62 Journal of Agricultural Research voi.xv. no. a 



subsample successively through a4-mm. and a i/20-inch (1.27 mm.) sieve. 

 The fine soils obtained from the latter sifting were in each case com- 

 posited so as to represent the four series of subsamples obtained from 

 each plot, and the composites were then submitted to partial analysis. 

 The results of this study of sampling methods have also been elsewhere 

 reported in detail.* 



The net results for the proportion and amount of fi^ie soil (passing a 

 I /20-inch sieve) in the surface of the two plots, which contain a good 

 many cherty fragments, were : 



Plot 1. Plot 4. 



Fine soil (per cent of air-dry sample) 9°- 39 9i- 59 



Fine soil (pounds to the acre 7 inches) i, 890, 644 i, 865, 947 



The analyses of the several series composites from the same plot show 

 composition differences somewhat greater than appear in duplicate analy- 

 ses of the same composite. These series differences are rarely more than 

 double the differences shown by duplicate determinations upon the 

 same composite. They are not of such magnitude as to approach the 

 differences exhibited by the analytical figures, either detailed or aver- 

 age, for the respective plots. For this reason the several sets of detailed 

 analyses are not presented in full in the present account; but, when the 

 repetitions of a determination upon composites from the same plot are 

 sufficiently numerous, they have been used, by application of the Gauss 

 formula, to calculate the probable error of the determination. 



Because of the exceptional care employed in taking, preparing, and 

 compositing the samples, it is believed that the materials used for this 

 study are in an unusual degree representative of the soil areas from 

 which they were obtained. 



All analytical results stated in this paper are expressed in terms of 

 the air-dry soil, without recalculation to a water-free basis. Deter- 

 minations of hygroscopic moisture made when the analytical work was 

 begun showed a range of 0.785 to 1.088 per cent for the eight series 

 composites, with plot averages of 0.816 and 0.829 per cent, respectively. 

 Another set of such determinations made about the close of the analytical 

 work showed an average increase of only 0.05 per cent of moisture in 

 the soils. 



TOTAI. POTASH 



Portions of the series composites were reduced to an impalpable 

 powder by grinding in an electric mortar mill of agate. The total 

 potash content in each was determined in duplicate by the J. Lawrence 

 Smith method as described by Washington (9, p. 129). The final 

 weights of potassium platinichlorid were ascertained by first weighing 

 the dried precipitate in a Gooch crucible, then washing out the soluble 



1 Frear, William, and Erb, E. S. a study in son, sampling. To be published in Proc. Asset-. Ofif. 

 Agr. Chem. 191 7. 



