USE OF WATER BY SPRING WHEAT ON GREAT PLAINS. 23 



was produced with a total water use of 10.21 inches. Plats on all 

 sides of the one studied yielded much less, although their water con- 

 tent at the time of the determinations was nearly the same. An 

 uneven drifting of snow was noted in the spring of this year, and 

 it is possible that portions of this plat, which was in stubble, con- 

 tained more water at the beginning of the season than the average 

 shown by the soil samples taken. 



The larger quantity of water needed at southern stations to produce 

 a crop is clearly shown in this study of the total use of water at the 

 different stations. For example, at Assinniboine, the northernmost 

 station reported, small yields of grain and straw have been obtained 

 from the use of 4.52 and 4.11 inches of water, respectively. At Het- 

 tinger, a little farther south, the use of 4.32 inches resulted in a total 

 failure of both grain and straw. At Belle Fourche the use of be- 

 tween 6 and 7 inches of water has resulted in yields of grain ranging 

 from nothing to 4 bushels per acre, and in yields of straw a little 

 higher than those produced by the lesser quantities of water at As- 

 sinniboine. At Scottsbluff, Nebr., there has been no total failure, but 

 yields approximating those at Assinniboine have required from 7 to 

 8 inches of water. At the four stations farthest south. Garden City, 

 Dalhart, Amarillo, and Tucumcari, respectively, total failures of 

 both grain and straw have resulted from the use of quantities of 

 water varying from 4.12 to 9.46 inches. At none of these stations 

 has a yield of grain or straw been secured when less than 6 inches of 

 water has been used. Dalhart shows the poorest adaptation for wheat 

 of any station under study. A yield of grain or straw has not yet 

 been obtained there when less than 10 inches of water has been used. 



In a few cases stations have matured a crop on less water than 

 stations north of them, but the figures presented certainly justify 

 the general statement that the quantity of water required to produce 

 a crop of wheat depends upon geographical location and on the Great 

 Plains increases with the distance south except as this is modified by 

 altitude. 



In order to obtain a mathematical expression for the relation that 

 exists between the total use of water and the yield of wheat, a series 

 of studies was made in which the total use of water was correlated 

 with the yield of grain, yield of straw, and total weight of crop 

 produced. All of these correlations have been determined by the 

 product-moment method. 



In these correlations the average yields for each station and the 

 average water use for each station were used as bases for computing 

 departures. As the quantity of water required to produce a crop is 

 influenced by the location, this method of procedure was necessary 

 in order to eliminate station differences. While this method may 

 not be mathematically correct, it at least gives a result that can not 

 be far wrong. It is the only method practicable, since there are only 

 a few stations where there have been enough years to make indi- 

 vidual-station correlations between yield and water use. 



The results of the studies are as follows : 



The correlation between the total quantity of water used and the yield of 

 grain in bushels is 0.76, with a probable error of ±0.03 ; the correlation between 

 the total quantity of water used and the yield of straw per acre is 0.69, with 

 a probable error of ±0.03; and the correlation between the total quantity of 

 water used and the total weight of grain and straw is 0.78, with a probable 

 error of ±0.03. 



