When reduced to unit area the constant K becomes the evasion coefficient 8 of Bohr, or the coefficient of escape of Adeney and Becker. These authors, who have studied respectively the rate of escape of carbon dioxide from solution and the rate of absorption of the gases of the air in water, have assumed that they obtain complete mixing and have assigned definite values to ẞ or ƒ. I think that a study of the effect of varying the rate of agitation is necessary before limiting values can be assigned to the evasion constants of gases.

This work on the rate of absorption cannot be properly treated in a rapid survey, but it is of the greatest value in that it indicates that true significance of any observed concentration, e.g., of dissolved oxygen in the water of a stream or even of a lake or pond. Given motion of the water, such as will occur in a tidal or fairly rapid river, the rate of absorption of gases from the air is directly proportional to the saturation difference,

as

Brunner calls it. In other words, the difference between 100 and the observed percentage saturation with, say, oxygen, is a direct measure of the rate at which the water is, at the time of experiment, absorbing oxygen from the air as compared with the limiting rate. This is exactly what is required to be known in all work on river pollution, and it was the need of this information which led Dibdin to undertake his little-known work on the subject in 1894, which he boldly applied to the solution of his problem of the Thames.

We see that there is a limit to the solubility of a gas in water; that is, for every set of conditions a point is reached when the water will absorb no more gas. It is clearly necessary to ascertain that point for varying conditions. The effect of gas pressure was studied by Henry and Dalton. The other determining factors are temperature and the presence of other matters-such as the salts of sea-water-already dissolved in the water. Water will dissolve more gas at its freezing point than at any other temperature. All the gases as well as other solutes separate on freezing.

The solubility of gases of air in water has, since Bunsen's work in 1855, been determined by many workers whose results are on the whole very consistent. The effect of salts in reducing solubility of gases is well shown in the case of sea water.

For the actual determination, two classes of methods are available; either (1) measured quantities of gas and liquid are agitated at a constant temperature and the decrease of the gas phase noted from time to time until the mass of the residual undissolved gas is constant, whence the

mass dissolved in the water at the observed temperature and pressure can be deduced; or (2) the water is saturated with the gas and the resulting saturated solution is examined as to its content of gas. The former or absorptiometric method as usually applied is suitable only for pure gases, the analytical method is applicable to mixtures such as air.

A slide was exhibited showing the apparatus used by Adeney and Becker for the absorptiometric determination of the gases in distilled and sea water. It may be taken as a type of such apparatus but possess the novel feature of utilising the surface tension of the water for obtaining a bubble of constant surface and volume which also acts as a stirrer.

It is usual, following Bunsen, to reduce the results of determination of solubility of gases to the volume of gas at a pressure of 760 mm.-that is the gas itself, not gas plus water vapour-which one cubic centimetre of water would dissolve at the temperature of experiment. A few such values are given in the table. This value is usually styled the absorption coefficient a

Since air contains only 78 per cent of nitrogen, I per cent of argon, and 21 per cent of oxygen, with about 003 per cent of carbon dioxide the amounts of these gases dissolved from it are reduced from those given proportionated to these amounts, and since also air over water is always saturated with water vapour-at least at the interface-which as we have seen is the locus of the mutual action, the amounts are further reduced from this cause, a fact which is not always remembered.

One very convenient result of this is the fact || that water dissolves from air just about one-half as much oxygen as nitrogen, so that one gas is a simple measure of the other. This ratio is not maintained throughout the whole range of the temperature from o° to 100° C., but Tornoe showed that the proportion of oxygen decreased more rapidly that that of nitrogen as the temperature rose, and Winkler deduced the following equation::

Percentage of oxygen in)

air boiled out from water =35'47-00338 showing that the limiting percentages between o and 100 are 35'47 and 32:09, or ratios of oxygen to nitrogen plus argon : 182 and 1: 211.

The curve showing the solubility of atmospheric oxygen in water from 0-100 has been calculated from Winkler's tables for an observed pressure of air plus water vapour of 760 mm., hence the zero value at 100.

The volumes of gases dissolved are not large when considered as volume per volume, but the weights of gases dissolved are very small indeed. For example, distilled water which is fully saturated with air at 15° C. and normal barametric pressure, contains only one part per hundred thousand of Oxygen-o'001 per cent.

Ordinary river water would dissolve as much, but

-ea

water only contains about 80 per cent of this mount-0'0008 per cent. Yet, as we shall see ater, these very small concentrations have fareaching effects on some of the properties of water.

The absorptiometric method is useful as a means of determining the solubility of a gas in water, but we need an analytical method by which he actual amount of gas dissolved in any water can be determined.

A sample of water must first be obtained which nas not been exposed unnecessarily to the air. The process of filling a bottle in the ordinary way will obviously tend to aerate the water which is collected.

This can be avoided by using a canister which contains a bottle of much smaller capacity. Water entering through a central tube fills the bottle which overflows, a current of water passing through it until the canister is full. The air in the system escapes at the side tube and the bottle finally contains water which has not been in contact with air.

The oxygen and nitrogen in water can easily be removed by boiling, as is being done in the large flask which some here will remember having seen Mr. Streatfeild use in the laboratory. It is impossible to remove all carbon dioxide by this means, except by making the water acid and boiling for rather a long time. The method was used by Bunsen, who got over the difficulty of loss of water by expansion and of gas by solution in condensed water, by using a gas-collecting tube above the flask of water, which tube he separated by tying the rubber connection and then boiled out the air and sealed the top, afterwards removing the ligature and boiling the water to be examined, so that the gases all collected in the tube, which could be disconnected. Jacobson improved on this by substituting for the ligatured rubber tube a tube with a side hole, which, after the collecting tube was freed from air and sealed, could be pushed down into the water to be examined, which was then boiled and the gas collected and sealed off. This apparatus was used by Buchanan on the Challenger. It must have required much skill to L use on a small but lively vessel. Several useful pumping out methods have been used, e.g., by MacLeod, Dibdin, and Adeney, but it is desirable to avoid contact of mercury with the water to be examined, or loss of oxygen may occur.

1

The most practical gasometric method for oxygen and nitrogen is Winkler's. The gases are evolved in a stream of CO, generated from calcite by hydrochloric acid in the measuring vessel and collected over potash solution in a graduated tube, in which they are measured, and the oxygen absorbed by "pyro." This can easily be used on a steady ship and could be arranged for using on a lively one.

Of many

Oxygen, from its great chemical activity, can be determined by chemical means. methods the best is certainly Winkler's manganese process, which is well known. Carbon dioxide may be determined by the usual method of titration of the base combined with it by acid and methyl orange and of the free and bicarbonate CO, by sodium carbonate and phenol-phthalein. Another method which is applicable to a long series of determinations is by comparing the tint produced in the water in question by such indications of alkalinity with those produced in a

series of "buffer mixtures." Mixtures of boric acid and borax, mono- and di-hydrogen alkali phosphates, or other salts of weak di. or tribasic acids, may be used. The carbonate acid bicarbonate, bicarbonate-carbonate equilibrium, has been studied by Prideaux and the alkalinity of such mixtures is well defined. The experiment of passing carbonic acid into a very dilute solution of sodium carbonate coloured purple with a mixture of a-naphthol phthalein and phenolphthalein will illustrate the effect of an increasing proportion of acid to base. First the crimson colour imparted to phenol-phthalein will disappear as the liquid becomes less alkaline by absorption of carbon dioxide, then the blue of a-naphthol phthalein which continues until the point of absolute neutrality is nearly reached. We may now add a neutral point indicator such as neutral red, and until the gradual reddening as the concentration of carbonic acid increases. The reverse series of changes will occur if air is blown through the acid solution.

What, if any, is the significance of all this to anyone but the physical chemist? This is, of course, a vulgarly materialistic question, but such questions are asked.

It is a noteworthy fact that many of the people who have investigated this matter have not been at all interested in physical chemistry, but rather in oceanography, marine biology, public health, or engineering, and have struggled with a subject which fell necessity forced them to study as best they could.

All the animals of the sea and rivers except warm-blooded air-breathing ones have to depend upon the oxygen dissolved in the water for respiration and are able to pump into their gills a sufficiency of water for this purpose. Hoppe Seyler and Duncan, and more recently Addyman Gardner, have examined the respiratory requirements of a variety of fishes which as they have no definite body temperatures to keep up are much less than those of land animals. Gardner found that the respiratory quotient of brown trout went up so rapidly with rising temperature that the fish becomes incapable of sufficient muscular effort to force the required amount of water through its gills. It will be remembered that the amount of oxygen available in the water, unfortunately for the fish, decreases rapidly with the rising temperature. This explains the distress of fish in shallow water during very hot weather. Conversely, respiration slows down to hybernation point at low temperature. The rather high oxygen requirement of the Salmonidæ generally accounts for the inability of salmon to frequent the Thames, where they would, in their migration to and from the sea, encounter a zone of very low aeration than they could tolerate in the region of the sewage outfalls.

Natural water is not entirely dependent upon the air for its dissolved oxygen. It is well known that vegetable life in water will enable it to support a larger population of fish. Owing to the ability of green plants under the influence of sunlight to take up the carbon dioxide given off by animals, utilising the carbon for building their tissues and giving out the oxygen which is not needed. I have found enormous excesses of oxygen in water containing plenty of weeds, on a bright warm day-up to So per cent excess over saturation. This becomes greater as the day

advances and falls during the night partly by escape into the air and partly because it is spent in oxidising organic matter.

The carbon dioxide also has a very important effect on life in water. Very small difference of hydrion concentration due to changes in the carbonate-bicarbonate equilibrium seem to have an important function in determining the migration of various living organisms which if not themselves of economic value are food material for others that are.

A sea without dissolved gases would certainly be a dead sea.

The relative amounts of gases in water are small, yet if we take Sir J. Murray's estimate of the masses of the seas of the globe they should contain about 10 billion tons of dissolved oxygen if approximately separated. The proportion of carbon dioxide is about 98 mgrms, per litre, or in the whole ocean about 120 billion tons. The ocean has been shown by Schloesing to act as a great equaliser of the content of this gas in the atmosphere by reason of its carbonate-bicarbonate system.

Apart from its biological value in conserving higher aquatic life, dissolved oxygen is of importance in maintaining water in a reasonable state of purity, although there are strong reasons for sewage disposal by discharge into streams; but the introduction in a stream of a great mass of oxidisable matter frequently leads to putrefaction and offence. The changes which refuse undergoes in a river are, in sum, similar to those occurring in a dust-destructor-that is, complicated organic compounds are oxidised to stable and inoffensive products. Since they occur at temperatures several hundreds of degrees lower the rate is much less rapid and the intermediate stages are more readily discovered. In both cases a large proportion of oxygen is necessary for the combustion. In the fiery burning oxygen is obtained direct from the air; the mechanism of combustion is less simple than is shown in the usual equations. In a river all the necessary oxygen must first be dissolved, mostly from the air, partly perhaps from plants, and then interact with the organic matter through the agency of living organisms, bacteria, and others.

It frequently happens that through slowness of flow, inadequate volume, or other causes, the rate of absorption of oxygen in a sewage-polluted stream is insufficient for the desired changes to proceed with sufficient rapidity, and the products of incomplete oxidation produce nuisance, as will those from a destructor of faulty construction. In such cases, warning is given beforehand by a decrease in the proportion of dissolved oxygen, which as it is taken up by the polluting matter and the water becomes more unsaturated, can be absorbed from the air at a more rapid rate. If no dissolved oxygen is found in the water on analysis it means that the river has reached the limit of oxygen supply for its then condition of flow. Usually a more or less stable equilibrium is established at some point intermediate between saturation and zero where absorption from the air is balanced by the bio-chemical demand of the oxidisable matter.

Nitrogen and argon dissolved in natural waters appear to be as inert as the name of the latter implies but oxygen and carbon dioxide account for much of the geological change which leads

Thes

to the disintegration of rocks and the presence dissolved salts in the water. The boiling out carbon dioxide causes a separation of calcium carbonate which gives engineers so much troubis when unsoftened water is used in boilers. two gases are also responsible for much of the trouble experienced in the corrosion of iron azi of lead. This is too large a subject to touch o to-day, but I mention it to show that the gases in water may be harmful as well as beneficial.

Two practical considerations arise out of all this.

First. It is sometimes desirable to accelerate the process of absorption of gases from the air usually with a view to oxidation of matters in the liquid. This can be done if practicable—t] usually is not-by increasing the air pressure or more easily by increasing the area of the interface. In the most modern processes of sewage treatment this is done rather by passing air through the liquid in very small bubbles, thereby imparting to a relatively small mass of air a large surface (the usual activated sludge method of diffusing tiles) than by increasing tank area. friend and colleague, Colonel W. Butler, has suggested the inverse process of spraying the mixture of sewage and activated substance through the air. Another method, when area remains constant, is to promote mixing so as to reduce Brunner's layer of concentration gradient to a minimum. This has been successful at Sheffield, where Mr. Haworth has used it for purifying sewage on a very limited area of land.

MA

I

Second. It is frequently desired, for example, in steam raising to keep out air from water. air is not already dissolved in the water, the surface of exposure to air if such cannot be avoided. | must be kept as small and as undisturbed as possible. The surface temperature may be raised to decrease tension of the air gases and also their solubility and rate of absorption.

If the water already contains air, a rise of temperature or decrease of pressure with increase of surface will aid escape. Condenser water from a steam engine is usually gas free, as it is condensed over a large surface in a partial vacuum. Oxygen can be removed by chemical means. The water of a hot-water pipe system has been found to be oxygen free, with atmospheric nitrogen as a supernatent gas in radiators on the top floor of a large building. This removal of dissolved oxygen by corrosion of iron might be regulated at will by using large anodic surfaces of iron in a feed water system.

It may be noted that in the lungs, in a sewage works, on a metallic surface the most active gas of the air oxygen only becomes active when dissolved in water.

THE SPECTRA OF ABSORPTION OF PHOSPHORUS FOR X-RAYS.-Utilising the Siegbahn apparatus with revolving gypsum glass, Mr. Bergengren directed his researches to ammonium phosphate, phosphoric acid, and red commercial phosphorous, The anti-cathode of the bulb was constituted by a tungsten plate, the current not exceeding So milliampères under 16,000 volts. The wave length of the limits of absorption are thus different for different varieties of phosphorus and the chemical state of an element has therefore a certain influence on its spectrum of X-rays.Comptes Rendus, September, 1920.

A CONVENIENT FORM OF THE PERIODIC
CLASSIFICATION OF THE ELEMENTS
By C. R. NODDER.

THE accompanying arrangement of the elements, while possessing no essentially novel features, is probably the form in which their natural relationships can best be illustrated without recourse to a three-dimensional arrangement. It is practically Harkins' spiral arrangement adapted for a representation in one plane. In the form of a wall chart it should be a distinct improvement on the usual form for teaching purposes. Corresponding radii of the primary and secondary spirals are parallel to one another and are similarly numbered. In the case of the five missing elements (four of which lie on radii 1, 1, and 7') the atomic numbers are indicated in the circles. The missing rare-earth element of atomic number 61 is the third element after cerium, just as elements

C.R.N. 1920.

of atomic number 43 and 93 are missing after zirconium and thorium. Neglecting the rare earths, moreover, the third element after cerium is missing. The dotted lines in the diagram will serve to recall the interesting resemblances of lithium and magnesium, beryllium and aluminium, boron and carbon. The double circles are intended to indicate pleiads of isotopes which have been discovered by a study of radioactive transformations. Lastly, barbs have been placed between those elements whose chemical properties require a reversal of the order suggested by their atomic weights.

Cambridge, October 30, 1920.

TECHNICAL INSPECTION ASSOCIATION.-A Paper on "Some Features of Tensile Fractures" will be read by Dr. G. H. Gulliver at the Rooms of the Royal Society of Arts, on Friday, December 10, at 7.30.

A grm. of the rock powder is moistened with water and dissolved in dilute hydrochloric acid (or acetic, if the former acid is likely to act appreciably on the silicates) in a covered beaker till all effervescence ceases. Moderate heat is needed if the effervescence is so weak as to indicate a rock of dolomitic character. The solution is filtered through a 7-cm. filter, and the residue washed with water, or hot dilute hydrochloric acid should gypsum be present and the silicates resistant. The paper, with its contents, is ignited moist in platinum (blast needed only if the amount is considerable), and after weighing it is fused with sodium carbonate and analysed like a silicate rock. If it contains hydrous minerals the original water content of the residue is best determined by a separate test on a fresh portion of the sample and not by drying and weighing on a counterpoised filter and then igniting. The weight of water found should be added to that of the ignited residue in order to get the true weight of the insoluble matter.

If the ignited residue is wholly quartz, or is very insignificant in amount, it may be treated at once with a drop of sulphuric acid and a few drops of hydrofluoric acid, and the acids removed in the radiator. If then a slight residue still is visible, it is well to repeat the treatment with acids and evaporation, for it cannot be too insistently pointed out that quartz resists the action of hydrofluoric acid far more than many silicates, and several evaporations may be needed to volatilise the crystallised mineral unless it has first been reduced to a most impalpable state of division. When the weight after exposure for a few minutes to the full burner heat no longer changes, the loss represents silica. If a slight residue still shows (it is most likely to be mainly alumina), the subsequently obtained precipitate of alumina, &c., is added to it and ignited in the same crucible.

If the original residue is small, and accurate knowledge of its composition is desired, several grms. of the rock may be dissolved. In this case, the whole filtrate from the insoluble matter is best treated for dissolved iron, aluminium, and manganese, but then only an aliquot part of the subsequent filtrate should be used for the determination of calcium and magnesium.

In case the filtrate from the residue is coloured perceptibly by dissolved organic matter, a condition that may arise occasionally, this organic matter must be effectually destroyed before proceeding to the precipitation of iron and aluminium, for otherwise incomplete precipitation of

one or both will be the result. The complete removal of this disturbing material cannot always be effected by ordinary oxidising agents, but only by evaporation and heating to its carbonisation point. In doing this it is necessary, in order to avoid possible loss of iron as chloride, to evaporate with nitric acid to dryness a couple of times and then to heat over a free flame gently till the desired result is achieved. If this seems undesirable for any reason, it may be better to start with a fresh portion of the limestone and to treat it by one of the methods given under b below, for the bulk analysis.

b. When the Inorganic Residue is not to be Separately Analysed.

The usual case does not involve separate analysis of the inorganic residue. As said above, the residue may consist of clay or other silicates, carbonaceous matter, pyrite, &c., with or without quartz. There are two ways open to render the noncombustible part of this soluble.

a.

By solution in acid after strong ignitionSolution in acid after strong ignition is the best method to employ if the ratio of insoluble to soluble compounds is not less than that in an argillaceous limestone which is directly suited for burning to Portland cement-that is to say, if the silica does not much exceed 15 per cent and the oxides of iron, aluminium, and titanium together are not in excess of 6 per cent. The exact allowable limits have not yet been determined, nor is it known what the proportions may be in dolomites and dolomitic limestones. This is a subject for further investigation.

Limestones, however, in which the above percentages of silica, alumina, &c., are not exceeded may be converted in 10 to 15 minutes by a good blast capable of giving an effective temperature of 1100° to 1200° to a product that is wholly soluble in hydrochloric acid, provided the rock was first reduced to a very fine powder.

A grm. of the powder is heated in a covered platinum crucible by an inclined blast. If a limestone, the flame may be applied at once, as a rule, without fear of loss, or after short exposure to a full Bunsen flame. Highly magnesian limestones, if this method is applicable to them, must, however, be heated with the greatest caution, for their temperature of decomposition is far below that of limestones, and violent projection of material often begins far short of visible redness.

The strong heating is usually stopped after 10 or 15 minutes, when the shrunken product may have the appearance of a sintered or even clinkered mass that detaches for the most part readily from the crucible. (The changes that take place during the ignition comprise loss of all carbon dioxide, water, and carbonaceous matter; oxidation of all pyrite, with retention of the whole of the sulphur as calcium sulphate. Prolonged heating will gradually expel all the sulphur trioxide from the calcium sulphate and later the alkalies, which can be wholly volatilised in an hour or less by a powerful blast. With an inclined blast the alkalies condense in one part on the under side of the lid as a soluble and powerfully alkaline deposit, sometimes weighing several milligrms., but in the time above set no loss of alkali appears to occur). It is transferred to a beaker or evaporating dish and moistened with