V. L. Whittles Kjti 



shown with great advantage. In this connection the u.se of ultra-filtration 

 methods might give valuable information. 



The curves for the fighter soils rise steeply and then flatten out, those 

 for the heavier soils are flat at first and rise steeply later. Two particular 

 appfications of these curves are of importance : 



{]) Two soils, analysed according to different systems of grouping of 

 particles may be compared exactly. 



(2) The analytical results may be transformed from one system to 

 another by reading oft' the values at the points at which the curve cuts 

 the mean diameter of the selected classes. 



Baker (7) has devised two values for describing a soil from its mechani- 

 cal analysis plotted in this way. This is of value for catalogue and descrip- 

 tive purposes, but is not so well adapted for the preparation of drift 

 maps as is the triangular method. In this method only three clas.ses of 

 particles can be considered, which may be conveniently termed Coarse, 

 Medium and Fine respectively. The question now arises as to which two 

 limiting values shall be chosen for the separation of the Medium class 

 from the Coarse and Fine grades respectively. Considering Fig. 1 we 

 find that: 



U.S. Bureau of Soils take -0:3 and •()02!'). 

 Wilsdon(58) takes -02.5 and -Odl. 



It is here suggested that, taking into consideration the known 

 properties of the various grades, the curves are best characterised by 

 their intercepts on the lines for -12 and -006. On this basis we have, on 

 the usual English system: 



Coarse (fine gravel + coarse sand) 

 Medium (fine sand + silt) 

 Fine (fine silt -1- clay). 



Triangular Methods. 



In the method adopted by the U.S. Bureau of Soils the proportions of 

 clay and silt are plotted along each of two axes at right angles. By joining 

 the 100 points on these lines a right-angled triangle is formed. This is 

 then conventionally divided up into compartments as shown in Fig. 2. 

 The scheme of classification is given in Table III. 



Wilsdon has suggested a modification of this method whereby the 

 proportion of the third constituent — the sand — may be read off directly 

 from the diagram. The percentages of sand, silt and clay are plotted on 

 an equilateral triangle. The detailed procedure is given in connection 



12—2 



