In the following pages are described methods which can be used for the estimation of methyl alcohol when present in mixture with ethyl alcohol. Certain processes are included here which are essentially qualitative, but which can nevertheless be made to yield fairly good approximate results by comparison with "control" experiments made on known mixtures. For example, the time-honoured method of Riche and Bardy, tedious but trustworthy, can be employed in this way.

1 Zeitsch. Nahr, Genussm., 1914, 27, 469, the methyl alcohol to formaldehyde, and the detection of the latter by means of Schiff's reagent in the presence of sufficient sulphuric acid to prevent the development of colour with any acetaidehyde formed from the ethyl alcohol during the oxidation.

As regards the best method to use in any particular case, much will depend upon the probable proportion of methyl alcohol in the mixture. No one method will suit every case. The refractometric process, for instance, would be useless in examining wood naphtha, on account of the acetone present; the iodide method used for wood naphtha would be inapplicable (without modification) when ethyl alcohol is also a constituent of the mixture examined; the oxidation process of Thorpe and Holmes is not suitable for very small quantities; and so on. Moreover-, apart from the question of proportion, it may happen that in particular circumstances one process will be advantageous as compared with another, though the latter may be usually the better. But so far as a general indication can be given, the following comments will serve as a useful guide.

For very small quantities, ranging up to 3 or 4 per cent., Deniges's process as modified by the writer (p. 184) will be found, in general, by far the best to use. Even when the proportion is as much as 5 to 10 per cent. of the total alcohols, the method can often be used to advantage, though of course any error in matching the colours is then multiplied rather largely, since only about 0 002 to 0 004 gram of methyl alcohol is employed in the actual comparison. With moderately large proportions of methyl alcohol - say from 5 to 15 per cent. - the process proposed by Thorpe and Holmes (p. 188) gives quite satisfactory results. With still larger proportions fair estimations may be made by means of the immersion refractometer (see Chap. VII). But in ordinary practice the cases are few which cannot be treated by one or other of the two methods first mentioned, confirmed if necessary by the shortened form of the Riche and Bardy process, or by such of the remaining methods as may best suit the circumstances. (For the determination of methyl alcohol when no ethyl is present, see under " Wood naphtha" and "Alcoholometry").

Deniges's method,1 modified by the author.2 and 5 c.c. of magenta-sulphurous acid solution (Schiff's or Gayon's reagent), and the colorations obtained are compared after the lapse of one hour. The pectin content of the sample is ten times the content of methyl alcohol.

The principle of this very convenient method is the oxidation of

1 Compt. rend., 1910, 150, 832,

2 The Analyst, 1912, 37, 16.

The following solutions are required: -

Potassium Permanganate

2.0 Grams of KMnO4 per 100 c.c.

Oxalic Acid

9.5 Grams of the crystallised acid per 100 c.c.