THE SOIL PAliTlCLE 91 



any vessel that is relatively deep will do for the deter- 

 mination. The larger particles, or sand grains, will of 

 course settle first, and the finer silts and clays may be 

 dec-anted off. As the sands carry finer particles down 

 with them, the suspension and subsidence must be re- 

 peated a number of times. The finer particles, separated 

 thus and decanted, may be further subdivided in the 

 same manner. The time necessary for such decantation 

 as will leave in suspension only particles below a given 

 size is determined by the examination of a drop of the 

 suspension under a microscope fitted with an eyepiece 

 micrometer. In this way the size of the particles decanted 

 may be accurately measured. 



The three steps in this method of separation are : 

 deflocculation of the sample; separation by successive 

 subsidence and decantation; and evaporation to dryness 

 of the separates and their calculation to a percentage 

 based on the original sample. The method, however, is 

 slow, as the time necessary for each subsidence of the 

 finer particles is very great and the number of individual 

 subsidences is large. Neither is the method capable of 

 the refinement of separation which is possible with cer- 

 tain of the elutriators. As a consequence it has been 

 superseded by methods that utilize centrifugal force for 

 the finer separations while retaining gravity for removing 

 the various grades of sand. 



69. Atterberg's modified Appiani l silt cylinder (Fig. 

 14). This method 2 is similar to the beaker method in 



1 Appiani, G. Ueber einen Schlammapparat fiir die Analyse 

 <i<T Boden- und Thonarten. Forsch. a. d. Gebiete d. Agri- 

 Physik, Band 17, Seite 291-297. 1894. 



- Atterberg, A. Die Mechaniscke Bodenanalyse und die 

 Klassifikation der Mineralboden Schwedens. Internat. Mitt, 

 f. Bodenkunde, Band II, Heft 4, Seite 312-342. 1912. 



