The reaction or _titre_ of meat extract, medium, or any solution estimated in the foregoing manner, is most conveniently expressed by indicating the number of cubic centimetres of normal alkali (or normal acid) that would be required to render _one litre_ of the solution exactly neutral to phenolphthalein. [Illustration: FIG. 99.--Stock bottle for dekanormal soda solution.] The sign + (plus) is prefixed to this number if the original solution reacts acid, and the sign - (minus) if it reacts alkaline. For example, "meat extract + 10," indicates a sample of meat extract which reacts acid to phenolphthalein, and would require the addition of 10 c.c. of _normal_ NaOH per litre, to neutralise it. NOTE.--Such a solution would probably react alkaline to litmus. Conversely, if as the result of our titration experiments we find that 25 c.c. of meat extract require the addition of 5 c.c. n/10 NaOH to neutralise, then 1000 c.c. of meat extract will require the addition of 200 c.c. n/10 NaOH = 20 c.c. n/1 NaOH. And this last figure, 20, preceded by the sign + (i. e., +20), to signify that it is acid, indicates the reaction of the meat extract. NOTE.--The standard soda solutions should be prepared by accurate measuring operations, controlled by titrations, from a stock solution of 10N NaOH, which should be very carefully standardised. If a large supply is made or the consumption is small this stock solution must be kept in an aspirator bottle to which air can only gain access after it has been dried and rendered free from CO_{2}. This may be done by first leading it over H_{2}SO_{4} and soda lime, or soda lime alone, by some such arrangement as is shown in figure 99, which also shows a constant burette arrangement for the delivery of small measured quantities of the dekanormal soda solution. STANDARDISATION OF MEDIA. Differences in the reaction of the medium in which it is grown will provoke not only differences in the rate of growth of any given bacterium, but also well-marked differences in its cultural and morphological characters; and nearly every organism will be found to affect a definite "optimum reaction"--a point to be carefully determined for each. For most bacteria, however, the "optimum" usually approximates fairly closely to +10; and as experiment has shown that this reaction is the most generally useful for routine laboratory work, it is the one which may be adopted as the standard for all nutrient media derived from meat extract. Briefly, the method of standardising a litre of media to +10 consists in subtracting 10 from the initial _titre_ of the medium mass; the remainder indicates the number of cubic centimetres of normal soda solution that must be added to the medium, per litre, to render the reaction +10. ~Standardising Nutrient Bouillon.~--For example, 1000 c.c. bouillon are prepared; at the first titration it is found