Four columns wcrc ttsod. Eac11 had en internal cross-sectional arca of 3.8 sq.cm and was filled to 14 cm with rcsm AG $-Xl), zoo to zoo mesh. Each column was provided with a water jacket connected HO that the resin bells tn all columns could be hcatcd srmultancously with ono circulating pump. The AC rcsm WRS obtained from UK+Had Laboratories, 1300 Dclawarc Street., Berkeley, California: it 1s Dowcx r-S8 that has been carefully sieved and washed with acids, bases, and complcxmg agents. The shipment rcccivcd in this laboratory was found to give nnrrower elutron graphs and hence better separations than the resin previouslyl,* used, DOWCX 1-Xx0, 100 to zoo mesh. The fraction collector, pi meter, spcctrophotomoter, and most of the reagents have been dcscribedl3. The eluant dcsignatcd as 0,250M potassium chloride of pll 9.00 was actually 0,240M with potassium chloride and o.oxoM with ammonium chloride and contained sufftcicnt ammonia to give a PH of g.oo. Stmilarly, a mixture of o.4goM potassium chloride, o.o~oM ammonium chloride and ammonia IY designated as o.gooM potassium chloride of prr 9.00. Guanidinium tctraphosphate monohydratc was prepared by the method of QUIMBY~ except that the quantity of formamide used was 2003, greater than recommended by Qur~n~.The analysis of this product is given in Table I. Tbc Graham's salt was prepared by TRISTLLR * 4, who also found its number-average degree of polymerization to be 190.