November 23. 1893] 



NA rURE 



89 



I now form the flame at the top of our cone separating 

 apparatus, and supply a certain amount cf air along with the 

 cyanosjen. Vou observe the rose-coloured cone contracts 

 somewhat. The gas burning there now gets its air supply ea-ily, 

 and has not to wander outwards. If I still further increase the 

 air supply, and make the ascending mixture explosive, you see 

 the inner cone begins to descend into the tube, and passes down 

 until its progress is checked at the narrow tube, where the up- 

 rush of gas is more rapid. We have now got the cyanogen 

 flame dis-ected, and by taking out a sample of the gases from 

 iliis interconal space and analysing it, we shall find what 

 chemical change has taken place in the inner rose-coloured cone. 

 The analysis shows that what takes place is the combustion of 

 the carbon of the cyanogen to form carbon monoxide almost 

 exclusively ; the carbon monoxide then ascends, and when it 

 meets with more oxygen in the outer air, burns in a second cone 

 to form carbon dioxide. 



Reverting then to the flame of the pure unmixed gas burning 

 at the top of a tube, we see that the gas and air will interpene- 

 trate. When there is just enough oxygen to burn the gas to 

 carbon monoxide, we get the rose-coloured cone, and outside it, 

 where this carbon monoxide gets more air, we have a second 

 cone. The two-coned structure corresponds then to two chemical 

 stages of combustion. 



Nov/ we might go further and anticipate that if we supplied a 

 very large quantity of air to the cyanogen, as in a blowpipe, the 

 two-coned structure would disappear, for the carbon should 

 be burnt up at once to the ultimate product, carbon dioxide. 

 We can easily try this. I will separate the two cines again in 

 our apparatus, and increase the air supply s'ill further. When I 

 do 50 you observe that the second cone gradually fades away, 

 and now the whole of the combustion is taking place at the end 

 of the inner tube. Though this is so, the flame is not quite a 

 simple cone. It is, as you see, surrounded by a greenish halo. 

 This hal > is due, I believe, as Prof. Dixon has suggested, to the 

 fact that the nitrogen of the cyanogen is not, strictly speaking, 

 incombustible. This has been proved by Mr. Crookes in his 

 beautiful air flame, and besides, the greenish iialo is frequently 

 noticeable in cases of combustion where oxides of nitrogen are 

 ]")re=;ent. 



Keeping to our list we ought next to deal with ths combustible 

 hydrogen sulphide or sulphuretted hydrogen. This gas, you 

 remember, is composed of two separately combustible elements, 

 each burning in one stage. The flame is, as you might expect, 

 two-coned, buf I will not dwell upon this case — partly because 

 it is not yet fully worked out, and partly because any prolonged 

 experimenting with this gas would, I feel sure, be resented even 

 by the most indulgent audience. 



I am obliged, therefore, to pass to compounds of carbon and 

 hydrogen, in which there are not only two combustible ele- 

 ments, but one of them, as we have seen, combustible in two 

 chemical stages. Here we have an almost unlimited choice of 

 materials, for we come amongst the combustibles which ordi- 

 narily supply us with light. I shall, for the sake of convenience, 

 use coal-gas. This is really a very complex combustible, con- 

 sisting one half of hydrogen, the other half of at least a dozen 

 different compounds of carbon and hydrogen. But experience 

 has shown that the chemical phenomena attending its com- 

 I bustion are quite of the same character as those to be observed 

 1 with a single compound of hydrogen and carbon. 



It will, I imagine, be scarcely necessary for me to point out 

 the various parts which are to be seen in the flame of a candle or 

 of coal-gas. There is on the diagram (Fig. 2, c) before you the pic- 

 ture of a somewhat small coal gas flame, produced at a circular 

 orifice. It is, of course, enormously enlarged in the diagram. 

 jFour distinct parts are to be recognised. First, the central and 

 I darkest part ; this contains the unburnt gas, just as we saw in 

 the case of the carbon monoxide flame. Perhaps it is wrong to 

 (speak of this at all as part of the flame, for it is really a region 

 iof no flame. At the base of the flame are two blue strips em- 

 ibracing the lower portion of the flame. This appearance you 

 will understand results from the mode in which we view the 

 Jilame. The strips are really due to a sheath which goes right 

 jround the flame like an uninterrupted calyx. It looks bright 

 (where we view it edgewise. When we look through, as in the 

 jmiddle of the diagram, it is very pale indeed. Next we have to 

 inotice the bright yellow patch, so bright in the reality as to 

 nia>k the other parts. Though it looks bright and dense, it is 

 merely a hollow sheath. Lastly, there is surrounding the 

 whole flame a pale mantle of flame of very slight luminosity. 



and of an almost indescribable tint, which perhaps we may call 

 lilac. These parts are discernible in all ordinary name--. They d > 

 not always occupy the same relative space. In the flame given 

 by a good gas-burner the yellow part is made by intention as 

 large as possible ; in the flame of a piece of string or a spirit - 

 lamp you will see the outer investing mantle very ditinctly 

 developed. 



If we are to understand flame, then, we must find an intel- 

 ligible explanation of the existence of these distinct pans of its 

 anatomy. One important point we can settle at once. An ordinary 

 flime owe; its differeniiated structure to the slowness with 

 which it gets the oxygen necessary for combustion. If there is 

 an immediate and sufficient supply of air, the characteristic 

 structure disappears. This we can secure bv making the stream 

 of gas sufficiently rapid. I have here a steel cylinder containing 

 coal-gas at very high pressure. If I allow the gas to escape 

 slowly, we get a flame in which we should find the ordinary 

 parts. But if now I allow the gas to is>ue rapidly, the admix- 

 ture with air is so rapid, and, as you see, we have a pale flame 

 quite undifferentiated in structure. We reach the same result 

 by introducing a strong current of air into an ordinary flame, 

 I as in the blast blow-pipe. The flame, you see, is then homo- 

 geneous, as in the previous case. 



We see then that the structure of an ordinary gas flame is 

 largely dependent upon the slowness with which the gas gets 

 the air necessary for combustion. There is s'ill one other 

 evidence of this. It is obvious that a very small flame will have 

 a much better chance of getting its oxygen quickly than a larger 

 flame. It is, I am sure, within everyone's knowledge 

 that a very tiny gas flame is blue, and, as a matter of fact, we 

 can learn a great deal about flame structure by carefully watch- 

 ing the development of a very small flame. I am going to show 

 you on the screen a series of photographs of actual flames. The 

 photographs hjave been tinted as faithfully as pissible. 



The first slide (Fig. 3, a) shows a liny gas flame burning at the 

 end of a glass tube ; it consists of a bright blue cone surrounded 

 by a fainter one. Both are quite continuous. By puitinsj in 

 another slide, and using the " dissolving view" arrangement of the 

 lantern, I will show you the effect of turning on the gas. The flame 

 (Fig. 3, b) you see is larger, and now is observed a third region 

 in the flame— namely, a patch of bright yelliw at the tip. The 

 original cones are still there, but are slightly interrupted at their 

 apices. Turning on more gas, the flame (Fig. 3, f) again enlarges, 

 the yellow patch increases in size, and the original cones are further 

 broken into. But you see the yello.v jiatch is indented at points 

 corresponding to the inner cone, which, as it seems, is striving 

 to maintain its integrity. Turning on still more gas, we have 

 now a great preponderance of yellow, the original blue cone 

 is reduced to mere vestiges, and the outer cone forms a faint 

 surrounding to the whole flame (Fig. 3, d). This is flame as we 

 ordinarily know it. I wish now to show you another series of 

 changes. We must suppose the gas supply fixed, and the photo- 

 graphs I will show represent the changes which take place in 

 the flame when air is gradually added beforehand to the coal- 

 gas. The supply of coal-gas is, I repeat, the same in all cases. 

 The first change seen is, you will notice, that the yellow 

 patch diminishes in extent (Fig. 3, f). If I add more air 

 it diminishes still more, and the inner cone is growing 

 in distinctness (Fig. 3, /). If I add a trifle more air, 

 the yellow patch disappears altogether, and we have now 

 complete and distinct inner and outer cones (Fig. 3, g). I think 

 you will admit that these two sets of photographs show a close 

 correspondence, and you can see it more plainly if I throw them 

 on to the screen in a group. There is really nothing sarprising 

 in this similarity. The smallest gas flame has obviously the 

 best chance of getting air, and when it gets enough it burns 

 in a two-coned flame. The same effect is reached by adding 

 air to the gas before it is burned. If we have a larger gas flame 

 it has, of course, less chance of getting its oxygen rapidly, and 

 we see that in whatever way we starve the flame of oxygen, we 

 lose the simple structure, and come upon the yellow patch. 



Now, when we come to inquire into the chemical changes 

 occurring in such a flame, we may, I think, feel confident that 

 the chemical actions which determine the existence of the blue 

 cone and the outer cone are the same, whether these cones are 

 complete, as in a small flame, or fragmentary, as in a larger one. 



If that is so we can soon make progress, for, as I have shown 

 you, we can easily separate these cones and find what is going 

 on in each. I again use the cone-separating apparatus. First 

 we have an ordinary luminous gas flame at the top of the outer 



NO, 1256, VOL. 49] 



