5o6 



NATURE 



\Oct. 17, 1872 



foCf of Anaxagoras, tlie ui'ax'"? of Demokritos, the actuating 

 form of fire of' llerakleilos, the moving etiier of Aristotle. Tlie 

 links whicli bind togetlier ancient and modern pliysical thought 

 are strong and enduring, and since tliey have lasteil during tlie 

 rise and fall of many nations, and during the most prolound 

 changes in the mode and tone of thought, it is not unlikely that 

 they will endure as long as the chain itself. 



G. F. 'RODWELL 



THE DIA THERM AC Y OF FLAME 

 T HAVE just read Mr. Ericsson's paper on "The Sun's 

 •^ Radiant Heat " in Nature, October 3, p. 458, and find that 

 he has made some experiments on the diathermacy of flame 

 closely resembling those which I made in 1S69, and described in 

 chap. viii. of "Ihe Fuel of the Sun" published January 1S70. 

 Although the object of our investigations was identical and the 

 method of proceeding very similar, the results obtained are so 

 contradictory that one of us must be quite wrong, and therefore 

 I think the subject demands discussion. 



Referring the reader to the engraving illustrating Mr. Eric- 

 sson's paper, I may easily describe my apparatus. Like Mr. 

 Ericsson's, there was a gas-pipe from the side of which projected 

 a row of burners, each provided with a separate stop-cock. My 

 burners, however, differed from his in being perpendicular to the 

 main pipe which was always used in a horizontal position. My 

 blackened bulb thermometer was similarly fixed at one end of a 

 chamber or vessel, the other end of whicli was open to receive 

 the radiations from the lUmes. This, however, was much simpler 

 than Mr. Ericsson's. It had not the double chamber with inter- 

 vening wall, nor was it surrounded by water, but was simply a 

 thin tube of tin plate polished inside to prevent absorption of 

 radiant heat. The thermometer was insulated from metallic 

 contact with this tube, and thus could only receive heat from it 

 by radiation, which the polishing reduced to a minimum. Tlie 

 sectional form and opening of tlie tube was made to correspond 

 nearly with that of the presented side of the gas fiames, but w-as 

 somewhat larger. 



At first I used Bunsen-burner flames, then flat flames like 

 t'nose figured by Mr. Ericsson, afterwards simple jets formed 

 by the gas issuing from a small pin-hole, the jets being far 

 enough apart to be quite independent ; finally a row of such jets 

 so near to each other that they came in contact, coalesced fully, 

 and formed one sheet of flame, the edge of which was pre- 

 sented to the mouth of the polished tube containing the ther- 

 mometer . 



Guided by results obtained in a previous series of photo- 

 metric experiments on the transparency of flames to their own 

 special radiations, and by the first experiments I made on 

 diathermacy, I relied on the arrangement last described, viz. the 

 coalescing jets. The reason for this will presently appear. 



My mode of proceeding differed in another respect from Mr. 

 Ericsson's. Instead of lighting one jet at one end and then another 

 and another in succession towards the thermometer, I always 

 worked with an odd : number of flames, and began with the 

 middle jet, then lighted one on each side, next one on each 

 side of those three, then one on each side of those five, and so 

 on. My flames were thus maintained at a constant mean distance 

 from the thermometer. 



By means of a well-constructed experimental gas meter, with 

 micrometic regulator, ami a minute alarum clock, the supply of 

 gas was accurately adjusted, so that each additional jet, or pair 

 of jets, should consume an exactly equal differential increase of 

 gas. The results obtained were as follows : — 



Here, then, is a serious discrepancy. I get an increase of 4" 

 by the first addition of two flames, and by eight of such additional 

 pairs obtain an increase of 34°, instead of the 32° due to theoretical 

 diathermacy. These 2° of excess (being due to the latter end of 

 the series) I attributed to the increased temperature of my 

 apparatus. 



Mr. Ericsson obtained an increase of only 70-9 instead of I7°'6, 

 the theoretical requirement. 



Without any disposition to underrate the value and importance 

 of Mr. Ericsson's researches, I think that in this matter he has 

 been deceived by overlooking some important sources of fallacy. 



I. He tells us nothing about the quantity of gas consumed. 

 His jets all issue from the same main pipe, which he describes 

 as supplied with "gas at ordinary pressure." Now with such a 

 supply the quantity of gas burning from each jet would steadily 

 diminish as he turned on the .additional jets. On turning the 

 secondjet the first would diminish; when the third was turned 

 the supply to both first and second would be reduced ; and so on, 

 to an extent depending upon rates of sectional area of the supply 

 pipe to that of the jet holes. If Mr. Ericsson's drawing is 

 made to scale, the error due to this was of great magnitude. 



A second source of error is described in Mr. Ericsson's own 

 words ; he says, " It will be observed that the prolongation of 

 the axis of tiie conical vessel upwards passes througli the central 

 portion of the fl.iiiies at the point of maximum thickness and in- 

 tensity." Now the point of maximum thickness of a flame is 

 just that part which is hollow, and consists of a central core of 

 unburnt gas with an outer coating of true flame, and the central 

 portion of such a flat flame as Sir. Ericsson represents includes 

 much of the blue portion of the flame, consisting of hydro- 

 carbon not yet in full combustion. Mr. Ericsson, therefore, was 

 not experimenting upon the diathermacy of ten flames, but upon 

 the diathermacy of ten discs consisting of a mixture of flame 

 proper and uubiirnt hydrocarbon. Now Tyndall has demon- 

 strated the remarkably high resisting or absorbing power of such 

 hydrocarbon in reference to the radiations from a flame produced 

 by hydrocarbon combustion. The flame itself might therefore 

 be perfectly diatherraous, and yet, when examined in this man- 

 ner, exhibit a considerable degree of athermacy. 



There is still a third source of error in .VIr. Ericsson's mode of 

 proceeding, the magnitude of which I am not yet able to estimate, 

 though some experiments made since publishing my fir»t results 

 lead me to suspect that it is sufficiently important to demand very 

 careful elimination. I allude to the arrangement of a series of 

 separated flat flames, with the broad surfaces presented to the 

 thermometer. 



What must we have between each of these separated flat 

 flames ? Obviously each flame is coated with a film of vapour, 

 the product of the combustion of those portions of llame lying 

 below it ; these vapours, though rapidly rising, must form a layer 

 of sensible thickness equal to an important fraction of the whole 

 thickness of such thin flames. When operating wilh the whole 

 eleven flames, there were twenty-one such films between the first 

 flame and the thermometer. Now, we know from the experi- 

 men:s of Tyndall, that a large proportion of the rays of heat 

 emitted from a hydrocarbon flame will be absorbed by such in- 

 tervening strata of aqueous vapour, carbonic acid, and carbonic 

 oxide. It is true that the middle or blue part of the flame, 

 having less combustion going on below it, must h.ive a thinner 

 coating of such vapours than the upper part; and thus in Mr. 

 Ericsson's arrangement this third source of error is diminished 

 in the same proportion as the second is increased. It was these 

 theoretical considerations, confirmed by resuhs of preliminary 

 experiments, that induced me to abandon the flat flames in favour 

 of the simple round jets, and finally to adopt the continuous 

 flame formed by the coalescent jets. 



As I stated on the first publication of the results of these ex- 

 periments on the diathermacy of tlame, I do not regard them as 

 suflicienUy delicate to be finally and quantitatively conclusive ; 

 the means at my disposal rendered them less satisfactory than 

 those I made on the transparency of flame. Still, I think they 

 are not open to any such serious sources of error as those I have 

 here pointed out. 



I hope that Mr. Ericsson will not be offended by the candour 

 of my criticism, nor by the egotism which is inevitable in an 

 unaffected defence of one's own philosoiihical bairns. 



My experiments, like those of Mr. Ericsson, were made with 

 the direct object of throwing some light upon the great mystery 

 of solar radiation; and the fact that we have arrived at such 

 pposing conclusions will, I hope, lead^^to further investigation, 



