Gr. W. Robinson 



311 



fractions, a to e, having velocities less than d/t. The ratio of the concen- 

 tration at depth d after time t to the total concentration at the beginning 

 of the experiment will thus give the proportion of material having 

 velocities less than d/t. By determining the concentration for diilerent 

 values of djt the data are obtained for a summation curve showing the 

 relation between percentage of material and log setthng velocity. 



Experimental. 



The method used consists in allowing a soil suspension of known 

 concentration to settle in a cyhndrical vessel and withdrawing samples 

 for appropriate values of depth/time. By suitable choice of depth and 

 time the concentration and hence the percentage of particles 

 corresponding to any desired velocity can be obtained. 

 Generally speaking a htre measuring cyhnder about 40 cm. 

 in height and 6 cm. in diameter is used. There is of course no 

 necessity to use a graduated vessel: any cyhnder of uniform 

 cross section and suitable dimensions may be used. Samphng 

 of the suspension is carried out by means of a 20 c.c. pipette 

 passed through a cork or shive and adjusted so that when 

 the cork rests on the top of the cyhnder the point of the 

 pipette is at the desired depth below the surface of the hquid 

 (see Fig. 2). The column having settled for the required time 

 the pipette, previously adjusted for depth, is closed at the top 

 with the finger, in order to avoid samphng the upper layers, 

 and lowered very carefully till the cork rests on the top of 

 the cyhnder. The finger is then removed and 20 c.c. of the 

 suspension withdrawn. Every precaution is of course taken 

 to avoid shaking or mixing the layers of the suspension at 

 the point of sampling. With a column of the dimensions 

 mentioned the withdrawal of 20 c.c. causes a fall in level 

 of about 7 mm. This probably represents the extreme error 

 in samphng. It is assumed that the 20 c.c. of suspension withdrawn 

 represents the concentration at the point of the pipette. Probably the 

 liquid comes mainly from above, but to some extent from below this 

 point. A separate experiment with a column which had settled for 

 several weeks and which had formed clearly defined strata, showed that 

 it was possible to pipette to within 2 to 3 mm. of a stratum without 

 disturbance. It will be shown later that an error of a few milhmetres 

 in samphng involves a neghgible error in the final result. Careful manipu- 

 lation is of course necessary in this operation. The 20 c.c. of suspension 



21—5 



Fig. 2. 

 Method of 

 sampling. 



