22 BULLETIN 1340, IT. S. DEPARTMENT OF AGRICULTURE 
germination until the plants are well above the ground. From that 
stage to the heading-out stage, there is a gradual increase in water 
requirements and a more rapid increase beyond this latter stage up 
to the time of the beginning of the milk stage of the grain. Beyond 
the hard-dough stage there is a sudden decrease in the quantity of 
soil moisture needed and this lessening requirement for moisture 
continues until the grain is harvested. 
This is shown in diagrams "A" and " B " of Figure 11, which 
represents the results of tank experiments by Fortier and Petersen 
at the Nevada Agricultural Experiment Station farm near Reno 
in the summer of 1914. Diagram "A" represents the mean of six 
experiments in which Marquis wheat was grown in tanks. 
In "A" the stepped line shows the quantity of water used at the 
various stages of growth. Every three or four days water was 
added to each tank to replace that used in transpiration and evapora- 
tion from the soil surface. Accordingly each horizontal step repre- 
sents the average quantity of water used by all the tanks during the 
period covered since water was last added. As nearly as could be 
determined, the ratio of the loss due to evaporation and transpira- 
tion was as 1 is to 2.44. Each tank on an average received 359 
pounds of water and about 71 per cent of the total was due to trans- 
piration. The tanks were fitted with false bottoms so that all deep 
percolation water was withdrawn through a tap and the quantity 
deducted from that applied to each tank. Soil moisture deter- 
minations were also made at seed time and harvest time and the 
gain or loss of soil moisture included in the final results. In like 
manner due allowance was made for rainfall. 
Diagram " B " represents the mean of five tank experiments for the 
purpose of determining the water requirements of barley at various 
stages, of growth. These experiments were conducted simultane- 
ously with those represented by Diagram "A" and the meteorological 
data are the same for both. 
The Utah Agricultural Experiment Station has demonstrated 
repeatedly (1) that if cereals can be watered but once in their period 
of growth, this one irrigation should be applied about the time of 
the early booting stage so as to supply ample moisture for the plant 
stage requiring most water; (2) that water applied after the dough 
stage is reached decreases the yield and tends to lodge the grain, and 
(3.) that the largest yields are obtained by maintaining an adequate 
moisture supply in the soil by the necessary number of irrigations 
until the soft-dough stage of the cereals is reached. The graphical 
representations of Figure 11 seem to give a reason for these demon- 
strated facts. 
Thorn and Holtz, in growing plants in tanks under irrigation at 
Pullman, Wash., during the seasons of 1911 and 1912, found that 
the daily quantity of water transpired by wheat, corn, oats, and peas 
increased until about the beginning of the ripening period and then 
decreased until harvested. The daily quantity of water transpired 
by each of these plants, expressed in pounds, is shown in Figure 12. 
The same authors found that in growing wheat, corn, oats, and 
peas on plots adjacent to similar control plots which were uncropped, 
the averaged ratio of the quantity of water transpired to that evapo- 
rated from bare soils was approximately as 3 is to 1. The rainfall 
