274 



NA TURE 



\JJan. 22, i 885 



shown that a substance capable of emitting light of all 

 wave-lengths may be capable of reflecting at the same 

 time light of any wave-length, yet I have not been able 

 to show whether or not a substance emitting light of one 

 definite wave-length may not be able to reflect light of 

 that same wave-length, though I have proved that it can 

 reflect no other. For instance, the light given by sodium 

 is absent from that of the sun, so that my experiment 

 proves nothing with regard to it ; yet that particular 

 light is not transmitted through hot sodium vapour, but 

 is stopped by it. One would think it must either be re- 

 flected or its energy must be used up in some way on the 

 vapour itself. I have been unable to get access to the 

 electric light, and no other light I know is strong enough 

 for this experiment. I have wished also to try whether 

 sodium burnt under pressure, or at a very high tempera- 

 ture, would or would not have the power of reflecting 

 light ; but in this direction I am again stopped by lack 

 of apparatus. 



Experiment 8. — The spectrum of the light transmitted 

 through the lac solution is complementary to that reflected 

 by it, i.e. the reflected light is bluish, and the transmitted 

 yellowish-brown ; in the latter case the spectrum is 

 weakened towards the violet, and in the former towards 

 the red. I desired to see if this was so with flame. I 

 arranged a metallic screen with a slit one-fourth of an 

 inch long and one-twentieth wide, close to the candle, so 

 that all light falling upon the spectroscope must first have 

 passed through the luminous portion of the flame, and 

 then with a mirror directed the sunlight into the instru- 

 ment. Pure sunlight was thrown into the upper half of 

 the field for comparison, by means of the reflecting prism. 

 Having adjusted the light so that no difference could be 

 detected between the upper and lower halves of the field 

 of view, the candle was placed in position in front of the 

 slit. There was a very definite general absorption, most 

 noticeable in those rays that are deficient in lamplight, 

 especially about F and G, where also the spectrum of the 

 refle:ted sunlight is brightest. The experiment is difficult 

 owing to the necessity 'of reducing the brilliancy of the 

 sunlight without so far reducing the angle of the illuminat- 

 ing ray that the hot air-currents may vitiate the result. 

 But after many trials I satisfied myself that the more 

 refrangible rays of light transmitted through a luminous 

 flame are to some extent absorbed, the effect being 

 stronger in proportion as the smoky part of the flame is 

 approached. 



Experiment 9. — It seemed evident that the reflection 

 of the sunlight from the flame was due to its superior in- 

 tensity ; I therefore judged that, if I could lower the 

 temperature of the carbon somewhat, I might get a 

 visible reflection with light from other sources. I held 

 an iron nail in the flame, and focused on the resulting 

 smoke the light from a petroleum lamp. The spot of 

 light was plainly visible, only not of a bluish white as 

 with sunlight, but of a dirty yellow colour. It could be 

 seen not only on the cold smoke, but also where it was 

 of a bright cherry red ; beyond that it became lost against 

 the brightness of the incandescence. But the smoke, 

 whether hot or cold, polarised the light exactly as the 

 fine precipitates did. 



Experiment 10. — In order to get rid of the disturbing 

 effects of the light from the candle itself, I punched a 

 hole in the middle of a tin plate, and placed it over the 

 candle. The column of smoke coming up through the 

 hole completely polarised the light thrown on it, whether 

 from a lamp or from the sun, at right angles to the line of 

 incidence. I then placed a little tuft of asbestos satu- 

 rated with melted paraffin upon the hot plate. It gave 

 off a dense smoke, indistinguishable to the eye from that 

 of the burning candle. On applying the spectroscope, 

 however, the difference was manifest. The light reflected 

 by it was not polarised. I would therefore suggest that 

 this polarisation test be the distinction between "steam," 



however dense, and a true " smoke." I have reason to 

 believe that a polarising smoke only arises where the heat 

 causes decomposition. 



Experiment n.— I placed the under side of the tin 

 plate in the light, and found that the soot upon it re- 

 flected plane-polarised light in all directions at right 

 angles to the line of incidence. 



I now desired to ascertain if this power of reflecting 

 light is confined to substances burning in the inner flame. 

 It is difficult to make accurate observations as to the 

 spectrum of the inner flame with an ordinary Bunsen 

 burner, from the fact that it is completely surrounded by 

 the outer flame ; and this last, being but feebly luminous, 

 gives only a very faint spectrum. I wished to make an 

 arrangement by which the spectra of the two flames 

 could be completely separated, while at the same time 

 their intensity should be increased. Accordingly, I made 

 a Bunsen burner with a rectilinear aperture, two inches 

 long by an eighth of an inch wide, in place of the usual 

 round tube. This gave me a broad flat flame, the edges 

 of which I allowed to play each against a piece of well- 

 annealed glass, so that I could look through the glass 

 and see the flame edgeways. In this way I got a very 

 strong spectrum of both the inner and the outer flames, 

 perfectly distinct from each other, the ends of the flame 

 being cut off by playing against the glass. The inner 

 flame with its bright lines was thus completely separated 

 from the outer with its soft, apparently continuous, spec- 

 trum : under sufficient pressure, the separation extended 

 to the eighth of an inch or more. I could see no lines 

 across this intervening space, except perhaps that in the 

 violet : as to which I am not quite sure. Of the others 

 I am certain, and I think the space is perfectly dark. As 

 the glasses soon crack, I substituted another arrange- 

 ment, which I hope still farther to perfect. In this flame 

 I burnt a number of substances, keeping the image of 

 the sun upon it all the while, and having the spectroscope 

 with polarising prism, &c, arranged as in Experiment 5. 

 1 here give the results of two of the most interesting of 

 these experiments. 



Experiment 12. — I burnt on a piece of wire a mixture 

 of copper sulphate and ammonium chloride. This com- 

 pound, as is well known, gives a very beautiful and com- 

 plex spectrum. When the mixture is held in the inner 

 flame it turns dark, bubbles up, and burns like a piece of 

 pitch, giving a continuous spectrum ; and upon this 

 flame, which never passes beyond the inner flame, the 

 reflection of the sunlight may be seen and the Frauen- 

 hofer lines distinguished. There is also, at the same 

 time, in addition to the beautiful blue-violet coloration 

 of the outer flame, a curious " red smoke " right on the 

 outer edge of it. But though in a dark room this looks 

 far more like a solid precipitate, or true smoke, than the 

 bright flame— though by daylight it looks so "smoky" 

 that I thought it surely must give what I sought, a reflec- 

 tion in the outer flame — yet the sunlight passes through 

 it without the slightest effect, save that it renders it in- 

 visible. The spectrum of this apparent smoke consists 

 of groups of lines in the red. 1 



Experiment 13. — I now sought a substance that should 

 be volatile in the inner flame and give anon-volatile oxide 

 in the outer. I placed some zinc, which I found to be 

 the most manageable metal for this purpose, in a small 

 iron cup in the very centre of the flame. As soon as it 

 boiled, flashes of white light appeared in the outer flame, 

 and I was enabled to ascertain that these flashes gave a 

 continuous spectrum and were also capable of reflecting 

 sunlight, the reflected light being polarised, as in the 

 other cases, in all directions at right angles to the line of 

 incidence. 



[n a recent experiment this "red smoke" gave a " soft " 

 l from the extreme red to the yellow a little beyond D. 

 , and seems to be produced when the fused mass is dra 



