The estimation of the amount of sizing material in a paper is a very complicated process and one which demands considerable chemical experience for its proper conduct. The amount of gelatine present is best ascertained by determining the amount of nitrogen present by combustion with soda-lime, and from this, calculating the amount of gelatine. Pure gelatine contains 18·16 per cent. of nitrogen (Muntz). The comparison of one paper with another with a view to ascertain the relative degree of sizing, is usually performed in a more or less rough and ready way by moistening the samples with the tongue for a certain time, and noticing the degree of transparency produced, which is of course inversely to the degree of hardness. A more accurate method consists in placing a drop of a mixture of alcohol and water containing some colouring matter in solution and determining the time necessary for the colour to make its appearance on the other side. In this way more trustworthy comparisons can be made. _Colouring Matters._—The chemical reactions of the chief colouring matters have been already described (p. 141). {205} CHAPTER XIV. GENERAL CHEMICAL ANALYSIS FOR PAPER-MAKERS. A knowledge of the methods usually employed for the qualitative and quantitative analysis of the various chemical substances met with in paper manufacture is of considerable importance. The scope of this work will not allow us to enter fully into the necessary details of manipulation; for these the reader is referred to standard works on analysis; we will merely indicate the methods. _Caustic Soda, Soda Ash, Recovered Soda, &c._—These substances are always valued by the amount of real alkali (Na_{2}O) that they contain, which is determined by titration with a standard acid. In testing recovered soda, it is necessary to _boil_ the finely powdered ash for some time with water; unless this be done, the amount of alkali will probably be under-estimated, as part of the soda is present as a difficultly soluble silicate. It is sometimes desirable to determine the amount both of sodium hydrate and carbonate in a sample of alkali. For this purpose an excess of a solution of barium chloride is added to a solution containing a known quantity of the substance under examination, and the mixture made up to a definite volume. It is then filtered, and an aliquot portion of the filtrate titrated with standard alkali. The filtrate contains only the alkali that was originally present as hydrate, the carbonate of soda having formed with the barium chloride, insoluble barium carbonate, and sodium chloride.