For many purposes it is desirable to have a method for the determination of the cation exchange capacity of clay minerals based on principles as free from ambiguity as possible. Most of the many methods previously proposed have the drawback of entailing a great deal of work if one is to make sure that the mineral is completely saturated with the reagent in use, or that an excess of the reagent is absent. What might be called the complete chromatographic method by its very nature does not suffer from this difficulty. A known weight of clay is suspended on an appropriate inert filter aid (such as Gooch asbestos), packed into a glass tube, and the exchangeable ions eluted with a solution of cesium chloride traced with radioactive Cs 137. The effluent is collected in a series of appropriately sized samples, and the specific radioactivity of each of these determined. When the activity of the effluent has risen to that of the influent solution, one is certain that complete exchange has taken place. A simple computation gives the total cesium holdup of the column, and if the column was initially dry, the exchange capacity of the clay sample.

As a routine method, particularly when results of the highest precision are not needed, the effluent may be collected in only two samples, one large sample and finally one small sample to check for complete exchange. This procedure requires a minimum of work and attention but suffers from the disadvantage that the result depends upon the difference of the two counting rates measured and is thus subject to greater uncertainty.

These methods have been compared with the Bower-Truog method (Ind. and Eng. Chem., Anal. Ed., 12, 411 (1940)) on Chambers, Arizona, montmorillonite, a nontronite, and an attapulgite. Generally the Bower-Truog method gave more variable and somewhat lower results. Thus for the Chambers montmorillonite the Bower-Truog method gave 0.97 meq./gm.; the chromatographic method, 1.054 meq./gm.

The method has been further tested by initially saturating columns with manganous ion, determining the weight of solution held up in the column, and then eluting with traced CsCl. Effluent Mn was determined colorimetrically and the Cs in the Geiger counter. The results agreed within experimental error for the following clays of the API series: montmorillonites H19, H23, H24; illite H36; and nontronite H33b.