light on the question at issue. It is my intention to extend the work to other typical hydrocarbons in the hope that the gradual accumulation of experimental facts may at some future time provide a sure basis for a general theory of hydrocarbon combustion. Meanwhile, it seemed to me that the meetings of the Chemical Section afforded a fitting opportunity of communicating and discussing these new observations, of obtaining suggestions for future work, and possibly, also, of arranging some form of co-operation among those workers who are specially interested in this field of inquiry.

It seems to me unnecessary to make more than a passing reference to the theories which up to the present have been advanced to explain the mechanism of hydrocarbon combustion. I must, however, say a word with regard to two of them which involve the idea of the preferential' combustion, either of hydrogen or of carbon. The older idea, that in a defective oxygen supply the hydrogen of a hydrocarbon burns preferentially to the carbon, is unsupported by experimental evidence, and I suppose now hardly finds acceptance among chemists at any rate. On the other hand, the opposite view, that the carbon burns preferentially to the hydrogen, was put forward, originally by Kersten in 1861 I believe, to explain the well-known fact that when such a hydrocarbon as ethylene is exploded with just sufficient oxygen to burn the carbon to carbon monoxide, the cooled products consist of carbon monoxide and free hydrogen—

Professor Smithells in 1892 was led to indorse this view as the result of his analyses of the interconal gases of hydrocarbon flames.

It seems to me that the idea of preferential combustion,' whether of hydrogen or of carbon, is closely allied to the old doctrine of elective affinity,' and that it is hardly to be reconciled with modern conceptions of the nature and conditions of chemical change in a homogeneous system. Furthermore, it may be pointed out that the evidence usually adduced in support of the contention that carbon burns preferentially to hydrogen is wholly derived from experiments on the oxidation of hydrocarbons at very high temperatures, either in the flame, or in the explosion wave. Under these conditions it is practically impossible, by any means at our command at present, to distinguish the character of the primary oxidation in the case of a hydrocarbon, for since the velocities of all the reactions concerned are enormously great, the firal state of equilibrium is almost instantaneously established.

The experiments on the slow combustion of methane and ethane, which have led me to make this communication, have been carried out at temperatures far below the ignition-points of the gases that is to say, at temperatures where the oxidation velocities are sufficiently small to allow of their being easily measured. It is also important to observe that either of the hydrocarbons in question interacts with oxygen at temperatures below those at which the velocities of any of the undermentioned possible secondary changes become appreciable :

(iv) Reduction of CO2, or of HO, by carbon.

Therefore, by suitably choosing our temperature conditions we have been able to exclude the possibility of these reactions occurring, and so to prevent the complete masking of the primary reaction by secondary changes.

The details of these experiments either have been, or shortly will be, published elsewhere,' and we need therefore only here indicate the general character of the results.

In the first place, we should say that the mixtures of methane (or ethane) and oxygen employed usually contained just sufficient oxygen to burn the carbon of

1903.

1 Trans. Chem. Soc. (1902), 81, 535; 83, 1903.

SS

the hydrocarbon to carbon monoxide (e.g. mixtures of two volumes methane with one volume oxygen, or of equal volumes of ethane and oxygen).

The lowest temperature at which such mixtures of methane and oxygen interact, when sealed up in a borosilicate glass bulb at atmospheric pressure, and afterwards placed in a constant temperature air-bath, is somewhere about 300°; in the case of the mixtures of ethane and oxygen it is about 225°. At all temperatures ethane is oxidised much more rapidly than is methane, other conditions being equal.

Under such conditions a portion of the hydrocarbon is burnt to, finally, carbon dioxide, carbon monoxide, and steam, without any liberation of free hydrogen or separation of carbon, while a portion of the original hydrocarbon always remains intact.

Below are tabulated the analyses of the products from two typical experi

We next devised an apparatus in which the reacting gaseous mixtures can be continuously circulated day and night, at a practically uniform rate, (1) over a surface maintained at a constant temperature; and (2) through suitable washing and cooling arrangements for the removal of soluble or condensable intermediate products. A manometric arrangement enables us to take pressure records of the gas in the apparatus at regular time-intervals throughout a given experiment, which may often extend over many consecutive days and nights. The records so obtained show, in the case of both methane and ethane, a regular and continuous fall of pressure throughout the oxidation.

The experiments with methane reveal the fact that formaldehyde plays an important role as an intermediate product; that, indeed, the oxidation involves at least two distinct stages, namely :—

1. A primary oxidation to formaldehyde and steam—

2. The subsequent further rapid oxidation of the formaldehyde to carbon monoxide, carbon dioxide, and steam. This may best be considered as the result of two simultaneous reactions, namely :—

(b) H·C : 0 + 0 : 0 + H·C: 0 = 2CO + 2H2O.

Possibly the latter may involve the formation and very rapid decomposition of formic acid.

Thus
H

0: CH+0: 0+H·C: 0=2

In the case of ethane we are able to distinguish the successive formations of (1) acetaldehyde, and (2) formaldehyde, as intermediate products. The experimental results are consistent with the following view of the case, namely:— 1. That the primary oxidation involves the formation of acetaldehyde and

steam

CHş'CII, + O,=CH-CHO + H,O.

3

3

2. That the acetaldehyde is further rapidly oxidised to carbon monoxide, steam, and formaldehyde

3. That the formaldehyde suffers further oxidation as indicated above.

These views, it may be stated, are supported by experiments on the oxidation of acetaldehyde.

I wish it to be understood that I have provisionally adopted the explanations just given of the oxidation stages of methane and ethane as a convenient working hypothesis because they express most simply the observed facts. Professor Armstrong has recently given us 1 ી very suggestive general theory of combustion which embodies his dictum that chemical interchange and electrolysis must be regarded as interchangeable equivalent terms. Applied to hydrocarbons (e.g. methane) the theory involves the successive hydroxylation' of each hydrogen by an indirect process, the oxygen being transferred electrolytically across 'conducting' water, as indicated by the following scheme :—

The hydrogen peroxide formed being in part decomposed by heat, and in part acting as depolariser.

The hydroxylated molecules thus produced may decompose, as for instance :

(5) CH2(OH)2 = CH2O + H2O
formaldehyde ;

and then the formaldehyde is further indirectly oxidised to (1) formic acid (2) carbonic acid, thus:

(7) HO>C : 0 + OH, +0,2 = HO>C : 0 + H2O,

The formic and carbonic acids thus produced then decompose, as follows:-(8) HO>C: 0=CO+H2O

It does not come within the province of this paper to discuss this electrolytic theory of chemical change; it should, however, be pointed out that Professor Armstrong's views demand the formation of an alcohol in the primary oxidation of a saturated hydrocarbon. Although I have never failed to obtain a marked

1 Trans. Chem. Soc. 1903, 83, 1088.

formation of aldehydes in my experiments on methane and ethane, I have so far searched in vain for alcohols; if the latter are produced during the primary oxidation, they are very rapidly further oxidised to the corresponding aldehyde, which must be presumed to be more stable under these conditions.

The question now arises whether these reactions which undoubtedly occur at low temperatures also occur at the higher temperatures of hydrocarbon flames. My own view is this: the velocities of these 'low temperature' reactions will rapidly increase as the temperature rises, and so long as aldehydes can exist, aldehyde formation will occur. But aldehydes themselves decompose at high temperatures; thus acetaldehyde is known to yield carbon monoxide and methane

CH,CHO +CH,+CO

and similarly formaldehyde yields carbon monoxide and hydrogen-

H.CHOCO + H2

and possibly within certain temperature limits these reactions are reversible. The production of formaldehyde in the oxidation of methane, for example, will only be limited by the temperature at which formaldehyde is incapable of existence, whatever that may be.

We shall have to take into account similar considerations in discussing other probable changes, as, for example, (1) the further oxidation of aldehydes, and (2) the purely thermal decomposition of hydrocarbons. All these possible reactions call for further careful investigation. As yet we have so few wellestablished data that it seems premature to formulate general theories. The subject is very complex, and is beset with many and great experimental difficulties, but it is surely within our power to overcome them, especially if a sufficient number of workers will co-operate.

2. Fluorescence as related to the Constitution of Organic Substances. By JOHN THEODORE HEWITT.

A distinction must be made between substances which are simply coloured and those which exhibit the phenomenon known as fluorescence. Whilst both classes of substances select radiant energy of certain wave-lengths, the fate of this energy is different in the two cases. A merely coloured substance degrades the energy it absorbs to a confused mixture of relatively slow vibrations, so that the substance or its solution tends to rise in temperature. A fluorescent solution largely transforms the absorbed energy and emits it with an altered frequency, in most cases still sufficiently high for the emitted energy to appear as light.

Both the absorption and the fluorescent spectrum are composed of bands which in the fluorescent spectrum are usually broader than in the absorption spectrum. Dark-line absorption spectra or bright-line fluorescent spectra are not to be expected in the case of a solution; the molecules of the solvent must exert an influence on the vibrations of the molecules of dissolved coloured substance, and, this influence not being uniform for all the molecules of dissolved substance, both spectra can only be expected to consist of bands and not of lines.

In the case of a gas the emission spectrum varies with the pressure; should the gas be sufficiently rarefied, the molecules perform their vibrations in an unfettered manner and the spectrum consists of bright lines corresponding to definite rates of vibration. But on increasing the pressure of the gas the molecules must mutually influence one another, with the result that their rates of vibration are affected. Since at any instant different molecules will not be affected to the same extent, they will execute their vibrations at somewhat varying rates and the lines in the spectrum will broaden into bands. A fluorescent-line spectrum could only be found in the case of a gas; whether any sufficiently fluorescent rarefied gas exists appears very doubtful.

The ultimate cause of fluorescence has naturally attracted attention. Stokes 1 was inclined to attribute a peculiar sensibility to the molecules of substances exhibiting this phenomenon. Lommel started with the assumption that light of a certain frequency may give rise to vibrations of varying amplitudes in the molecules of a substance. If the frequency depends on the amplitude, the emitted light will not be homogeneous and the substance may be considered as fluorescent. Two grave objections to Lommel's theory are, that there seems to be no possibility of a source of light remaining homogeneous whilst it fades in intensity, and that all coloured substances should be fluorescent. Both deductions are at variance with actual facts.

3

Fluorescence must of necessity attract the attention of organic chemists, chiefly on account of the fact that so many fluorescent substances are organic compounds of known constitution. Richard Meyer attempted to connect the fluorescence of organic dyestuffs with the presence of certain atomic groupings which he termed 'fluorophors.' Amongst such fluorophors, the pyridine, pyrone, and paradiazine rings may be mentioned. For fluorescence to be developed it is necessary that the fluorophor be attached to heavy carbon groups, usually aromatic nuclei. Meyer's theory gives no explanation of the influence of solvents and of the differences frequently observed in the case of isomeric compounds.

4

The present author has started from a fundamentally different point of view, which may be stated as follows. If in the case of a tautomeric compound the passage from one to the other configuration can be effected by two equal but opposite atomic displacements, the molecules will vibrate between the two extreme positions of less symmetry, passing through the intermediate more symmetrical configuration. Energy absorbed when the molecules possessed one configuration could then be emitted when they had the other configuration; and as the two configurations would certainly correspond to different vibration frequencies, one has the necessary conditions for the exhibition of fluorescence.

Consider the fluorescence phenomena in the case of the following compounds :I. Fluoran, C0H1203

V. 45-Dinitrofluorescein, CH (NO2)2O5.

10

20

VI. 45-Dinitro-27-dibromofluorescein, CH(NO),Br.O..
VII. 27-Dinitro-4-5-dibromofluorescein, CH (NO2), Br2O.

Of these substances, I. is colourless, and in neutral solvents gives colourless, non-fluorescent solutions. It fluoresces, however, if dissolved in strong sulphuric acid. II. III. and IV. all fluoresce, especially in alkaline solution.

The alkaline solutions of V. VI. VII. do not fluoresce at all. Meyer's theory gives no explanation of these differences. The theory now brought forward agrees

Phil. Trans., 1852, 463.

Zeitschr. physikal. Ch. (1897), 24, 468.

2 Wied. Annalen, 3, 268.

Proc. Ch. Soc. (1900), 16, 3; Zeitschr. physikal. Ch. (1900), 34, 1-19. 5 Berichte (1891), 24, 1412; (1892), 25, 1385; Annalen (1882), 212, 349. J. Chem. Soc. (1900) 77, 1324; (1902), 81, 893.

Ibid. (1902), 81, 893.