelot,^ and Deherain,' have undertaken to determine from examina- 

 tions of the drainage water the time required to render soluble the 

 chemical principles in fertilizers, the optimum economical quantity of 

 fertilizers to be used, the loss of nitrates iu the soil, etc. There still 



' Warinjiton, .lour. Roy. Agl. Soc, 17 (1S81), p. 241. 



2Coiiipt. Reiul., 105 (1«87), p. 690. 



sDehdrain, Ann.Agron., 16 (1890), p. 337; 17 (1891), p. 49. 



