c.m. long and 4 c.m. in diameter, with a spherical bulb at the end. Tubes of this kind are commonly used to heat Victor Meyer's vapour-density apparatus. For brevity we have called it a "Meyer" tube.

The outside of the tube is thickly padded with asbestos wool, with the exception of the lower half of the bulb, and of a short length of 3-5 c.m. at the top, which serves as a condenser. The tube is filled with sulphur to a level of 3 or 4 c.m. above the bulb, and is heated by a Bunsen burner. The gas is adjusted so as to keep the level of the vapour near the top of the tube, which is covered with asbestos card to prevent the sulphur catching fire.

Our experiments have shown that a thermometer inserted in an apparatus of this kind will not attain the actual temperature of the vapour, unless it is protected from radiation to the sides of the tube, and from the condensed liquid which runs down the stem. lowering of temperature due to radiation, &c., may readily amount to upwards of 2° at the boiling-point of sulphur.

The

The method which we have adopted for screening the to its stem a short distance above the bulb. Two coaxial tubes are hung on to this umbrella to screen the thermometer from radiation. We have found that glass is not sufficiently opaque to heat radiation at this temperature. The inner tube at least should be of metal.

thermometer is to bind an umbrella of asbestos card on

To avoid superheating of the vapour, it is necessary to make sure that the level of the liquid sulphur stands well above that part of the bulb which is exposed to the flame.

Using these precautions, we have found that the temperature by normal air thermometer at constant pressure of the saturated vapour of sulphur, boiling freely under a pressure of 760 m.m. of mercury at o° C. and g = 980'61 C.G.S. (sea level in lat. 45°), is—

t=444 53° C.

The value given by Regnault is nearly 4° higher than this; but in the account which he gives of his experiments he has pointed out several sources of error, and it is evident that he did not place much confidence in his results.

The close agreement between the air thermometer experiments of 1887 and the present series leads us to conclude that the number above given is probably correct to a tenth of a degree, and that it may be safely used for standardising platinum thermometers.

The method which we recommend for standardising platinum thermometers is briefly as follows:-Observe the value Rs of the resistance in sulphur vapour in an apparatus such as we have described. Calculate the value of pts by the formula

100 (Rs Ro)/(R100R。). Find the temperature t of the sulphur vapour, corresponding to the corrected barometric pressure Ho, from the formula

t = 444'53+0'082 (H。-760).

NEWS

carefully determined as the boiling-point of sulphur. The fixed points given in the above have not been so They rest entirely on the assumption of the accuracy of the formula, and have not been directly referred to the air thermometer. We believe, however, that they are probably correct to o'r° C., and that they may be safely used to standardise thermometers of limited range, in cases where it may happen to be inconvenient to make use of the sulphur point.

In comparing the platinum and air thermometers between 0 and 100° C., observations were taken at intervals of 5° all the way up. The mean deviation of the observations from the parabolic formula (d) is only 0.006°. This corresponds to the limit of accuracy of the barometric readings, and there is no reason to suppose that the d-formula may not represent the difference even more closely than this.

The same platinum thermometer has been compared with several mercury thermometers standardised at Kew. The result seems to show that the Kew standard reads o'1° C. lower than our air-thermometer at 30'.

THE VARIATIONS OF ELECTROMOTIVE FORCE OF CELLS CONSISTING OF CERTAIN METALS, PLATINUM, AND NITRIC ACID.† By G. J. BURCH, B.A., and V. H. VELEY, M.A., the University Museum, Oxford.

THE description of the apparatus, the capillary electrometer, and the method of working are given fully in the paper. The following conclusions are drawn from the results of the experiments :

I. When the metals copper, silver, bismuth, and mercury are introduced into purified nitric acid of different degrees of concentration, and a couple made with platinum, the electromotive force of such a cell increases considerably until it reaches a constant and (in most

to the production of nitrous acid by the decomposition of the nitric acid, and the final value is considered to be due to the former acid only, while the initial value is due for the most part to the latter acid, though it is affected to a remarkable degree by the amount of impurity of nitrous acid, either initially present or produced by minute and unavoidable uncleanliness of the metallic strip and the containing vessel.

II. If nitrous acid has been previously added to the nitric acid, then the maximum E.M. F. is reached at once. III. If the conditions, namely, increase of temperature, of impurity, and of concentration of acid, are such as

Griffiths," Brit. Assoc. Report," 1890.

Abstract of a Paper read before the Royal Society.

would favour a more rapid solution of the metal, and consequently a more rapid production of nitrous acid, then the rise of E.M.F. is concomitantly more rapid.

IV. Conversely, if the conditions are unfavourable to the production of nitrous acid, the rise of E.M.F. is less rapid.

V. If any substance, such as urea, be added which would tend so destroy the nitrous acid as fast as it may be formed, then the rise of E.M.F. is extremely slow, being dependent upon the number of molecular impacts of the nitrous acid upon the surface of the metal.

Thus the results obtained by the electrometer and by the chemical balance are in every way confirmatory the one of the other.

The authors propose to conduct further investigations on cells containing other acids, to determine whether the action of them upon metals is conditioned by the presence of their products of electrolysis.

THIS paper is in continuation of a preliminary communication on the same subject; the main points contained in it are as follows:

I. The metals copper, mercury, and bismuth do not dissolve in nitric acid of about 30 per cent concentration (the acid commonly employed for the preparation of nitric oxide gas) and heated to a temperature of 30° C., provided that nitrous acid is neither present initially nor formed subsequently. To prevent this it is necessary in the cases of copper and bismuth to add a small quantity of some oxidising substance, such as hydrogen peroxide or potassium chlorate, or, as less efficacious, potassium permanganate, or to pass a current of air or, lastly, such a substance as urea, which destroys the nitrous acid by its interaction.

II. If the conditions are such that these metals dissolve, then the amount of metal dissolved and the amount of nitrous acid present are concomitant variables, provided that the nitric acid is in considerable excess. Change of conditions, such as concentration of acid and variation of temperature, which increase the former increase also the latter.

III. If the conditions are such that these metals dissolve, it would appear that the metallic nitrite is at first formed, together with nitric oxide; the former is decomposed by the excess of nitric acid to liberate nitrous acid, while the latter reduces the nitric acid to form a further quantity of nitrous acid.

Eventually the net result is the product of two reverse chemical changes represented by the equations

(i.) 2NO+HNO,+H20=3HNO2,
(i.) 2HNO2=2NO+ HNO3+ H2O.

The nitrous acid is thus destroyed as fast as it is generated.

IV. If the conditions are such that metals dissolve in nitric acid, then nitrous acid is invariably the initial product of reduction.

V. The metals copper, mercury, and bismuth dissolve very readily in a 1 per cent solution of nitrous acid; under these conditions nitric acid present in slight excess interferes with, rather than promotes, the chemical change. This result is probably due to the greater stability of nitrous acid in the presence of nitric acid.

VI. Hydrogen gas reduces nitric to nitrous acid in presence of cupric or lead nitrate; it also converts

* Abstract of a Paper read before the Royal Society.

mercuric into mercurous nitrate, but does not produce any change in solutions of bismuth and zinc nitrates dissolved in nitric acid.

FRICTION OF GASES IN PIPES.

By Dr. L. C. LEVOIR, Polytechnicum, Delft.

THE introduction of gaseous fuel in factories compels lecturers in technical colleges to discuss a series of quite new laws. In all the old treatises on Physics and Chemistry no experiment is related that might give a large audience an idea of the diminution in pressure on conducting gases through tubes. In the CHEMICAL NEWS for August 3rd, 1861 (vol. iv., p. 60), and Dingler's Polytechnisches Journal (German), 1861, I described an experiment to show that friction of smoke in a chimney is higher in a conical one when the widest end is used as inlet.

Two small gas-flames on the same T-shaped gas-pipe, with very low pressure of the gas, gave me an opportunity to show that one of the flames was nearly invisible on the opposite side from where the chimney (1 metre in length, and 2 and 5 centimetres in diameter) drew with the highest force.

But there is a still more sensitive test for difference in exhausting power, that is, in every Bunsen's burner of great size the inlet for the air must be regulated in the stand, that the flame cannot burn on the top, but detonates always downwards from too great admixture of air. A small end of tube placed in the air-holes directly stops that explosion, by causing friction therein and diminishing the quantity of air admitted.

Evidently this is a way to show to a large audience the friction of smoke in chimneys and gas in tubes, which is so difficult a point both in the education of physicis and engineers.

ON GASEOUS ILLUMINANTS.* By Professor VIVIAN B. LEWES.

I.

JuST 200 years after Van Helmont (in the seventeenth century) first used the term "" gas "to describe aëriform bodies, Faraday defined a gas as being the vapour of a volatile liquid, existing at a temperature considerably above the boiling-point of the liquid, and that the condensing-point of the gas was merely the boiling-point of the liquid producing it. This definition was contested at the time, as several of the gases had not then been condensed; but now it is known that the condensation of any gas to the liquid form is merely a question of suffi ciently intense cold, and pressure.

Hydrogen and the gaseous compounds of carbon and hydrogen have so strong an affinity for the oxygen contained in the atmosphere, that the heat emitted by a burning match is generally sufficient to determine combination between the gases; and where the heat evolved by the combination is sufficient to raise the gases or vapour to incandescence, the phenomenon of flame is the result. Some flames have the power, under certain conditions, of emitting light, while others have no photometric value; and it is a matter of the gravest importance to the gas world that as clear a conception as possible should be obtained of the conditions and cause of luminosity in flames. A visible flame may either be solid-that is to say, composed of a solid mass of incandescent particles-or it may have a distinctive internal structure, and show zones in which varying phases of combustion are taking place; and it is to this latter class

* Abstract of the Cantor Lectures delivered at the Society of Arts Communicated by the Author.

that all flames produced by a gas issuing from a burner belong.

NEWS

His theory, however, has gradually been altered by frequent quotation, until it is more often given as "the presence of solid particles suspended in the flame (or in immediate contact with the burning gas) is essential to its luminosity - an idea which Davy never had, as is shown by him later in the paper defining flame as follows:-" Flame is gaseous matter heated so highly as to be luminous;" and again: "When in flames, pure gaseous matter is burnt, the light is extremely feeble." Moreover, he alludes to "common flames "-evidently meaning the flames of candle, lamps, or gas; in all which cases I think it can be proved beyond a doubt that his theory, as expounded by himself, was perfectly correct. On June 11, 1868, Professor E. Frankland read a communication before the Royal Society, in which he described experiments which led him to doubt Sir Humphry Davy's theory. He points out that the deposit of soot formed when a cold surface is held in a gas or candle flame is not pure carbon, but contains hydrogen, which can only be got rid of by prolonged heating in an atmosphere of chlorine. Also that many flames possessing a high degree of luminosity cannot possibly contain solid particles. Arsenic burnt in oxygen gives a bright white light; yet as arsenic volatilises at 180° C., and the arsenic trioxide forms at 2180 C., it is evident that at the temperature of incandescence (which is at least 500° C.) there can be no solids, but simply vapours present in the flame; and for the same reason, the intense light resulting from the burning of phosphorus in oxygen cannot be explained by the solid particle theory. From these results Dr. Frankland considers that "incandescent particles of carbon are not the source of light in gas and candle flames, but that the luminosity of these flames is due to radiations from dense but transparent hydrocarbon vapours ; and he further shows that non-luminous flames, such as that produced by carbon monoxide and hydrogen, can, when burning in an atmosphere of oxygen, be rendered luminous if the ordinary atmospheric pressure is increased to 10 atmospheres, so as to prevent or retard as far as possible expansion during combustion. From Dr. Frankland's experiments, there is no doubt that the luminosity of a flame is increased by pressing around it the atmosphere in which it is burning, and also that rarefaction has the opposite effect-a point also worked at by Davy; but his experiments do not show that incandescent particles of carbon are not the principal source of luminosity in a gas flame. He also shows that the higher the density of the vapours present in a flame, the more likely is it to be luminous.

In 1874, Soret attempted to demonstrate the existence of solid particles in a luminous hydrocarbon flame, by focussing the sun's rays on the flame, and examining the reflected light by means of a Nicol prism; but neither his research nor that of Burch, who repeated his experiments, using the spectroscope instead of the prism, showed more than that solid particles are present. Herr W. Stein, in considering Dr. Frankland's objections to Davy's theory, pointed out that the soot which is deposited from a candle or gas flame, and which Frankland looked upon as a condensed hydrocarbon, contains 991 per cent of carbon and only o'9 per cent of hydrogen, which is about the quantity of hydrogen one would expect to be occluded by carbon formed under these conditions, and he also pointed out that if the soot were a heavy hydrocarbon condensed by a cold surface, cooling the vapour present in the flame, it ought to again become volatile at a high temperature, which it does not. The next steps in the

controversy were the attempts made by Hilgard, Landolt, and Blochman to trace the actions taking place in various flames by withdrawing the gases from different parts of the flame and determining their composition. Exhydrogen which burns first, whilst the heavy hydrocarbons become gradually reduced by the heat of the flame into simpler compounds until in the luminous zone of the flame they are broken down into carbon and methane, and it is the carbon in excessively minute particles which at the moment of liberation is heated to incandescence, and " principally" gives the light of the flame-the marsh gas originally present, and also that formed from the heavier hydrocarbons, adding its quota to the luminosity by still further decomposition during combustion, and finally becoming carbon dioxide and water. In 1876, Dr. Karl Heumann made a most important contribution to the theory of luminous flames in some papers published in Liebig's Annalen, in which he carefully went over the work of previous observers, and, by a large number of original experiments, proved that Davy's theory was correct, but that other causes also affected the degree of luminosity in a gas or candle flame.

In the ordinary atmospheric burner in which a mixture of coal-gas and air burn with a non-luminous flame, it was supposed that the admixture of air, by supplying oxygen to the inner portion of the flame, caused immediate and complete oxidation of the hydrocarbons, without giving time for the liberation of carbon in the flame, and consequenily luminosity. More modern researches, however, have proved this to be utterly wrong. The loss of luminosity is due to two causes-first, to the diluting action of the air introduced; secondly, to the fact that when a gas is so diluted it requires a far higher temperature to break up the hydrocarbons present than when the gas is undiluted, and therefore the temperature which serves to liberate carbon and render the undiluted gasflame luminous, is totally insufficient to do so in the diluted gas. Consequently the hydrocarbon burns to carbon dioxide and water without any such liberation, and hence with a non-luminous flame. The truth of this theory can be easily proved by the fact that diluting the gas with nitrogen, carbon dioxide, or even steam, serves to render it non-luminous, and therefore more rapid oxidation has very little or nothing to do with it, while the non-luminous flame can again be rendered luminous either by heating the mixture of air and gas just before combustion, or by heating the air with which the gas is diluted. This being so, it is evident that in the nonluminous flame we have the same hydrocarbon present as in the luminous flame; and anything that will tend to break them up, and liberate the carbon before the hydrocarbons are consumed, should again make the flame luminous. That heat will do this has been already shown; but it can be demonstrated in a still more striking way. It is well known that chlorine gas and bromine vapour will both support the combustion of a gas containing much hydrogen, but that the combustion is very different from that of the same gas burning in air, as the chlorine or bromine, having no affinity for the carbon, combines with the hydrogen only, and deposits the carbon in clouds of soot; in other words, at the temperature of flame, chlorine will break up the hydrocarbons and liberate solid carbon. If now a small quantity of chlorine is led into the non-luminous Bunsen flame, it at once becomes luminous; proving conclusively that luminosity is due to solid particles of carbon liberated in the flame. Again, Heumann points out that a small rod held in the luminous flame becomes rapidly covered on its lower side with a deposit of soot; that is to say, the soot is present in particles in the flame, and the uprush of the gas drives it against the rod and deposits it there. If the soot were present in the flame, as Frankland supposed, in the state of vapour, and the rod merely acted by cooling and con. densing it, the soot should be deposited on all sides of the