rinsings being added to the contents of the large flask. Some starch paste is now added, and then iodine solution until a permanent blue coloration is produced. The final reaction can be looked upon as conclusive only if the coloration does not disappear again upon addition of solution of carbonate, or, better still, bicarbonate of soda. On disconnecting the apparatus, neither the large nor the small flask must emit the odor of chlorine; otherwise the experiment is a failure. To afford

some indication of the quantities to be used, I may remark that, in the analysis of bichromate of potassa, for instance, 0.2-0.5 grm. of that salt may be taken, to 40-100 c.c. of arsenite of soda solution in the receiver.

Upon an attentive consideration and comparison of the two preceding methods, a and b, it will be seen that in the former the quantity of iodine is to be determined which has been liberated by chlorine; whilst in the latter, on the contrary, we have to deal, not with the arsenious acid which has been oxidized, but with the excess of the acid which has not been oxidized. Now, although this is of no great consequence in cases where the amount of chlorine is comparatively large, it is of very considerable importance in cases where the amount of chlorine is only small, more particularly where it unexpectedly turns out to be so, and where, accordingly, a pretty large quantity of arsenate of soda solution has been employed in the analytical process. For, in such cases, it may happen that the small difference which lies within the limits of the

errors of observation, and which, under the circumstances, on account of the indefinite and uncertain quantity of carbonate of soda added, is likely to rise to as high as 0.5 c.c., amounting to,, and even more of the difference between the quantities of the iodine solution respectively consumed in the experiments, before and after addition of the chlorine water; which may lead to very serious errors in the results, to the extent, indeed, of 10 or 20 per cent., and even more.

For this reason Bunsen's method (a) alone deserves to be recommended for the estimation of smaller quantities of chlorine.

c. With Solution of Protoxide of Iron and Permanganate of Potassa. The chlorine solution is mixed with an excess of solution of protochloride of iron, or sulphate of protoxide of iron and ammonia of known strength, in a stoppered flask; the mixture is allowed to stand for some time, and the iron, still left in the state of protochloride or protoxide, determined by solution of permanganate of potassa (§ 112). It must be borne in mind, in the calculation, that 2 eq. of protochloride of iron are converted into sesquichloride by 1 eq. of chlorine. This method is suited, indeed, for the estimation of chlorine in aqueous solution; but it is much less adapted to effect the determination of chlorine when evolved in gaseous form, as the gas is absorbed with comparative slowness by the protochloride of iron solution.

2. Gravimetrical Method.

The fluid under examination, which must be free from sulphuric acid, say, for instance, 30 grammes of chlorine water is mixed in a stoppered bottle, with an excess, say 0.5 grm., of hyposulphite of soda, the stopper inserted, and the bottle kept for a short time in a warm place; after which the odor of chlorine is found to have gone off. The mixture is then heated to boiling with some hydrochloric acid in excess, to destroy the excess of hyposulphite of soda, filtered, and the sulphuric acid in the filtrate determined by baryta (§ 132). 1 equivalent of sulphuric acid corresponds to 2 equivalents of chlorine (Wicke, "Annal. d. Chem. u. Pharm.," 99, 99).

In fluids containing, besides free chlorine, also hydrochloric acid, or a metallic chloride, the chlorine existing in a state of combination may be determined, in presence of the free chlorine, in the following way :—

=

a. A weighed portion of the fluid is mixed with ammonia in excess nitrogen escapes, and the solution contains the whole of the free chlorine as chloride of ammonium (3 Cl + 4 N H, N+3 [N H, Cl]). By precipitating now with solution of silver, we learn the total amount of the chlorine. The quantity of the free chlorine is then determined in another weighed portion, by means of iodide of potassium, or by some other method; the difference gives the amount of chlorine which the analysed fluid contained in a state of combination.

b. A weighed portion of the fluid is mixed with solution of sulphurous acid in excess, the mixture acidified with nitric acid, and the whole of the chlorine precipitated as chloride of silver. The further operation is conducted as in a.

Having thus seen in how simple a manner the quantity of free chlorine may be determined by Bunsen's method, and for larger amounts

*If chlorine water is mixed at once with solution of nitrate of silver, ths only of the chlorine are obtained as chloride of silver : 6 Cl + 6 Ag O 5 Ag Cl + Ag 0, C10 ̧ (H. Rose; Weltzien, "Annal. d. Chem. u. Pharm.," 91, 45).

=

also by Mohr's method, it will be readily understood that all oxides and peroxides which evolve chlorine when heated with hydrochloric acid, may be analysed by heating them with concentrated hydrochloric acid, and determining the amount of chlorine evolved. For the modus operandi compare § 130, d, B, and § 142, b.

a. Free hydrobromic acid is precipitated from its solution with nitrate of silver, and the further process conducted as in the case of chlorine (§ 141). For the properties of bromide of 'silver, see § 94, 2. The results are perfectly accurate.

b. Heine's colorimetrical method.* The bromine is liberated by means of chlorine, and received in ether; the solution is compared, with respect to color, with an ethereal solution of bromine of known strength, and the quantity of bromine contained in it thus ascertained. Fehling ("Journ. f. prakt. Chem.," 45, 269) obtained satisfactory results by this method. It will at once be seen that the amount of bromine contained in the fluid to be analysed must be known in some measure, before this method can be resorted to. As the mother liquor examined by Fehling could contain at the most 0.02 grm. of bromine, he prepared ten different test fluids, by adding to ten several portions of 60 grammes each of a saturated solution of common salt increasing quantities of bromide of potassium (containing respectively from 0.002 grm. to 0.020 grm. of bromine). He added an equal volume of ether to the test fluids, and then chlorine water, until there was no further change observed in the color of the ether. It being of the highest importance to hit this exact point, Fehling prepared three samples of each test fluid, and then chose the darkest of them for the comparison. 60 grammes are now takent of the mother liquor to be examined, the same volume of ether added as was added to the test fluids, and then chlorine water. Every experiment is repeated several times. Direct solar light must be avoided, and the operation conducted with proper expedition.

c. Figuier's colorimetrical method ("Annal. de Chim. et de Phys.," 33, 303, and "Journ. f. prakt. Chem.," 54, 293), proposed as a useful method to effect the determination of bromine in mother liquors, &c.

This method is based upon the circumstance that 1 equivalent of chlorine (added in the form of chlorine water), liberates from a solution of a metallic bromide 1 equivalent of bromine, and that bromine imparts a yellow color to an aqueous solution, and escapes readily upon boiling, the yellow tint of the solution disappearing again with the escape of the bromine.

To carry this method into effect, the strength of the chlorine water is determined at the moment of its application, by making it act upon a solution of bromide of sodium of known strength, acidified with a few drops of hydrochloric acid (or by one of the methods given in § 142), and then applying it to the mother liquor. The latter is heated in a flask

"Journ. f. prakt. Chem.," 36, 184, proposed as a useful method to effect the determination of bromine in mother liquors.

The best way is to take them by measure.

nearly to ebullition; chlorine water is then added from a burette covered with black paper, and the mixture heated for about 3 minutes, whereupon the yellow tint imparted to the fluid by the addition of the chlorine water will disappear again; the mixture is now allowed to cool for 2 minutes, after which some more chlorine water is dropped into it, heat again applied, and the same process repeated until further addition of chlorine water fails to impart a yellow color to the fluid. Should the experiment last several hours, the strength of the chlorine water must be determined once more at the end of the process, and the calculation of the results based upon the mean of the two experiments. Alkaline fluids must be slightly acidified with hydrochloric acid. Protoxide of iron, protoxide of manganese, iodine, and organic matters must not be present. Mother liquors colored yellow by organic matter should be evaporated to dryness, the residue gently ignited, then treated with water, and the fluid filtered. In evaporating the solutions to dryness, carbonate of soda must be added, since chloride and bromide of magnesium evolve hydrochloric and hydrobromic acids in the process.

II. Separation of Bromine from the Metals.

The metallic bromides are analysed exactly like the corresponding chlorides (§ 141, II., a to d), the whole of these methods being applicable to bromides as well as chlorides. In the decomposition of bromides by sulphuric acid (§ 141, II., d), porcelain crucibles must be used instead of platinum ones, as the liberated bromine would injuriously affect the latter.

Supplement.

Determination of Free Bromine.

§ 144.

Free bromine in aqueous solution, or evolved in the gaseous form, is determined in the same way as free chlorine (see § 142).

A

Another method has been proposed by Williams ("Chem. Gaz.,” 1854, 432). It is based upon the fact that free bromine suffers decolorization by the action of oil of turpentine, the bromine replacing the hydrogen in the latter-34 parts of oil of turpentine (1 equivalent) decolorize 79.97 parts (1 equivalent) of bromine. solution of perfectly pure oil of turpentine in absolute alcohol is used as test fluid, 20 grammes of the oil being dissolved to 200 c.c. of fluid. The fluid containing the free bromine should be in a stoppered bottle. The test fluid is added drop by drop, the bottle being shaken after every addition, and the operation continued until the mixture is quite colorless. Every 34 c.c. correspond to 8 grms. of bromine. For small quantities of bromine a more dilute test fluid must be used. Results satisfactory.

The determination of free bromine in presence of hydrobromic acid or metallic bromides is also effected in the same manner as that of free chlorine in presence of hydrochloric acid or metallic chlorides (see § 142, 2). The addition of ammonia to the bromine requires caution. If bromine in solution is to be converted by ammonia into bromide of ammonium, the solution is poured into a capacious flask, a tolerably large quantity of water added, and then the ammonia through a funnel tube. The nitrogen gas escaping is transmitted, by means of a bent tube, through dilute ammonia; the two fluids are then mixed

together, and the remaining part of the process is conducted as directed in § 142. By this means all loss of substance is effectively guarded against.

a. If you have hydriodic acid in solution, precipitate with nitrate of silver, and proceed exactly as with hydrochloric acid (§ 141). For the properties of iodide of silver, see § 94, 3. The results are perfectly

accurate.

b. The following method, recommended first by Lassaigne, is resorted to almost exclusively to effect the separation of hydriodic acid from hydrochloric and hydrobromic acids, for which purpose it is extremely well adapted. Acidify the solution slightly with hydrochloric acid, and add a solution of protochloride of palladium, as long as a precipitate forms; let the mixture stand from 24 to 48 hours in a warm place, filter the russet-black precipitate off on a weighed filter, wash with warm water, and dry at a temperature from about 158° to 176° F., until the weight remains constant. The drying may be greatly facilitated by replacing the water (after the operation of washing) by some alcohol, and the latter fluid again by a little ether. For the properties of the precipitate, see § 94, 3. This method gives very accurate results, provided the drying be managed with proper care; but if the temperature is raised to near 212° F., the precipitate smells of iodine, and a trifling loss is incurred.

Instead of simply drying the protiodide of palladium, and weighing, it in that form, you may ignite it in a crucible of porcelain or platinum,* and calculate the iodine from the residuary metallic palladium (H. Rose).

c. Kersting's volumetrical method ("Annal. der Chem. und Pharm.," 87, 25). This method is based upon the precipitation of iodine from iodide solution by protochloride of palladium. The process requires,—

a. A solution of pure iodide of potassium containing exactly 1 part of iodine in 1000 parts of fluid. This is prepared by dissolving 1.308 grm. of ignited iodide of potassium in water, and diluting the solution to 1 litre of fluid.

B. An acid solution of protochloride of palladium containing exactly 1 part of palladium in 2370 parts of fluid. This is prepared by dissolving 1 part of palladium in nitrohydrochloric acid, with application of heat, evaporating the solution to dryness at 212° F., adding 50 parts of concentrated hydrochloric acid and 2000 parts of water, and allowing to deposit. The exact strength of the clear solution is then ascertained by means of the solution of iodide of potassium, in the manner described below (Analytical Process).

y. The solution of the iodide to be analysed. Dissolve the iodide in water, if possible, and determine the amount of iodine in it approximately, in the manner described below (Analytical Process); dilute the rest of the solution until it contains 1 part of iodine in about 1000 parts, and then determine the exact amount of iodine in it by the same method.

Should the iodide be insoluble in water, or not well adapted for direct

This substance is not injured by the operation.