(115) Solid Residue.

100 c.c. of the water are placed in a platinum dish on the waterbath, evaporated to dryness, cooled and weighed; the weight of the empty dish being subtracted from the combined dish and residue of course gives the weight of the residue. After weighing, the latter may be ignited over a Bunsen burner, and it should be noticed whether there is blackening or scintillation.

(116) Ammonia.

Nearly all waters are alkaline and therefore the ammonia present may be simply expelled by boiling; the reaction of water is best ascertained by a solution of methyl orange; if the water is not alkaline it will be necessary to add a very small quantity of recently ignited sodic carbonate, and some chemists always add this and return the ammonia as "expelled by sodic carbonate." Half a litre or a litre of the water is placed in a tubulated retort adjusted to a Liebig's condenser, and 100 c.c. distilled over, a cubic centimetre of the Nessler solution is added to the distillate; if no ammonia at all is present, no perceptible yellow or amber tint will be noticed, but in any other case a pale yellow, amber, or dark brown colour or even a precipitate, according to the amount of ammonia present, will be produced. In these cases the maximum of colour requires a certain time, from ten minutes to a quarter of an hour, for its development. The exact tint of colour produced is to be imitated by running into 80 or 90 c.c. of pure distilled water, a carefully measured volume from a burette of the standard ammonic chloride solution, adding 1 c.c. of Nessler, and making up with pure distilled water, to the same volume of the distillate : this tint is not likely to be hit off at the first trial, either the comparison liquid is too dark or too light. It is of course open to the analyst to repeat the process by making other comparison fluids with varying quantities of the ammonic chloride solution until the right tint is obtained, but this is tedious and unnecessary. It is easier to reduce the length of the darker liquid, until it balances the lighter, and then to calculate the amount. For instance, supposing the Nesslerised distillate measuring exactly 100 c.c. is darker than the comparison liquid, which latter contains 2 c.c. of the ammonic chloride solution, but it is found that by pouring away

30 c.c. of the darker liquid, the two discs seen by looking down through the two cylinders placed side by side on a white surface such as a porcelain plaque, are exactly matched, then the following proportion sum will give the number of c.c. of ammonic chloride which the distillate is equal to—as 70:2::100 = 2·85 c.c.

Various forms of tubes are to be bought facilitating this method, such as Nessler cylinders with taps half-way down and the tubes graduated into divisions, others have mirrors attached below and are fixed in a stand with proper arrangements for shutting off extraneous light. But for practical purposes glass cylinders of pure white glass, a white porcelain slab, and an upright narrow glass measure graduated into c.c. are quite sufficient for the purpose. Of course the number of c.c. of ammonic chloride to which the distillate is equal to gives at once the value for ammonia; for example, in the case quoted the ammonia would be equal to 2:85 c.c. x .0001 that is 000285 grm. of ammonia. With water containing so little ammonia as the above, the whole will practically come over in the first 100 c.c. but it is safe to distil a second quantity of 100 c.c., and to test by adding again the Nessler solution; waters which contain so large a quantity of ammonia as to give a distillate of a deep amber colour, must be distilled in less quantities than the above, for as Nesslerising is only accurate for the estimation of small quantities of ammonia, a quarter of a litre or less should be taken of waters known to be impure; or any water which by a preliminary testing with Nessler before distillation is shown to be very impure, is better dealt with by distilling a quarter of a litre over, colouring the distillate with a c.c. of methyl orange solution, and dropping into it from a burette deci-normal hydrochloric acid until it just changes colour, every c.c. of d.n. hydrochloric 0017 NH.

=

The free ammonia thus estimated, the next process is to add 50 c.c. of alkaline permanganate to the liquid in the retort and to again distil. It is well to boil up the 50 c.c. for some minutes in an open dish before adding it, and the addition of the hot alkaline solution is best poured through a funnel passed through the tubulure of the retort. The distillate is collected, about 100 c.c. at a time, and the various fractions Nesslerised until all ammonia is proved to have come over. The result is returned as albuminoid ammonia."

(117) Chlorine.

Chlorine is estimated by the volumetric solution of silver nitrate. 100 c.c. of the water are either placed in a beaker resting on a white slab, or in a porcelain dish, just coloured with a solution of potassic chromate, and then the silver nitrate solution dropped in from a burette, until a reddish tinge is produced. The explanation of the test is as follows: silver chromate is a red salt, it is not formed until all the chlorides are first used up; upon its appearance, therefore, it is certain that all the chlorine is combined with silver. Each c.c. of the silver solution equals a milligrm. (001 grm.) of chlorine, so that the number of c.c. used multiplied by 001 equals the chlorine in 100 c.c. of water. (The reaction is made more delicate by observing the tint through a solution of chromate.)

(118) Nitrates.

The quickest method of estimating nitrates is by indigo-carmine. It is fairly accurate when the nitrates are in quantities—such for example as over a grain per gallon, but save with special precautions it is not a good method for the estimation of small quantities.

4 grms. of sublimed indigotin are digested for some hours with five times their weight of Nordhausen sulphuric acid, the liquid is diluted, and made up to two litres. A normal nitre solution is made by dissolving 1'011 grm. of pure potassic nitrate in one litre of water. From this solution, solutions of,,,, normal are prepared. An assay is now made by mixing, say 20 c.c. of the nitre solution with any amount of the indigo solution run from a burette deemed sufficient, in a wide-mouthed flask of 150 c.c. capacity. Oil of vitriol is run into a test-tube, the volume being equal to the united volumes of the indigo solution. The contents of the test-tube are now tipped into the mixture and the flask heated (best by a chloride of calcium bath) to 140°. If the indigo solution is insufficient, the liquid is suddenly decolourized, but in other case the liquid will continue blue. Less or more is therefore taken for the next experiment. As the amounts of the indigo solution are not proportionate exactly to the amount of nitric acid present, it is necessary to standardize by means of different

strengths of the nitre solution and the results should be all marked on the label of the bottle. To make a nitrate determination in a water, 20 c.c. of the water are taken, some indigo solution run in and an equal bulk of pure sulphuric acid, the water being kept hot the while, and the indigo dropped in from a burette, until there appears a faint blue or blue green colour. The approximate strength of the water being thus known, a second determination is made and the value of the indigo obtained from the nearest number which has been determined previously in standardizing.

Estimation of Nitrates as Ammonia.-This may be done by either making the water strongly alkaline by dissolving in it little by little metallic sodium, then distilling until free from ammonia. The retort is cooled, a small piece of aluminium foil is dropped in and the retort allowed to stand over-night. The water is again distilled, any ammonia produced is derived from the reduction of nitrates or nitrites.

Estimation by the Copper-Zinc Couple is more convenient.-Pieces of clean zinc foil, well crumpled, are immersed in a solution of copper sulphate (3 per cent.), when the zinc is coated, the liquid is poured off, the couple washed with a little distilled water, and then any convenient quantity of the water poured on to the couple and the whole allowed to stand in a warm place over-night. The water may be distilled and the ammonia estimated, subtracting what has been previously found as "free" ammonia, but in most cases the ammonia may be estimated directly by taking 10-25 c.c. of the water, diluting it up 100 c.c., and then Nesslerising direct. Here again the same correction for ammonia pre-existing will have to be made.

(119) Hardness.

(A) Total Hardness.—100 c.c. of the water to be examined are placed in a suitable bottle which can be either stoppered or corked and the standard solution of soap run in from a burette little by little; after each addition the water is violently shaken to see whether it "lathers" or not. By a very little practice the stages of the process can be detected by the ear quite as readily as by the sight; so long as the soap solution is insufficient, the shaking produces a sharp sound, whereas directly the soap has used most of the earthy salts up, the sound changes to a soft muffled note.

Sufficient soap solution has been added when a good foam or froth has been produced which remains pretty well the same for five minutes. There is a difference in the time that it takes to soften waters which contain little or no magnesian salts; and those which contain magnesian salts; the latter are decomposed with considerable slowness nor is the end reaction so sharp. Waters which take per 70 c.c. more than 16 c.c. of the soap solution require to be diluted, for the results are found not to be accurate when such comparatively strong solutions of earthy salts are treated. In these cases therefore a preliminary trial is made and then the water diluted accordingly, the best dilution being that which uses from 14 to 16 c.c. of the soap solution per 70 c.c. of the diluted water.

(B) Permanent Hardness.-By boiling water holding in solution carbonate of lime, the carbon dioxide, that is the gas which dissolves carbonates of lime and magnesia, is expelled, hence these carbonates are thrown down as a precipitate. The boiling should be brisk and continued until at least a third of the water has disappeared, the water is then made up to its original volume by means of distilled water, filtered and treated with soap solution just as before, the result is returned as "permanent hardness," that is hardness due to soluble salts like lime or magnesic sulphates or nitrates.

By subtracting the number of degrees of "total hardness" from the number of degrees of "permanent hardness," the degrees of "temporary hardness" are obtained.

(120) Alkalinity.

It is of great use to measure the alkalinity of the water. 100 c.c. of the water are placed in a Nessler cylinder, coloured yellow by means of a solution of methyl orange and deci-normal hydrochloric acid run in from a burette until the reddish hue shows that the liquid has become neutral, each c.c. of deci-normal acid is equal to 005 grm., that is 5 milligrms. of lime carbonate, to which salt the alkalinity of water is usually due.

(121) Sulphates.

The quantitative determination of sulphates is very simple: a quarter of a litre or if the water should be soft, a litre of the