WATER, as everybody knows, is a generic term which includes a great variety of different substances. But when we compare any two species we always find more of agreement than of difference in properties which suggest that all waters consist essentially of the same thing, which is only modified differently in the several varieties by the nature or proportion of impurities. This surmise is confirmed by the results of scientific inquiry. In all ordinary waters, such as are used for primary purposes, the impurities amount to very little by weight— as a rule to less than -jQ-th of 1 per cent.
Of all natural stores of water the ocean, is by far the most abundant, and from it all other water may be said to be derived. From the surface of the ocean a continuous stream of vapour is rising up into the atmosphere to be recondensed in colder regions and precipitated as rain, snow, or sleet, -&c. Some T\ths of these precipitates of course return directly to the ocean; the rest, falling on laud, collects into pools, lakes, rivers, &c, or else penetrates into the earth, perhaps to come to light again, or to be brought to light, as springs or wells.
As all the saline components of the ocean are non-volatile, rain-water, in its natural state, can be con-taminated only with atmospheric gases—oxygen, nitrogen, and carbonic acid. So we should presume, and so it is, except that these gases, having different rates of solubility, are not associated in rain-water (or natural water generally) in the proportions characteristic of their source. Thus, for instance, while atmospheric air contains 21 per cent, by volume of oxygen, a solution of air in water contains about 34-0 per cent, of that gas. Besides, rain-water, for the reason stated, contains perceptible traces of ammonia, combined as a rule, at least partly, with nitric acid, which latter is being produced wherever an electric discharge per-vades the atmosphere. This electrically-produced nitrate of ammonia forms no doubt the primary source of all organic nitrogen.
Lake waters, as a class, are relatively pure, especially so if the mountain slopes over which the rain collects into a lake are relatively free of soluble components. As an example we may refer to the water of Loch Katrine (Scotland), which is almost chemically pure, apart from small, but perceptible, traces of richly carboniferous matter taken up from the peat of the surrounding hills, which impart to it a faint brownish hue, while really pure water is blue.
River water varies very much in composition even in the same bed, as a river in the course of its journey towards the ocean passes from one kind of earth to others; while, compared with spring-waters, relatively poor in dissolved salts, rivers are liable to be contaminated with more or less of suspended matter.
Spring waters, having been filtered through more or

less considerable strata of earth, class, clear of
suspended, but rich in dissolved, mineral matter. Of ordinarily occurring minerals only a few are perceptibly soluble in water, and of these carbonate of lime, sulphate of lime, and common salt are most widely diffused. Common salt, however, in its natural occurrence, is very much localized; and so it comes that spring and well waters are contaminated chiefly with carbonate and sulphate of lime. Of these two salts, however, the former is held in solution only by the carbonic acid of the water, as bicarbonate of lime. But a carbonate-of-lime water, if exposed to the atmosphere, even at ordinary temperatures, loses its carbonic acid, and the carbonate of lime crystallizes out. The fantastically shaped " stalactites " which adorn the roofs and sides of certain caverns are produced in this manner.
In the relatively rare cases where a spring water in the course of its migrations meets with a deposit of common salt or other soluble salts, it dissolves more or less of these and becomes a salt-water. Most salt-waters are substan-tially solutions of common salt (chloride of sodium), associated with only little of salts of potash and magnesia. But there are exceptions; in the so-called " bitter waters" the dissolved matter consists chiefly of sulphate of magnesia and other magnesia salts.
Immense quantities of carbonic acid gas are being con-stantly produced in the interior of the earth, probably by the action, at high temperatures, of steam on the carbonates of lime and magnesia. Some of it collects and is stored up temporarily in cavities, but the bulk streams out into the atmosphere, invisibly as a rule, through what one might call the capillaries of the earth's body; but here and there it unites into veins and arteries and comes forth, it may be, as a mighty carbonic acid spring. Carbonic acid being one and a half times as heavy as atmospheric air, it may collect into pools or even lakes ; the " Dog's Grotto," near Naples, and the " Valley of Death," in Asia Minor, are illustrative examples ; but, carbonic acid being a gas, and consequently diffusible, such lakes can maintain their level only if there is a constant and abundant supply of the gas from below.
A far more frequent occurrence is that a mass of water and a mass of carbonic acid meet within the earth. As a rule, the pressure on the gas is more than one atmosphere, and the supply of the gas is abundant. The water then takes up considerably more carbonic acid than it would under ordinary atmospheric pressure, and if an outlet be provided, perhaps artificially by a boring, a frothy mass of carbonic-acid water comes forth as a fountain, sometimes of great volume. Such carbonic-acid waters are met with in many parts of the world,—chiefly, however, in Germany. The well-known Apollinaris water is an example. In this connexion we may refer also to the sojfioni of Tuscany— jets of steam charged with ammonia or vapour of boric acid, which condense in the air and collect into " lagoons " of (substantially) a very dilute solution of boric acid. Boric acid waters, however, appear to be entirely confined to a limited district in Tuscany to the south of Volterra.
In addition to its natural components, water is liable to be con-taminated through accidental influxes of foreign matter. Thus, for instance, all the Scottish Highland lochs are brown through the presence in them of dissolved peaty matter. Rivers flowing through, or wells sunk in, populous districts may be contaminated with excrementitious matter, discharges from industrial establish-ments, &c. Our instinct rebels against the drinking of a contam-inated water, and it guides us correctly. Not that those organic compounds are in themselves hurtful. An otherwise pure water, contaminated with, say, one ten thousandth of its volume of urine, might be drunk with perfect confidence. Yet the presence of especially nitrogenous organic matter is a serious source of danger, inasmuch as such matter forms the natural food or soil for the development of micro-organisms, including those kinds of bacteria which are now supposed to propagate infectious diseases. Happily nature has provided a remedy. The nitrogenous organic matter dissolved in (say) a river speedily suffers disintegration by the action of certain kinds of bacteria, with formation of ammonia and other (harmless) products of simple constitution; and the ammonia, again, is no sooner formed than, by the conjoint action of other bacteria and atmospheric oxygen, it passes first into (salts of) nitrous and then nitric acid. A water which contains combined nitrogen in the form of nitrates only is, as a rule, safe organically; if nitrites are present it becomes liable to suspicion; the presence of ammonia is a worse symptom ; and if actual nitrogenous organic matter is found in more than microscopic traces the water is possibly (not necessarily) a dangerous water to drink.
Wanklyn's method of water analysis is based upon these ideas. Starting with half a litre of the water, he distils it with a small quantity of carbonate of soda, and in the distillate determines the ammonia eolorimetrically with Nessler's reagent (see MEECTJEY, vol. xvi. p. 31). When all the saline ammonia (the free ammonia) is thus eliminated, alkaline permanganate of potash is added to the residue and the distillation resumed. Part of the nitrogen of the organic matter is eliminated as ammonia ; it is determined in the distillate, by Nessler's reagent, and reported as albumenoid am-monia. The results are customarily referred to one million parts of water analysed. To give an idea of the order of quantities here dealt with, let us say that a water yielding O'l part of free and 0-l part of albumenoid ammonia would be condemned by any chemist, as possibly dangerous. The peaty waters of the Scottish Highlands, however, contain much of both ammonias, and yet are drunk by the natives with perfect impunity.
All waters, unless very impure, become safe by boiling, which process kills any bacteria or germs that may be present.
Of the ordinary saline components of waters, soluble magnesia and lime salts are the only ones which are objectionable sanitarily if present in relatively large proportion. Carbonate of lime is harmless ; but, on the other hand, the widely diffused notion that the presence of this component adds to the value of a water as a drinking water is a mistake. The farinaceous part of food alone is sufficient to supply all the lime the body needs ; besides, it is questionable whether lime introduced in any other form than that of phosphate is available for the formation of, for instance, bone tissue.
The fitness of a water for washing is determined by its degree of softness. A water which contains lime or magnesia salts decom-poses soap with formation of insoluble lime or magnesia salts of the fatty acids of the soap used. So much of the soap is simply wasted; only the surplus can effect any cleansing action. An excellent and easy method for the determination of the hardness of a water has been devised by Clark. A measured volume of the water is placed in a stoppered bottle, and a standard solution of soap is then dropped in from a graduated vessel, until the mixture, by addition of the last drop of soap, has acquired the property of throwing up a peculiar kind of creamy froth when violently shaken, which shows that all the soap-destroying components have been pre-cipitated. The volume of soap required measures the hardness of the water. The soap-solution is referred to a standard by means of a water of a known degree of hardness prepared from a known weight of carbonate of lime by converting it into neutral chloride of cal-cium, dissolving this in water and diluting to a certain volume. A water is said to have " 1, 2, 3, . . . 11 degrees of hardness," if it is equivalent in soap-destroying power to a solution of carbonate of lime containing 1, 2, 3, . . . ii grains of this salt per imperial gallon.
That part of the hardness of a water which is actually owing to carbonate of lime (or magnesia) can easily be removed in two ways. (1) By boiling, the free carbonic acid goes off with the steam, and the carbonate of lime, being bereft of its solvent, comes down as a precipitate which can be removed by filtration, or by allowing it to settle, and decanting off the clear supernatant liquor. (2) A method of Clark's is to mix the water with just enough of milk of lime to convert the free carbonic acid into carbonate. Both this and the original carbonate of lime are precipitated, and can be removed as in the first case. See p. 409 infra.
Prom any uncontaminated natural water pure water is easily pre-pared. The dissolved salts are removed by distillation ; if care be taken that the steam to be condensed is dry, and if its condensation be effected within a tube made of a suitable metal (platinum or silver are best, but copper or block tin work well enough for ordinary purposes), the distillate can contain no impurities except atmospheric gases, which latter, if necessary, must be removed by boiling the distilled water in a narrow-necked flask until it begins to "bump," and then allowing it to cool in the absence of air. This latter operation ought, strictly speaking, to be performed in a silver or platinum flask, as glass is appreciably attacked by hot water. For most purposes distilled water, taken as it comes from the con-denser, is sufficiently pure.
Pure water, being so easily procured in any quantity, is used largely as a standard of reference in metrology and in the quantita-tive definition of physical properties. Thus a '' gallon " is defined as the volume at 62° F. of a quantity of water whose uncorrected


mass, as determined by weighing in air of 30-inch pressure and 62° F. of temperature, is equal to 10 lb avoirdupois. The kilo-gramme in like manner is the m,i3s of 1 cubic decimetre of water, measured at the temperature corresponding to its maximum density (4° C.). The two fixed points cf the thermometer correspond—the lower (0° C, or 32° F.) to the temperature at which ice melts, the upper (100° C. or 212° F.) to that at which the maximum tension of steam, as it rises from boiling water, is equal to 760 mm. or 30-inch mercury pressure. 30 inches being a little more than 760 mm., 212° F. is, strictly speaking, a higher temperature than 100° C, but the difference is very trifling. Specific heats are customarily measured by that of water, which is taken as = l. All other specific heats of liquids or solids (with one solitary exception, formed by a certain strength of aqueous methyl alcohol) are less than 1. The temperate character of insular climates is greatly owing to this property of water. Another physiographically important peculi-arity of water is that it expands on freezing (into ice), while most other liquids do the reverse. 11 volumes of ice fuse into only 10 volumes of water at 0° C.; and the ice-water produced, wdien brought up gradually to higher and higher temperatures, again exhibits the very exceptional property that it contracts between 0° and 4° C. (by about i-j^ni of its volume) and then expands again by more and more per degree of increase of temperature, so that the volume at 100° C. is 1 '043 times that at 4° C. Imagine two lakes, one containing ordinary water and another containing a liquid which differs from water only in this that it has no maxi-mum density. Imagine both to be, say, at 5° C, and cold winter weather to set in. Both lakes become colder, being cooled down by convection until they are at 4° C. The imaginary lake then continues losing heat in the same way ; in the real lake the colder water floats on the surface and the underlying mass of water can lose heat only by the slow process of conduction. The real lake retains its heat more tenaciously, but for the reason stated will draw a sheet of ice, while the imaginary lake is still on its way to 0° C. The latter, indeed, if the winter is short, may fail to freeze altogether, while the former does freeze superficially. In either the freezing is a slow process, because for every pound of ice pro-duced 80 pound-centigrade units of heat are set free. Let us now assume that, after even the imaginary lake had drawn a sheet of ice, warm weather sets in. In either case a layer of liquid wrater of 0° C. is formed on the ice, and through it the heat from the air travels, in the imaginary case by slow conduction, in the real case by quick convection. The real ice is the first to disappear, and the upper strata of relatively cold water will soon come up to 4° C. by the prompt effect of convection, wdiile in the case of the imagin-ary lake it takes a far longer time before all the mass has come up to 4° O. by conduction. From 4° C. upwards, heat is taken in at the same rate on both sides.
In former times water was viewed as an "element," and the
notion remained in force after this term (about the time of Boyle)
had assumed its present meaning, although cases of decomposition
of water were familiar to chemists. In Boyle's time it was already
well known that iron, tin, and zinc dissolve in aqueous muriatic
acid or vitriol with evolution of a stinking inflammable gas. Even
Boyle, however, took this gas to be ordinary air contaminated with
inflammable stinking oils. This view was held by all chemists
until Cavendish, before 1784, showed that the gas referred to, if
properly purified, is free of smell and constant in its properties,
which are widely different from those of air,—the most important
point of difference being that the gas when kindled in air burns
with evolution of much heat and formation of water. Cavendish,
however, did not satisfy himself with merely proving this fact
qualitatively; he determined the quantitative relations, and found
that it takes very nearly 1000 volumes of air to burn 423 volumes
of "hydrogen" gas ; but 1000 volumes of air, again, according to
Cavendish, contain 210 volumes of oxygen; hence, very nearly,
2 volumes of hydrogen take up 1 volume of oxygen to become
water. This important discovery was only confirmed by the sub-
sequent experiments of Humboldt and Gay-Lussao, which were no
more competent than Cavendish's to prove that the surplus of 3
units (423 volumes instead of 420) of hydrogen was an observa-
tional error. , (W. D.)



Footnotes

Compare CHEMISTRY, vol. v. pp. 483-485, and SEA WATER.