88 



SOILS: PB0PEBTIE3 AND MANAGEMENT 



quantity of soil is used, the mass is not kept properly 

 agitated; secondly, convex currents are set up in the 

 conical glass tube, which vitiates the results ; and, thirdly, 

 the separate particles tend to coalesce into granules. It 

 is evident that in any separation of soil particles all 

 granulation must be avoided. This is usually accom- 

 plished by shaking or boiling the 

 sample previous to the determi- 

 nation. The tendency toward 

 granulation during the process 

 of separation itself is fatal to 

 accuracy, as compound particles 

 carrying a large number of small 

 grains would fail to pass over 

 at water-current velocities cor- 

 responding to their component 

 parts. 



66. Hilgard's chum elutriator. 

 — The errors of the Schone ap- 

 paratus are obviated to some 

 extent by Hilgard's,^ the prin- 

 ciple of operation remaining ex- 

 actly the same. In Hilgard's 

 elutriator the deflocculated soil 

 sample is introduced into the 

 Fig. 12. — Hiigard's churn base of a cylindrical tube (see 



above ,01 mm. in diameter, rapidly revolving Stirrer. This 



(o]; .lTeeuV(a)%^^^^^ ^^ ^^^^^^^^ ^^ counteract convex 

 chamber; (o), outlet tube. Currents and to prevent the 



1 Hilgard, E. W. Methods of Physical and Chemical Soil 

 Analysis. Ann. Rept. California Agr. Exp. Sta., pp. 241--257. 

 1891-~1892. 



