Apr. 1. 1918 Capacities of Soils for Irrigation Water 5 



application of water. Also the chances are 823 out of i ,000 that increases 

 of 0.06 inch per foot are due to the irrigation, and 957 out of 1,000 that 

 increases amounting to 0.09 inch were caused by irrigation.^ 



The probable error of the percentage of water accounted for can be 

 only estimated since the water measurement data at hand do not make 

 possible an accurate determination of the probable error of this factor. 

 Moreover, it is dependent on uniform lateral surface distribution, which 

 is, in fact, seldom attained. The field observations upon some of the 

 tracts warrant the conclusion that lack of uniformity in lateral distri- 

 bution ^ is responsible for the apparent discrepancies in the percentage 

 of water accounted for. 



A further test of precision was attempted by making three borings at 

 one time, each within a distance of 6 feet of the other two, and calcu- 

 ating the mean deviation of each determination from the average of the 

 three. Thirteen sets of such determinations to a depth of 6 feet gave an 

 average deviation of ±0.62 per cent of moisture, an average probable 

 error for a single observation of ±0.66 per cent of moisture, which is 

 approximately one-half of the probable error found in connection with 

 six determinations scattered over an entire field representing an area of 

 approximately 5 acres. 



RESULTS ON TYPICAL ALFALFA FIELDS 

 EXPERIMENTAL CONDITIONS 



The soils of these fields are divided into four general classes, notwith- 

 standing some variation in physical properties of different types of soil 

 within one class. While examining the data presented in Tables III and 

 IV, reference should be made to Table I, which contains the volume 

 weights and moisture equivalents (5) of the soils considered. It is to be 

 noted that the volume weights vary from i.io to 1.75 and that the silt- 

 loam soils are lighter than the clay loams and clays. These observations 

 are not in accord with the ideas generally entertained concerning the 

 relation of volume weight to soil texture. 



' It must be remembered that, because of variations in soil, number of borings, and other less important 

 factors, these values of the probable error do not apply equally to.each type of soil given. Where the number 

 of borings differs greatly from 114 (the average used) a more accurate value of the probable error may be 

 obtained by multiplying 0.03 by the ratio of the square roots of the mean, 114, and the number of borings 

 made in a given type of soil. For example, the averages for the Willows experimental tract given at the 

 bottom of Table IV are based upon 284 borings. The probable error of the increase is therefore ±0.019 

 inch per foot of soil. The minimum difference observed is 0.04 inch. The ratio of this difference to the 

 average probable error of the difference is 2.1, which means that the chances are 843 out of 1,000 that the 

 minimum difference was due to the irrigation. 



' It is recognized that the degree of uniformity in lateral distribution which can be attained frequently 

 determines the quantity of water applied per irrigation. 



