150 Probable Erri>r of SdiiiplliKj in Soil Surveys 

 Field B. Sedentary Soil. 



Practically the same values for ]x, might have been obtained if 

 analyses had been repeated several times for each boring and averaged. 

 The values of p^ would then have been greatly reduced and consequently, 

 as in the case of the moisture and organic matter, 7>2 could be taken 

 as approximately equal to P. 



In view of the great time taken in making soil analyses it was felt 

 that the increase in the accuracy of the values of p^ which would thus 

 be obtained would scarcely be great enough to justify spending four 

 or five times as long over the laboratory work. Also, no time was 

 spent in doing chemical analyses other than the determination of i'o^^s- 



It is seen from the above tables that in the case of Field A, which is 

 too variable for ordinary survey purposes, p^ is as great as 34-5 per 

 cent, for fine gravel. Now this constituent is comparatively unimportant 

 in characterising a soil unless it 2^re<^ominates over other fractions. 

 An accurate knowledge of the hygroscopic moisture is also, for survey 

 purposes, of no great importance. Of the other dctiM-minations the 

 probable field errors of the fractions in the mechanical analysis are 

 about 10 per cent., while that of the P2O5 is 13-'J per cent. 



If therefore we count on a probable error of 1.3-1 j)er ccut. for tliis 

 soil this can be taken as a mii.ximum value. Similarly 7-."5 percent, can 

 be taken as the ma.ximum value of the field error for I<"ieKi B which is 

 a sedentary soil. Field A is as stated above of too variable a character 

 for ordinary survey purposes. If therefore operations are conducted on 

 the basis of a maxiumm field error of 10 per cent, it will probably serve 

 for survey work. 



The value of the total probable error of the final result can now be 

 reduced to any desired value by increasing the number of borings and 

 the number of determinations. 



