50 CULTURE MEDIA. 
of alcohol, should be added to one cubic centimetre of bouillon. The 
beginning of an alkaline reaction is indicated by the appearance of a 
faint rose color. Fuller (1895), who has made a careful investiga- 
tion of this subject, recommends a modification of the method of 
Schultz. He gives the following directions in his paper published 
in the Journal of the American Public Health Association (Vol. 
XX., p. 386): 
This indicator is prepared by dissolving five grammes of commercial 
phenolphthaleinin one litre of fifty-per-cent alcohol. It is not feasible to use 
this indicator on strips of paper as the alcohol quickly evaporates, leaving the 
powder. The colorless liquid, however, may be added in small quantities to 
solutions of an acid or neutral nature without any change of color, but alka- 
lies quickly change it toa purple red. This change from no color to one of 
purple red makes the indicator a very satisfactory one, owing to the ease 
with which the eye detects the so-called end-point. ale 
For the determination of the degree of reaction of nutrient media it is 
the custom to put five cubic centimetres (practically five drachms) of the solu- 
tion into a six-inch porcelain evaporating dish, together with forty-five cubic 
centimetres of distilled water. This liquid is boiled for three minutes, after 
which is added one cubic centimetre of the phenolphthalein solution. While 
the solution is still hot it is quickly Bteatet, against a twentieth normal solu- 
tion of caustic alkali. 
As already mentioned, Dahmen stated that this indicator is useless, ow- 
ing to the presence of carbonates, and of ammonia and its salts. On gen- 
eral grounds the point that this indicator is inaccurate under such conditions 
is well taken, but so far as its application to nutrient mediais concerned it 
has been found that by proper precautions these objections may be over- 
come. 
With regard to the amount of free and combined ammonia present in 
ordinary nutrient media at the times when their reaction is determined, it 
has been found that it does not exceed 0.003 per cent. Experiments show 
that this quantity is less than one-tenth of that necessary to interfere with 
the accuracy of the method. It may be added that the reason why no am- 
monia is produced by the addition of alkali to the nitrogenous bodies is that 
at no rae during the preparation is there an appreciable excess of free alkali 
present. s 
The chief point by which the presence of carbon dioxide is obviated is by 
the use of caustic soda instead of sodium carbonate for neutralization, as re- 
ferred to beyond. It has been learned by actual experiment that the carbon 
dioxide is practically all removed by heat during the preparation of the 
media and at the time of boiling just prior to the titration. In order to pre- 
vent atmospheric absorption of this gas the titration should be made quickly 
and in a hot solution. 
The remaining precaution concerns the solution against which the nu- 
trient media are titrated. All of the usual media react acid to phenolphtha- 
lein; hence the solution in question must be alkaline. Caustic soda serves 
the purpose well, and the strength may be conveniently one-twentieth normal, 
equal to two grammes per litre. For the sake of prevention of interference 
from carbon dioxide in the standard solution as it meets the indicator, it is 
well to adda small quantity of calcium hydrate in order to precipitate this 
gas as calcium carbonate and allow it to settle out in the reagent bottle. 
After this solution has been accurately prepared, care is necessary in order 
to keep it of uniform strength and free from carbon dioxide. This is best 
done by placing the stock solution bottle on a shelf from which the liquid 
may be delivered into the burette by means of a siphon that is connected 
tightly with the top of the burette. In the tightly fitting stopper of the bottle 
