58 Mineral Nutrition of Plants 



determined experimentally by activity measurements. This is not so 

 serious a weakness as might be supposed, since any future kinetic 

 treatments of ionic transfer between soil and plant will inevitably need 

 both a firm point of departure and a practical experimental method for 

 observing macro changes with time. Both are now available. 



Some simplifying assumptions will be made regarding the nature 

 of the soil systems. Sparingly soluble salts such as calcium carbonate, 

 tricalcium phosphate, and calcium sulfate will be omitted from con- 

 sideration. The arguments used below can readily be modified to 

 include them. The soil system will be assumed to consist of (a) an 

 inert skeleton, (b) negatively charged soil colloids holding cations in 

 a diffuse ionic atmosphere, (c ) soluble salts, and (d) water. 



Consider now a large mass of such an idealized soil at a given mois- 

 ture content. Suppose that a small amount of solution is forced out by 

 pressure, the removal causing no appreciable change in the original 

 soil. Then the relationship between the concentration of soluble elec- 

 trolytes in the solution thus expressed, and in the soil mass, is governed 

 by the Donnan equilibrium. It can therefore readily be determined. 

 This means that we can define the conditions under which the con- 

 tribution of the ionizing colloids becomes important and can evaluate 

 the relationship between the soil solution as defined in this paper, the 

 "Burdian" soil solution, and the solution expressed from the soil. 



To do this we must assume that single cationic activities have mean- 

 ing and are measurable and must use certain simplifications in order 

 to get an approximate over-all picture. 



Consider potassium-saturated soil in the presence of potassium chlo- 

 ride solution and let a,- represent the activity of the potassium ions 

 associated with the soil colloid, a Vt the activity of potassium ions in the 

 "Burdian" soil solution (i.e., due to potassium chloride alone), and a E 

 the activity of potassium ions in the expressed fluid which is free from 

 soil colloids. According to the Donnan equilibrium we then have 



(a c + a B )a B =z (a E ) 2 (1) 



assuming that chloride ions have the same activity coefficients as potas- 

 sium ions. What we need is the relationship of a Vj to a n when a c takes 

 on different values in relation to a v ,. It can easily be seen that for finite 



