oil thus treated sets into a pale straw-coloured, hard mass in an hour or two; while cotton and other seed oils assume a deep orange-red colour, and do not set like olive oil. He asserts that there is a regular gradation of colour, according to the percentage of adulteration, and that by imitation mixtures approximate quantitative results may be obtained. The amount of delicacy under the most favourable circumstances does not appear to be more than 5 per cent.

A similar oxidation test is that of saturating sulphuric acid with hyponitric acid, adding it to the oil, and noting the time in which it sets to a solid mass; 7 grms. of the acid thus prepared are added to 9 grms. of oil; at 5° olive oil sets to a solid mass in ten minutes, and after twenty-four hours is a white hard mass.

Arachis oil, rape oil, and cotton oil get solid later in about an hour; while sesame oil after twenty-four hours is still as soft as honey. Mixtures of oils of slow solidifying properties with olive oil, are in proportion to their percentages slower of coagulation. Here, as in all other cases, a pure sample of oil for comparison is an essential. M. Lipowitz has proposed the following as a special test for poppy oil: If 1 part of chloride of lime be added to 8 parts of olive oil, the latter, if pure, separates completely into two layers at the end of four or five hours, the temperature being from 17° to 18°; but if it is mixed with an eighth or more of poppy oil, the separation is incomplete, and takes place with extreme slowness. According to M. Lailler, every olive oil must be considered false which, when mixed with one-fourth of its weight of chromic acid, does not at the end of twenty-four hours present a perfectly opaque liquid. According to Chevallier, in one part of France the oil is sophisticated with honey, an adulteration not likely to take place in England. It is easily discovered, for the oil has only to be shaken up with water, and the water separated and submitted to the usual tests for

sugar.

PART IX.-EXAMINATION AND ANALYSIS OF

WATER.

§ 316. Pure water neither exists in nature nor even in the laboratory of the chemist, save on those rare occasions when, with immense expenditure of time and labour, water is purified either by repeated distillation over permanganates in a vacuum, or made synthetically. Nevertheless, however difficult it is to obtain even an ounce of water which shall consist of 1 part of hydrogen and 8 parts of oxygen by weight, and no other admixture, it may yet be very easily obtained sufficiently pure to warrant the epithet "pure" water-i.e., containing impurities only to be detected by reagents of great sensibility, or, what amounts to the same thing, by operating on a large quantity of water. In the analysis of water, therefore, it need scarcely be added that it is not the water per se which the chemist really analyses; but his researches are directed with the object of unveiling and determining the nature, and where possible the amount, of whatever may be present, foreign to water, whether in suspension or solution, whether of mineral origin or as one of the myriad forms of "life." The experimental and analytical methods in use mainly fall under the following divisions—

I. EXAMINATION BY THE SENSES.-Smell, Sense of Taste, and General Appearance.

II. PHYSICAL EXAMINATION.

III. CHEMICAL METHODS.

IV. BIOLOGICAL.-Embracing A, microscopical appearances; B, cultivation of fungi and dormant germs; C, experiments on animals and human beings; D, experiments on fish.

I. EXAMINATION BY THE SENSES.

§ 317. Water that is evidently turbid, that possesses an odour and an unpleasant taste, requires no analytical processes to condemn it entirely; such a water is unsuitable for drinking purposes. A water that even possesses any one of the enumerated bad qualities will, as a rule, be found to hold in solution sufficient impurities to make it decidedly objectionable. Most drinkingwaters when looked at, or tasted, or smelt, without special precautions, have neither colour nor odour; on the other hand, all water, if viewed through a sufficiently deep stratum, possesses colour.

Colour.-To ascertain the colour of water, it is usual for analysts

to be provided with a colourless glass tube, at least 2 feet in length, having the ends closed with plate glass, and a small opening in the side of the tube through which to pour the water. A cheaper method of securing an aperture through which to introduce the water is to have a segment cut out of one of the glass discs, or a segmental section out of the end of the tube itself; the most convenient diameter of the tube is 2 inches, but one greater or smaller will answer the purpose. To make an observation, the tube is half-filled with the water to be examined, and then directed towards a white surface, which may be a white cloud in the sky or an equally illuminated sheet of paper. The air-filled space above the water then affords an excellent semicircular disc of comparison, and renders it easy to detect the slightest shade of colour. The purest waters have the slightest tinge of blue; the next in order of purity have a just distinguishable shade of green. Decided green tints, London fog hues, amber yellow, and brown tints are those possessed by waters tinged with peat, containing suspended matters, of second class composition, or those of considerable impurity.*

Smell.-Half a litre of the water or more is warmed in a large corked or stoppered flask to 38° [100° F.]; a long glass tube of three-quarters of an inch in diameter is now inserted, and the water sucked up once or twice so as to wet the side of the tube thoroughly; then, without taking the tube out of the flask, one nostril is applied to the orifice of the tube, the other closed by the finger, and deep inspirations or "sniffs" taken.

Another simpler plan is to warm a quantity of the water, without removing the stopper, up to the temperature given, then shake, remove the stopper, and smell; a putrid odour denotes decomposing animal or vegetable matter. If the sample is much polluted by fresh sewage, a urinous odour is not unfrequently distinct. But, again, it may be specially noted that water quite unfit to drink may have no odour, hence the usefulness of the test is limited. A positive smell teaches volumes-a negative result is of little value.

Taste.-A few waters, and a few only, have a decided taste. It is scarcely to be recommended that analysts should taste samples derived from fever-stricken localities; but, on the other hand, when there is no suspicion of the samples having been the

*Messrs. Crookes, Odling, and Tidy, in their report on the London waters supply for 1881, describe an ingenious "colour meter," consisting of two hollow wedges filled, one with a brown and the other with a blue solution. Any desired combination of green and blue may be made by sliding the wedges across each other in front of a circular aperture in a sheet of metal, and thus imitating the tint of water under examination; each prism is graduated from 1 to 50, the figures representing millimètres of the thickness at that particular part of the prism.

cause of any illness, the palate may detect some not unimportant peculiarity.

II. PHYSICAL EXAMINATION.

$318. The physical examination is mainly microscopical. Dr. W. Russell and W. Laplace* have recently discovered, it is true, that with a column of pure water 6 feet in length there is a distinct single absorption-band; and hence it is probable that, at all events, waters containing desmids and green vegetable cells generally would show particular spectra, but this has not yet been worked out; it will be more convenient for our purpose to consider the microscopical appearances later. (See p. 340.)

III. CHEMICAL METHODS.

§ 319. A complete examination by chemical processes embraces the following determinations :

1. Total solid residue.

2. Estimation of the halogens, chlorine, and occasionally iodine, and in a few cases bromine.

3. Phosphates.

4. Nitrates and Nitrites.

5. Sulphates.

6. Oxygen consumed in the Forchammer process.

7. Free and albuminoid ammonia.

S. Hardness.

9. Alkalinity.

10. Organic Analysis

nitrogen.

Estimation of organic carbon and

11. Mineral analysis of water.

The ordinary analyses, sufficient in most cases to pronounce an opinion as to the fitness of a water for drinking purposes, embrace only 1, 2, 3, 4, 6, 7, and 8.

1. Total Solid Residue.-By the total solid residue of a water is meant the substances in solution, as determined by drying up a measured portion, and weighing the dried residue; if the water contain suspended matters, it should first be filtered, and a portion of the clear filtered liquid taken. The amount suitable for this determination depends upon the characters of the water. The soft Devon waters yield a very insignificant residue from 100 cc., and to obtain trustworthy results, at least a quarter of a litre is required; while, on the other hand, with calcareous waters, good results may be always obtained from 100 cc. With waters the characters of which are unknown, it will be best to operate on * Journal of the Chemical Society. April, 1881.

a quarter of a litre, or (if working with English measures) onetwentieth of a gallon. The water may be placed in a platinum dish, and evaporated down to a small quantity over a ring burner, taking care that the liquid in no case boils or even simmers; the last drops are driven off on the water-bath. It is recommended by the Society of Analysts to heat the residue up to 104°•4(220°F.) in the air-bath, and then to cool under a desiccator; but with waters of unknown composition, it will be best to weigh the residue, which has not been exposed to a greater heat than 100°, for it is always open to the chemist to expose the residue thus obtained to higher temperatures. The examination of the solid matters by the eye will often not unfrequently reveal much. Iron gives a coppery lustre to the dish, manganese a green to the ash, and very pure waters leave a residue almost white. The dish with its contents is next heated to a low redness, by the aid of a good Bunsen's burner, furnished with a rose, and then cooled and weighed. Note should be taken of any blackening or scintillation. The dish is again cooled and weighed, the loss of weight being returned as loss on ignition, and this final residue is dissolved in the manner to be described, and used for the qualitative determination of the phosphates.

2. Estimation of the Halogens.-The estimation of chlorine is an essential part of the ordinary scheme of water analysis; that of iodine is rarely (perhaps too rarely) performed, while so few waters contain an estimable amount of bromine, that it need not be here described.

Chlorine.-Chlorine exists in ordinary waters in the form of sodic chloride, and occasionally a small portion of the total chlorine is combined with potassium. It is always estimated volumetrically by a standard solution of silver-nitrate (See Appendix), using as an indicator neutral potassic chromate. Nitrate

of silver in presence of potassic chromate and alkaline chlorides (when the solution is neutral) first uses up or decomposes all the chlorides, and then attacks the chromates. Chloride of silver being white, and chromate of silver being red, the formation of silver chromate is indicated immediately by a red colour. At least 100 cc. of ordinary water (or, if grains are worked with, 140 grains) are to be taken for the determination of chlorine. With much-polluted waters, with those near the seashore or other places in which the ground is impregnated with salt, such a quantity may be inconvenient, and it will be necessary then to dilute with distilled water, taking of the diluted liquid a known quantity. In any case, the water is put into either a white porcelain dish or a beaker standing on a white slab. 1 cc. of the chromate solution (or 15 grains) is added to the water, and

!

1