ON THE SENSITIVENESS OF INDICATORS. 271 



It is evident that if we wish to define with greater exactness the ranges 

 of sensitiveness of indicators in this way we must measure as accurately 

 as possible their dissociation constants. This, however, is by no means 

 an easy task ; an indicator is generally both too weak and too insoluble 

 an electrolyte for us to be able to determine its dissociation constant by the 

 ordinary conductivity method. The only accurate method appears to be 

 a quantitative measurement of the depth of colour of the indicator in 

 solutions of different, accurately known, concentrations of hydrions. 

 This is only convenient when the indicator is only coloured in one form 

 or has two coloured forms which practically do not differ in tint, but only 

 in depth of colour. 4 Approximate determinations made by means of the 

 solutions already referred to are sufficient for most ordinary work, and it 

 would be of the greatest assistance if in future every discoverer of a new 

 indicator would test his product in this manner. 



Since the concentration of hydrogen ions in pure water is 10 -7 at 25°, it 

 follows that the exact neutral point is only indicated by an indicator .with 

 a dissociation constant of about 10" 7 (litmus). It does not follow that 

 this is the most useful indicator, in fact the contrary is true. Speaking 

 generally, however, we may say that the most sensitive indicators are 

 those which have dissociation constants not widely different from 10~ 7 ; 

 for evidently the change from 10~ 3 to 10 -5 (methyl orange), that is from 

 one-thousandth to one-hundred-thousandth normal, is more considerable 

 than the change from one-millionth to one-hundred-millionth normal. 

 An indicator must therefore be a weak acid or base, as Ostwald said. 

 But it must not be too weak ; an indicator with a dissociation constant of 

 10" n , for example, changes over between concentrations of hydrogen ions 

 of 10~ 10 and 10 -12 , that is to say between concentrations of hydroxylions 

 of 10 4 and lO" 2 (since C H x C 0H = 10~ 14 at 25°). Such a change is only 

 brought about by addition of a considerable amount of alkali. 5 



It is now possible to apply these conclusions to the actual process of 

 titration. If an alkaline solution be gradually added to an acid solution, 

 the concentration of hydrogen ions in the latter becomes smaller and 

 smaller until a point is reached when the indicator present begins to 

 dissociate appreciably. This point may or may not be the point when 

 exactly equivalent quantities of base and acid are present together ; that 

 depends obviously upon the indicator used. Now we can either stop the 

 titration directly we observe a distinct change in colour, or when further 

 slight addition of alkali has no more appreciable effect. In practice it 

 has been found most convenient to take as our end-point the last part of 

 the colour change when we titrate from a dark to a light colour, and the 

 first part of the colour change when we titrate in the reverse direction ; and 



4 See Journ. Chem. Soc, 1910, p. 2477. 



5 The range of sensitiveness of an indicator probably alters considerably with the 

 temperature. The ' apparent ' dissociation constants of all pseudo-acids and bases 

 have been found to have high temperature coefficients, and indicators should form no 

 exception to this rule. If a 200th normal solution of acetic acid (C H = 3x 10 ') con- 

 taining methyl orange be warmed from 0° to 25°, the colour of the solution becomes 

 distinctly paler, although the dissociation' constant of acetic acid alters very little 

 with the temperature, and in any case C H does not become smaller. The phenomenon 

 is best explained by an increase in the dissociation "constant of methyl orange. It 

 follows that when indicators are used for the colorimetric determination of hydrion 

 concentrations, care should be taken to keep the temperature constant — a precaution 

 which has not been thought necessary up to the present. 



