November 23, 1893] 



NATURE 



91 



ihat this yellow patch in the flime is dae to glowing carbon in 

 a solid and very finely-divided state. The very familiar fact 

 that a cold object introduced into the yellow part b'C>mes 

 coaled with a black solid dep isit, composed almost wholly of solid 

 carbon, confirms this view. That this carbon or soot is shIkI in 

 the tlame, is shown by the fact that it is deposited as a solid even 

 when a highly-heated object is jilaced in the flame, and there 

 are other proofs — some of them very pretty — which I cannot 

 show for lack of time and of a means of magnifyinti. 



Tiiis ex|)lanation is due to Davy, and constitutes hi^ most 

 celebrated discovery on the subject of flame. He desciibes it 

 in the following words : — 



" When a wire-gauze safe-lamp is made to burn in a very 

 explosive mixtui e of coal-gas and air, the light is feeble and of 

 a pale c 'lour, whereas the flime of a current of coal-gas burnt 

 in the atmosphere, as is well known by the phenomena of the 

 gas-lights, is extremely brilliant. ... In reflecting on the 

 circumstances of the two species of combustion, I was led to 

 imagine that the cause of the superiority of the light of the 

 stream of coal-gas might be due to the dcconiposiiion of a part of 

 the gas towards the inteiior of the flame where the air was in 

 smallest quantity, and the dep 'sition of solid charcoal wliich, fi st 

 by li^igiiilioii, andafterwards l>y its <:<?;;///« j//(7«, in creased in a hi h 

 degree the intensity of the light ; and a few experiments soon 

 convinced me that this was the true solution of the pri)i>Ifm. 



"I held a piece of wire-gauze of about 900 apertures to the 

 square inch over a stream of coal gas issuing from a small pipe, 

 and inflamed the gas above the wire-gauze which was abn 1st in 

 contact with the orifice of the pipe, when it burned with its 

 usual bright light. On raising the wire-gauze so a> tocau-e the 

 gas to be mixed with more air before it inflamed, the lii^ht 

 became feebler, and at a certain distance the flame assumed tnc 

 ]>iecise character of that of an explosive mixture burning within 

 the lamp, but though the light was so feeble in this last case, 

 '.he heat was greater than when the light was much more vivi^l, 

 and a piece of wire of platinum held in this feeble blue flame 

 became instantly white hot. 



"On reversing the experiment by inflaming a stream ofcoal- 

 gis and iiassing a piece of wire-gauze gradually from the summit 

 of the flame to the orifice of the pipe, the result was s'dl r,()re 

 instructive, for it was found that the apex of the flame inter- 

 cepted by the wire gauze afifirded no solid charcoal, but in 

 passing it downwards solid charcoal was given off in consider 

 able quantities, and prevented frqin burning by the cooling 

 agency of the wire-gauze ; and at the bottom of the flame, where 

 tlie gas burnt blue in its immediate contact with the atmosphere, 

 charcoal ceased to be deposited in visible quantities." 



Only one attempt has been made to disturb the conclusion 

 here draun by Davy. In 1S68 Prof. Edward Frankland, to 

 whom we are indebted for many important discoveries respcc - 

 ing flame, came to the cimclu>ion that the light-giving agen y 

 in flames was not solid carbon, hut certain complex vaporous 

 compounds of carbon and hydrogen. I regret very much that 

 time will not admit of my detailing the evidence in favour ol this 

 vie>v, or the counter evidence by means of which most chemists 

 have been persuaded that Davy's explanation was, after all, the 

 correct one. It is, however, right to remaik that Prof. Frank- 

 land not only adheres to his own view, but promises to adduce 

 further evidence in Us favour. 



Let us for the present, at any rate, stick to the opinion of the 

 majority, and a<lmit that the bright light of ordinary flames is 

 due to incandesc nt particles of solid car'ion. The next ques- 

 tion is, How does this carbon become separated? 



This question is dealt with by Davy, but in language of some 

 ambiguity. lie says, "I was led to imagine" .... that it 

 " might be due to ^he deconiposiliun of a part of the gas to^^ ards 

 the interior of the fl ime where the air was in smallest quantiiy, 

 and the deposit i n of solid charcoal which first by its ignition, 

 and afterwards by its combustion increased in a high degree the 

 intensity of the light." 



Whatever these words may have been intended to mein, or 

 whatever interpretation is the fair one, it is certain that Davy's 

 explanation was soon presented as if it implied lack of air to be 

 the chief cause of carbon separation. As there was a large 

 quantity of hydrocarbon, and only a small amount of oxygen in 

 the central parts of flame, the hydrogen, it was said, being the | 

 more inflammable element, will seize upon this oxygen and leave I 

 the carbon uncombined. The fact that this veision was given 

 by Faraday len Is som • countenance to the belief that it was a | 

 fair representation of Davy's view. ) 



Now this docirine was really incompatible with fans known, 

 though apparently not widely known, at the time. I have already 

 referred to the fact that Ddton at the beginning of the century 

 showed that when a hydrocarbon is exploded with a supply of 

 oxygen insufficient to burn both the hydrogen and the carbon, 

 it is the carbon, and not the hydrogen, which has the preference. 

 If, therefore, we f )llow Davy in regarding flrme as a tethered 

 explosion, we cannot explain the separation of carbon as being 

 due to the preferential coadnistion of the hydrogen. This fact 

 was clearly pointed out by Kersten in 1861, bu; notwithstanding 

 this, and other investigations tending to the same conclusion, the 

 old view has somehow kept its ground down to the present day. 

 We muU now turn to the alternative explanation. It is sup- 

 plied by the words, and I think, by the intention, of Davy. 

 He says that the carbon separation might be due to the decom- 

 position of the gas towards the interior of the fl^me. If this 

 decomposition be not due to chemical action, it mu-t be due to 

 heat ; and certain it is that hydrocarbons when strongly hea'ed 

 do decompose, and do deposit carbon. Here is a result of this 

 action occurring on the large scale. This gas-carbon, as it is 

 called, is deposited in gas-retorts owing to the action of intense 

 heat on the hydrocarbons of the gas. 



In another place Davy says: "I hive shown in the paper 

 referred to in the introduction, that the light of common flames 

 depends almost entirely upon the deposition, ignition, and 

 combustion of solid charcoal, but to produce this deposition 

 from gaseous substances demands a high teinperature." 



Tnis explanation of carbon separation in flames seems per- 

 fectly adequate and free from oi jection. There is, as we have 

 seen, surrounding all ordinary hydrocarbon flames a shell of 

 almost non-luminous com 'Ustion. The gas which passes up- 

 wards within this shell must be highly heated, and in the 

 absence of air will be decomposed so as to deposit solid carbon. 

 This carbon is intensely heated, and glows, and as it reaches 

 the air will burn to form carbon dioxide. The fact that the 

 upper parts of flame are the most luminous in itself indicates 

 that the more we roast the gas the more do we separate the 

 carbon ; and there are other proofs, which I cannot; stop to 

 explain. 



We have now got pretty well to the end of the explanation 

 of the structure of ordinary luminous flames, and I will show 

 you an experiment which epitomises the explana'.ions that 

 have been given. 



We turn once more to the cone-separating apparatus, and 

 use as fuel a substance particularly rich in carbon. This sub- 

 stance, benzene, is a liquid, so I shall have to vaporise it by 

 means of a current of air. When I apply a light to this current 

 of air strongly impregnated with benzene, we get, as you see, 

 a very bright fia ne. This flame exhibits the usual structure. 

 Thi^ is one extreme. Now I will reduce the amount of ben- 

 zene vapnir very rapidly wiihout altering the air, and we shall 

 get the otherextreme, that is, a scarcely luminous flame consist- 

 ing of one single cone. The whole of the combustion is now 

 taking place in a single cone of flame. If I still further reduce 

 the benzene, this flame enlarges slightly and becomes paler. 

 There is now excess of air. A little less benzene still, and 

 you see the flame rises from its perch and disappears ; we have 

 got past the limits within which combustion is possible. Let us 

 next move in the other direction, and gradually increase the 

 supply of benzene to the single cone. It becomes smaller and 

 brighter as we proceed up to a certain point. At leng'h we have 

 evidently got more benzene than there is air to burn, and now 

 appears the second cone at the top of the tube. By sliding 

 the tubes we can unite the flame and make a Bansen flame. 

 Separating the cones again, let us ad 1 still more benzene. The 

 result is very remarkable. The two cones remain intact, but 

 stretching between them are thin luminous streaks of glowing 

 carbon. The excess of benzene u being decomji )sed by the heat, 

 so that the carbon separates nnd glows. The more benzene I add 

 the broaier do these streaks become, until even ually the inner 

 cone ascends, the luminous streaks coalesce, and we have the 

 ordinary luminous hydrocarhon flame. 



I have now put before. you the considerations and methods 

 which will serve, I believe, for the elucidation of all problems of 

 flame strticturc. I am not aware, at any rate, of any flame 

 which does not accord with the general principles which I have 

 explained to you. 



There are many other flame problems besides that which 

 rel ite to mere structure. Of these one of the most interesting 

 concerns the colouration of flame. I will refer to it for a moment 



NO. 1 256 VOL. 49] 



