36 PRATT. 



given in this paper had been passed through a l-millimeter screen and 

 therefore group number 1 was calculated into the analysis from the 

 figures for detritus on a 1-millimeter screen; therefore, it includes all 

 material above 1.0 millimeter in diameter. There is only a negligible 

 difference between this and the "fine gravel" group of the analyses of 

 the Bureau of Soils; in fact, in a number of our analyses the group is 

 absent entirely and it is probable that in the others all of this group 

 would have passed a 2-millimeter screen. 



Other slight variations from those specified in Bulletin No. 24 of the 

 Bureau of Soils were used and are of sufficient interest to justify discus- 

 sion. Our shaker is an ordinary laboratory machine run at 90 throws 

 per minute instead of 100 and for this reason the shaking was continued 

 slightly longer. In the centrifuge modeled after that described in Bul- 

 letin No. 24 we use brass tubes turned to a smooth surface inside instead 

 of glass. Brass tubes are not subject to the occasional breakage that 

 occurs with the test tubes, even though the latter are well annealed. 

 They are quite as easily cleaned, but are open to the objection that they 

 are not transparent. 



The time or rotation of the centrifuge necessary to settle a particle 

 of a given diameter depends on the speed of rotation and the distance 

 of the end of the centrifugal tube from the shaft axis. Our centrifuge 

 rotates at 700 revolutions per minute at full speed and the distance of 

 the end of the centrifugal tube from the shaft axis is 22 centimeters. 

 The centrifuge is operated until an examination with a microscope shows 

 no particles above a certain average size and shape in suspension. Freak 

 grains or scales must be ignored and the average largest grain taken. 

 Three minutes is the average time required for the separation of all but 

 the clay group. Time is counted from the time full speed is attained, 

 fifteen seconds after the current is turned on. Once determined, the 

 time of centrifugating varies from a constant only as the specific gravities 

 of the particles of a given size vary, which is little. Occasionally, espe- 

 cially for the first period of rotation of a sample of heavy clay, a little 

 additional time is necessary. 



This is quite natural since much more clay is in suspension at first 

 than will be present after the first decantation has been made. When 

 much clay is in suspension it is probable a condition exists approaching 

 that which Richards ^ terms a "hindered settling condition," i. e., a grain 

 which tends to settle may be interfered with and held up by its floating 

 companions. In subsequent rotations, when most of the clay is gone, 

 such a grain settles unobstructedly. 



About thirty seconds of rotation in the centrifuge is necessary for the 



' Ore Dressing. New York, 2 ed. ( 1908) , 1 , 464. 



