88 



NATURE 



[November 23, 1893 



discovered that flames have separable regions of combustion, 

 and having armed ourselves with an appliance for dissecting the 

 flame, we may proceed to discuss the main question. 



I do not intend this evening to enter seriously into chemical 

 details, but there are one or two simple points to which I must 

 draw your attention. Flame, we see, is a region in which 

 chemical changes are taking'place with the evolution of light. 

 It is to be expected, therefore, that the character of a flame, 

 its structure and appearance, will vary according to the chemical 

 changes that give it birth ; and we should naturally anticipate 

 that the more complex the chemical changes the more complex 

 would be the flame. The kind of complexity to which I refer 

 is illustrated by the diagram. 



Products 



Name 



Composition 



Partial 

 Combuation 



Hydrogen ] waier 



Carbon monoxide ' carbon and oxygen \ carbon dioxide 

 Carbon ' 1 carbo monoxide 



Cyanogen carbon and nitrogen ! carban monoxide 



and nitrogen 

 Hydrogen sulphide ] hydrogen & sulphur (?) 



Hydrocarbons 



hydrogen & carban 



carbon m-^noxide 

 carbon dioxide 

 hydrogen & water 



Complete 

 Combustion 

 water 



carbon dioxide 

 carbon dioxide 

 carb m dioxide 

 and nitrogen 

 water and sul- 

 phur dioxide 

 carbon dioxide 

 and water 



In the first column are the names of five combustibles ; 

 their chemical composition is stated in the second column. 

 All these substances in burning combine with the oxygen 

 of the air. The case of hydrogen is the simplest. This 

 gas, when it burns, unites with half its volume of oxygen, 

 and forms steam. The process is incapable of any complica- 

 tion. We might predict, therefore, a very simple structure for 

 a hydrogen flame. The same is true for the next gas carbon 

 monoxide, which, although a compound, unites at once with its 

 full supply of oxygen and burns, forming carbon dioxide. The 

 third combustible, carbon, presents a new feature ; in burning 

 it can combine with oxygen in two stages, forming in the first 

 instance carbon monoxide, which, as we have just seen, can 

 itself combine with more oxygen to form carbon dioxide. We 

 cannot vaporise carbon and use it as a gas, so that we shall not 

 actually deal with this example. But the next combustible on 

 the list, cyanogen, will serve almost as well, for it is a compound 

 of carbon with nitrogen, and nitrogen is, under ordinary cir- 

 cumstances, practically incombustible. To use cyanogen is 

 thus much the same as to use carbon vapour. We may expect 

 some complexity in the cyanogen flame in consequence of the 

 fact that carbon can burn in two steps. The next combustible, 

 hydrogen sulphide, presents a further degree of complexity. It 

 is composed of two elements, each of which is coml^ustible on 

 its own account. Lastly, we come to the great class of hydro- 

 carbons, which includes all ordinary combustibles, oil, tallow, 

 wax, petroleum, and coal-gas. The carbon and hydrogen are 

 both separately combustible elements, and one of them — carbon 

 — is, as we have seen, combustible in two steps. 



We will now consider the problem in its simplest aspect. For 

 this purpose I choose the gas carbon monoxide. I should choose 

 hydrogen were it not for the fact that its flame is almost in- 

 visible. We will allow a stream of carbon monoxide to issue 

 from the circular orifice of this glass tube. Lighting the gas we 

 get a blue flame. On examining this flame closely we perceive 

 that it is simply a hollow conical sheath of pretty uniform charac- 

 ter. I need scarcely demonstrate that it is hollow, but I may 

 do so in a moment by using Prof. Thorpe's simple device of 

 thrusting a match-head into the centre of the flame — a pin 

 passii/g through the stick of the match, and its ends resting on 

 the tube. The match-head is now thrust well up inside the 

 flame, and you observe that it remains there sufficiently long 

 without burning, to make it quite clear that there is no combus- 

 tion within the cone. The conical form of the flame is easily 

 explained. As the stream of gas issues from the tube the out- 

 side portions become mixed with the air and burn. The inner 

 layers must successively travel further upwards, like the succes- 

 sive tubes of a tele-cope, before they can get enough air to burn, 

 and in this way we a-tive at the conical form. 



There still remains one thing to account for, and that is the 

 luminosity and culour of the flame. The questions here involved 

 are p'jrhaps the most interesting of all, but they are complicated, 

 and I will not say more than a few words about them. The 

 most (ibvious answer to the question, " Why is the flame 

 (uminous?"is to say that the heat developed during the chemical 



combination raises the product of combustion to a temperature 

 at which it glows — a " blue heat " in the present case. Now if 

 we put a thermometric instrument into the carbonic oxide flame, 

 it does not register at any point as high a temperature as 

 1500° C, but if we take carbon dioxide and heat it in a tube by 

 external heating to I500°C. we get i^o signs of luminosity what- 

 ever. On these grounds several eminent investigators have been 

 led to abandon the simple explanation above given, and to say 

 that the luminosity of a carbon monoxide flame must depend not ■ 

 on the heat of chemical combination, but on something in the 

 nature of electrical discharges between the combining substances, 

 which discharges produce the disturbances of the ether percep- 

 tible as light. This view seems to be fraught with a fundamental 

 error. The temperature registered by any instrument introduced 

 into a flame is an average temperature, uncorrected for losses by 

 conduction. It is not the temperature of the newly-formed gas, 

 but of the mixture of that and the unliurned gases. If we had 

 a very small instrument which we could apply to the particles of 

 newly-formed gas, we should undoubtedly find them at a very 

 much higher temperature than any indicated by the ordinary 

 thermometric apparatus, and it is not unlikely that the tempera- 

 ture would be several thousand degrees, approximating indeed to 

 the temperature at which we arrive by calculation from the heat 

 of combustion of the gas and the heat capacity of the product. 

 We cannot say that the flame is luminous from some other cause 

 than simple hotness, for we have no means of seeing whether 

 carbonic acid glows when raised by external heating to a tem- 

 perature of several thousand degrees. 



At the same time one cannot help remarking on the similarity 

 between such a flame as that of carbon monoxide and the ap- 

 pearance presented by an attenuated gas when submitted to the 



NO. 1256, VOL. 49] 



Fig. 2. — Typical Flames, (ir) Carbon monox'de, single coned ; {h) Cyanogen, 

 two coned ; (<) Small coal-gas flame. 



electrical discharge in a Geissler tube. I have here such a tube, 

 containing carbon dioxide, and I have placed a mask over it, so 

 that we see a long triangular piece of it. When I pass the 

 discharge you see it lights up and presents an appearance 

 strikingly like that of our conical flame of carbon monoxide. 

 There may be a close relationship between the phenomena, but 

 we cannot affirm it yet. No doubt we shall soon learn a good 

 deal more about both phenomena. 



We have now done with the simplest kind of flame. We see 

 that it consists of a single conical sheath of combustion, at 

 every point of which the same chemical change is taking place, 

 and every point of which in consequence has the same appear- 

 ance. 



We pass to the cyanogen flame. This flame is one 

 of remarkable beauty ; it consists, as you see, of two distinct 

 parts : one a rose or peach-blossom coloured cone, surrounded 

 by a paler cone, which is bright blue where it is near the inner 

 cone, and shading off to a kind of greenish grey. What 

 is the cause of this double structure? It might be that 

 part of the gas is burning round the orifice, the rest 

 further out in the second cone ; but a similarity of the 

 chemical processes in the two parts of the flame is here ren- 

 dered improbable by the difference in colour. The only satis- 

 factoiy way of answering the question is to separate the cones, 

 and analyse the gases in the intervening space. This we car> 

 easily do in the cone-separating apparatus. 



