The quantities of reagents mentioned should be measured accurately, especially in respect of the permanganate solution and the sulphuric acid. Within limits, the former governs the quantity of formaldehyde produced and therefore, primarily, the intensity of colour obtained. Too little total sulphuric 'acid would allow of colour developing from the acetaldehyde, as well as from the formaldehyde; whilst too much lessens the intensity of the colour from the latter. A considerable excess of sulphuric acid at the oxidising stage may result in some formaldehyde beiner produced even from ethyl alcohol itself. Again, if the final volume c' the liquid is increased considerably - as by the addition of 10 c.c. of Schiff's reagent instead of 5 c.c. - the concentration of the sulphuric acid may be so much reduced as to allow of colour developing from the acetaldehyde. But provided the directions given are adhered to with reasonable exactness there is no risk of any such errors arising. A useful criticism of the method has been written by G. Cecil Jones.1

To obtain the greatest degree of accuracy, it is necessary to treat the sample and the standards quite alike as regards quantities of reagents. The second quantity of sulphuric acid added, for example, may not be exactly 10 c.c. - it may be 09 c.c. or 1.1 c.c.; but so long as the same volume is added to sample and to standard this small variation will make no difference to the results. Provided l Analyst, 1915, 40, 221.

all the other conditions are rigidly adhered to, the best way of increasing the sensitiveness of the test is, in the opinion of the writer quoted, to use 5 c.c. of permanganate solution instead of 2 5, and as a consequence, 1 c.c. of oxalic acid solution instead of 05 c.c.

The purification of the distillate by extraction with petroleum ether as indicated should never be omitted, unless it is known that interfering substances are absent. According to E. Salkowski,1 many different alcoholic liquids yield a distillate which gives a reaction indicating the presence of methyl alcohol when tested by Deniges's method. These should therefore be purified as indicated.

Some three years or more after the foregoing process was published T. von Fellenberg described the same method with some slight modifications.2 He adapts it, for use with a distillate which may contain only aqueous methyl alcohol, by adding a mixture of ethyl alcohol and dilute sulphuric acid before oxidation. See his subsequent application of the method to the determination of pectin in spices, given a little later on.

Elias Elvove has applied the process to the detection and estimation of methyl alcohol vapour in air.3 He finds that the sensitiveness is increased for this purpose by reducing the proportion of ethyl alcohol present from 10 per cent. to 0.5 per cent., the quantity of oxalic acid being increased to 0.7 c.c. on account of the decreased amount of ethyl alcohol. In these circumstances, as little .as 0 1 mg. of methyl alcohol showed a coloration on standing for 40 minutes.

(This is, of course, a difference problem from the detection and stimation of methyl alcohol occurring in ethyl alcohol, as the vapour is absorbed in water.) A definite volume of air containing the vapour is suitably passed through water, and to estimate the methyl alcohol in this or any similar aqueous solution proceed as follows:

Ascertain by a preliminary experiment the approximate amount of methyl alcohol in the solution. If this shows that 5 c.c. of it contain more than 1 mg. of methyl alcohol, dilute so as to bring it within this limit. Mix 45 c.c. of this diluted solution with 0 5 c.c. of 4 per cent. ethyl alcohol. Similarly, prepare several 5 c.c. portions of standard methyl alcohol solutions by diluting the proper amounts of a 0.1 per cent. (01 gram per 100 c.c.) aqueous

1 Zeitsch. Nahr. Genussm., 1914, 28, 225.

2 In Mitt. Lebensmittelunters. Hyg., 1915, 6, i; J. Chem. Soc. (Abst.) 1915, 108, ii, 587.

3 J. Ind. Eng. Chem., 1917, 9, 295, solution of methyl alcohol to 4 5 c.c. with water and then adding 0.5 c.c. of 5 per cent. ethyl alcohol to each. These standards are made to vary by 0.1 mg. of methyl alcohol, and the limits are chosen so as to bring the unknown solution for estimation within their range. The liquids are then treated with permanganate, etc., as already described, but using 0.7 c.c. of the oxalic acid instead of 0.5 c.c. Before adding the Schiffs reagent, the solutions are allowed to cool to the ordinary temperature for the sake of uniformity in treatment; and after the addition, they are allowed to stand forty minutes before being compared.

The directions given by Elvove for preparing the Schiff's reagent are substantially similar to those already noted, but a weighed quantity of anhydrous sodium sulphite is used instead of a saturated solution of the bisulphite. 0 2 Gram of finely powdered fuchsine is dissolved in about 120 c.c. of hot water and cooled to the ordinary temperature; 2 0 grams of anhydrous sodium sulphite are dissolved in about 20 c.c. of water and added to the fuchsine solution; then 2.0 c.c. of hydrochloric acid (sp. gr. 1.19) are added, and the whole diluted to 200 c.c. with water. After standing for about an hour the solution is ready for use.

Another interesting application of the method has been made by T. von Fellenberg, who has used it for the determination of pectin in spices by estimating the methyl alcohol derived from the pectin.1 From 1 to 2 grams of the sample are placed on a filter and treated with small quantities of boiling 95 per cent. alcohol until the filtrate measures about 40 c.c. After a similar treatment with light petroleum, the contents of the filter are dried, transferred to a flask, and distilled with 40 c.c. of water, 20 c.c. of distillate being collected. The mixture in the flask is now treated, while hot, with 5 c.c. of 10 per cent. sodium hydroxide solution. After five minutes, 2 5 c.c. of dilute sulphuric acid (1:4) are added, and 16.2 c.c. are distilled over. This distillate is mixed with 5 drops of 10 per cent. silver nitrate solution and again distilled, 10 c.c. of distillate being collected and again distilled. The final distillate, measuring exactly 6 c.c, is weighed. Three c.c. of it are then treated with 1 c.c. of an alcohol-sulphuric acid solution (21 c.c. of 95 per cent. alcohol and 40 c.c. of conc, sulphuric acid diluted with water to 200 c.c), and 1 c.c of 5 per cent. permanganate solution. Comparison solutions containing 5 mg., 1 mg., and 0 3 mg. of methyl alcohol, respectively, are similarly treated at the same time. After two minutes, the mixtures are each treated with 1 c.c. of 8 per cent. oxalic acid solution, 1 c.c. of cone sulphuric acid.