EXPERIMENT STATION BULLETINS. 615 



1. 10.50 3. 9.60 5. 9.30 



2. 10.15 4. 9.80 



The same solution was then titrated using phenol red as indicator and 

 titrating- to a reaction of 7.4 witli Clark and Lubs standard I'olloAving the 

 technic to be described later. The following results were obtained : 

 1. 7.75 2. 7.80 3. 7.80 



To a 10 cc. sani})le of the citrate solution, 7.80 cc. of the ammonia 

 solution Avere added together with the proper amount of rosolic acid 

 and Ihe color produced comj>ared with that of the Eastman-Hildebrand 

 standard previously used. The latter was found to be more alkaline 

 than the citrate solution though both should have a P h of 7.4. As the 

 only difference between these standard solutions and those previously 

 compared with Clark and Lubs standards by means of phenol red was 

 in their having been diluted with about an equal volume of distilled 

 water, it seemed that this must be the source of trouble. Upon pre- 

 paring two solutions, one according to Eastman and Hildebrand's direc- 

 tions and the other in the same Avay but diluted as above, it was found 

 that the latter was actually the more alkaline of the two. This experience 

 serves to emphasize the admonition of the orginators of this method that 

 the prescribed details must be adhered to. 



The figures quoted speak for themselves, however, as to the possible 

 accuracy of the method. When a comparator is used duplicate titra- 

 tions may be nmde which agree much better than with the litmus 

 method. (See p. 14). Kevertheless there appears to be a small but dis- 

 tinct "salt effect" with rosolic acid which causes a little difference in 

 the character of the colors in the citrate and phosphate solutions making 

 them hard to match. 



While tlie solution proposed by Eastman and Hildebrand answere 

 the purpose, its use is restricted to one indicator, and admits of the 

 preparation of solutions of but one reaction. Furthermore, there is no 

 way of checking up its condition unless a hydrogen electi'ode and gas 

 chain or a series of indicators of known reliability is available. 



It is proposed to substitute for this standard one or more prepared 

 from potassium diliydrogen i^hosphate and sodium hydroxide accord- 

 ing to Clark and Lubs (10). The complete series covers a range 

 from Pho.8 to Ph 8.0 in steps of 0.2. The following table shows the 

 compositions of the various mixtures togetlier with their reactions in 

 terms of Ph ^ 



iJourn. Bact. II. p. 20. 



TABLE I. 



KH2PO4— NaOH Mixtures. 



