318 Mechcuiical Anali/six of Soils and oflirr Dixpersioiu 



1000 mins. respectively. The log f abscissae refer of course to the 

 summation curve AB. and tlie depth abscissae to the concentration 

 curves. 



It will he seen that the change in concentration witji depth after 

 any given time is very gradual below the first few centimetres. Thus, 

 after ten minutes, the concentration in the case illustrated only changes 

 about 2-5 per cent, of the total concentration between 10 cm. and 

 20 cm. With longer times the gradient becomes rather steeper. Errors 

 of the order of a few millimetres in depth of sampling have thus very 

 little effect on the concentration obtained. This of course only holds 

 so long as the material under experiment has a fairly smooth summation 

 curve. With a material having an irregular type of curve, depth errors 

 might be more serious. 



Similar considerations can be developed for the time and tem- 

 perature error. 



The errors introduced by the above variations in working conditions 

 though scarcely negligible are nevertheless not serious when the character 

 of the material is considered and it may be doubted whether they are 

 of significance from the point of view of the genetics and physical pro- 

 perties of soils and clays. The Oden method is of course less exceptionable 

 from the point of view of delicacy and must be used where critical data 

 are required. The method described in the present paper could of course 

 be made to give results strictly comparable among themselves. It is 

 however desirable to have a method which can admit of some latitude 

 to suit the convenience of individual worlvcrs and which has to that 

 extent and within reasonable limits the character of an absolute method. 



Agreement with Re.sults obtained by the Standard Method. 



In view of the enormous number of results accumulated by the older 

 method, it would be a doubtful advance to suggest a new method, even 

 though more convenient and accurate, if the results obtained by it could 

 not be used for comparison with the older results. A considerable number 

 of determinations were therefore made by the new method and the 

 results compared with those already obtained by the old method. In 

 the following table the residts of this comparison are given. For con- 

 venience of statement only three values are given, namely, clay, fine 

 silt, and silt. The same method of dispersion was used throughout except 

 in the case of certain clays, which contained practically no organic matter 

 and were dispersed in -02.5 per cent, sodium carbonate. 



