MECHANICAL ANALYSIS OF SOIL. 



37 



separation of all tlie particles greater than 0.05 millimeter in diameter, 

 i. e., all but the silt. This separation is about twice as fast as that 

 calculated on the basis of a specific gravity of 2.6 by one of the numerous 

 formulas for the velocity of bodies falling free in vrater. Only a part 

 of this difference can even be accounted for by a deviation in the specific 

 gravity. We believe the greatest variable is the additional, time required 

 for the careful decantation of the silt and the possibility of grains getting 

 out of suspension by settling against the side of the tube when inclined 

 in pouring. 



The Bureau of Soils specifies brass screens with round perforations 

 for the separation of groups numbers 1 and 2 and silk bolting cloth of 

 proper mesh for the separation of gi'oups numbers 3 and 4. The bolting 

 cloth of course has a square mesh. We used a nest of square-mesh brass 

 screens throughout. A series of measurements along the central wire in 

 each direction was recorded as follows: 



Size of 

 screen 

 (meshes 

 to the 

 inch). 



Mesh dimensions. 



Mesh 

 desired. 



Largest. 



Average. 



30 

 60 

 150 



mm. 

 0.57 

 0.28 

 0.01 



mm. 

 0.54 

 0.22 

 0.08 



mm. 

 0.50 

 0.25 

 0.10 



We believe the square-mesh wire screens preferable to the bolting cloth 

 in that they are more readily and perfectly cleaned. Since the meshes 

 for the two smaller grained sands are of necessity square, the advantage 

 gained in having those for two larger grained sands round is for the 

 greater part lost. 



With care the various grains may be sized satisfactorily, With a 

 sample of 5 grams an ei-ror of 10 miligrams amounts to only 0.2 per 

 cent, yet this perfection of manipulation can be obtained if glazed paper 

 is used, wedged grains are removed from the sieves and all operations 

 are carefully performed. In all this work much can be accomplished 

 by systematizing the operations. Eepeated drying to constant weight 

 may also be avoided. Experiments with both samples and different 

 separations show that a period of three hours at a temperature of 100° 

 to 112° is sufficient thoroughly to dry either. Two hours usually suffices 

 to dry the separated sands. 



The Bureau of Soils * has published the following table of eight anal- 

 yses of 5-gram samples of the same soil with the statement that it "may 



* Loc. cit. 



