266 Development of a Standardised Agar Medium 



developed, and secondly, to develop a medium whose change of reaction 

 during sterilisation would not reach this limit. 



In the measurements of H-ion concentration involved in this work, 

 the indicator method developed by Gillespie (12) was used. This method 

 depends on the assumption that, at any given H-ion concentration, a 

 definite percentage of the indicator is in the acid form and the remainder 

 in the alkaline form. If it is known what these proportions are for a 

 given indicator at a given reaction, the colour shown by the indicator 

 in a solution of this reaction, can be imitated by dividing the indicator 

 in the correct proportions between two solutions, one of which contains 

 excess of acid and the other excess of alkali. These ratios have been 

 ascertained by Gillespie for a number of indicators over a range of 79H 

 values. Colour standards prepared by this method consist of pairs of 

 tubes, one containing dilute acid and the other dilute alkali. Each pair 

 together contains ten drops of the indicator, these drops being divided 

 between the two tubes according to the ratio ascertained for the ^H 

 value required. 



Before employing this technique, it was thought advisable to test 

 the accuracy of readings obtained with it, as compared with the method 

 of Clark and Lubs(i3) in which the indicator colour standards are made up 

 in standard bufEer solutions of definite pK value. These tests of Gillespie's 

 method were carried out at Rothamsted by Mr E. A. Fisher, to whom, 

 also, I am indebted for much help and advice throughout the work con- 

 nected with the reaction of the medium. The following indicators were 



tested : 



Range 



Brom Cresol Purple (Dibromo-o-cresolsulphonphthalein) joH 5-6-^jH 6-8 

 Brom Thymol Blue (Dibromo-thymolsulphonphthalein) fH 6-5-_pH 7-7 



Phenol Red (Phenolsulphonphthalein) joH 7 -2-^011 8-3 



With each indicator, seven pairs of tubes were made up as colour stan- 

 dards, as described by Gillespie. The indicators were added in drop ratios 

 ranging from 8 acid: 2 alkaline, to 2 acid: 8 alkahne. The colour of each 

 pair of tubes was compared with a series of colour standard tubes pre- 

 pared by Clark and Lubs' method in each of which ten drops of indicator 

 were added to 10 c.c. of a buffer solution of known H-ion concentration. 

 The readings thus obtained are shown in Fig. 12, in which the actual 

 readings are plotted on "smoothed" curves. In those cases where the 

 actual readings lay off the smoothed curves, the pairs of drop ratio tubes 

 were made up several times and invariably gave similar readings. It is 

 therefore believed that these irregularities were due to sHght errors in 



