operation; six or seven washings usually suffice. Wash thoroughly with 95% alcohol to remove the ammonium chlorid; then dry and weigh the precipitate and calculate the results as in the older method.

PREPARATION OF ACIDULATED ALCOHOL.

To each 1000 cc. of 95% alcohol, add 75 cc. of concentrated hydrochloric acid, then pass dry hydrochloric acid gas into the mixture until 1 cc. of the alcohol neutralizes 2.25 cc. of normal potassium hydroxid, using phenolphthalein as an indicator. The hydrochloric acid gas may be prepared by using C. P. sodium chlorid and concentrated sulphuric acid, or by heating concentrated hydrochloric acid and first passing the gas through sulphuric acid and then into the alcohol.

The writer made several analyses of mixed fertilizers using the de Roode, official Lindo-Gladding, and former official Lindo-Gladding methods. The former official Lindo-Gladding method differs from the official method only in the method of obtaining the solution of the potash. The results obtained are far from satisfactory, as may be seen by consulting the following table. It will be noted that the majority of results obtained by the official Lindo-Gladding method are lower than the corresponding results by the former official method. Those obtained by the de Roode method are so unsatisfactory that they are presented for no other reason than to invite suggestions as to the cause of the writer's error or errors.

In the following table it will be noted that 5 cc., 10 cc. and 15 cc. of hydrochloric acid, together with hot water, were used, with varying results, in taking up the residue from the nitric acid evaporation. The method does not advise one in regard to evaporation to a "sirupy consistency" after the addition of the platinic chlorid but the writer in all cases, except the four given in the column labelled "evaporated to a thin sirupy consistency", made this evaporation. In the case of the four exceptions, only a partial evaporation was made, that is, to a very thin sirupy consistency. Although a thorough washing was made with the ammonium chlorid solution, the results obtained are so much higher than those obtained by the regular official method that it is evident the final precipitate contained some material in addition to the potash salt.

After the work was done, the results of which are reported in the table, it was thought that possibly the very high results obtained in some samples were due to ammonia salts not being completely removed before the platinum chlorid was added. Consequently several of the samples which contained ammonia were analyzed by the de Roode method, except that after evaporating with hydrochloric acid they were heated sufficiently to drive off any ammonia that might be present. The averages of the results obtained on eleven samples were: not ignited, 4.29 per cent of potassium oxid; ignited, 4.17 per cent. This would seem to show that the high results were not due to ammonia.

This method was successfully used on mixed fertilizers and potash salts by Keitt, Shiver and Padgett' with very satisfactory results. The chief advantages of the method are: Ease of manipulation; less time consumed; no loss by sputtering; and no chance for occlusion.

Potash, expressed as K2O, obtained by different methods.

REPORT ON WATER'.

By J. W. SALE (Bureau of Chemistry, Washington, D. C.), Referee.

A method for the determination of iodin2 and a method for the determination of bromin3 were selected for cooperative work. These methods gave promise of greater accuracy than is obtainable by the colorimetric methods of this association. The need for better methods for iodin and bromin has been felt for some time by analysts, especially those concerned with the examination of mineral waters and brines.

The method for iodin in the presence of chlorin and bromin follows:

REAGENTS.

(a) 5% solution of equal weights of sodium hydroxid and sodium carbonate.

Take such a quantity of the brine or water as will contain not more than 0.1 gram of iodin or more than 10.0 grams of total salts. Adjust the volume to 100–150 cc., add a sufficient quantity of the solution of sodium hydroxid and sodium carbonate, reagent (a), to precipitate the calcium and magnesium. Boil, filter off the precipitate of calcium and magnesium, and wash with hot water; introduce the filtrate into an Erlenmeyer flask, adjust the volume to about 100 cc., neutralize with dilute sulphuric acid, reagent (b), and add 1 cc. of the solution of sodium hydroxid, reagent (C). Heat to boiling, add an excess of potassium permanganate, reagent (d), about 0.5 cc. excess, continue heating until the precipitate begins to coagulate and then allow to cool. Add sufficient alcohol, reagent (e), or hydrogen peroxid to bleach the permanganate color and set the beaker on a steam bath. When the precipitate has settled, filter and wash the precipitate with hot water. After cooling, add 1-2 grams of potassium iodid, acidify with hydrochloric acid and titrate with N/20 thiosulphate. One-sixth of the iodin titrated represents the amount originally present (1 cc. of N/20 thiosulphate solution equals 1.058 mg. of iodin).

1 Presented by W. W. Skinner.

J. Ind. Eng. Chem., 1919, 11: 563.

3 Ibid., 954.

Assoc. Official Agr. Chemists, Methods, 1916, 47.

The method for bromin in the presence of chlorin, but not iodin follows:

Evaporate the sample of water or brine, which should not be too acid, to dryness or nearly so. Charge the reaction cylinder A, Fig. 1, by introducing glass beads to a depth of about 1 inch, followed by 15 grams of solid chromium trioxid, and finally enough glass beads to fill the cylinder half full. Add 20 cc. of the solution of sodium sulphite and sodium carbonate, reagent (a), to the first absorption cylinder and 5 cc. to the second. Dilute each to about 200 cc. Connect the three cylinders and draw a current of air through slowly. Wash the sample into the reaction cylinder with water sufficient to make about 25 cc. of solution. Aspirate until the contents of the reaction cylinder are in solution and thoroughly mixed, then discontinue, close the inlet tube with a small piece of rubber tubing and clamp, and reduce the pressure in the apparatus slightly by sucking out some air in order to guard against any possible escape of bromin at the ground glass stopper. Allow to stand overnight, then aspirate with a rather strong current of air (about to liter per minute) for 3 hours, adding four 2 cc. portions of 3% hydrogen peroxid at 30-minute intervals. Stop the aspiration and evaporate the contents of the two absorption cylinders nearly to dryness. Clean out the reaction cylinder and freshly charge with glass beads and 15 grams of chromium trioxid. Into the first absorption cylinder put 10 grams of potassium iodid dissolved in 200 cc. of water and into the second 3 or 4 grams in a like amount of water. Connect the apparatus, draw through a slow current of air and transfer the contents of the evaporating dish to the reaction cylinder by means of the small funnel, using 25 cc. of water. Aspirate until all of the bromin is evolved (about 1 hour) and titrate the potassium iodid solution with thiosulphate (1 cc. of N/20 thiosulphate 3.996 mg. of bromin).

=

It will be noted that the method for bromin is to be used only in the absence of iodin. The authors state that iodin may be removed from the sample in which bromin is to be determined subsequently, as follows:

Introduce the neutral or slightly acid sample, which should contain not more than 0.1 gram of bromin or 10 grams of total salts into a distillation flask and adjust to a volume of approximately 75 cc., add 1.5-2.0 grams of ferric sulphate and distil the liberated iodin with steam into 100 cc. of a potassium iodid solution (10 grams of potassium iodid per 100 cc.). The potassium iodid solution may be titrated with sodium thiosulphate solution and the result used to check the figure obtained by the permanganate method.