Dec. 1 6, 1886] 



NATURE 



165 



heat expanded the lower atmosphere, the upper cloud stratum 

 would be lifted, flattened, and broken into patches, the result 

 being a mackerel sky. Should, however, the expansion in the 

 lower atmosphere take place very slowly, it was possible that the 

 cloud, though thinned, would remain unbroken. Rapid motion 

 of the atmosphere would elongate the cloud in the direction of 

 motion ; and, if accompanied by expansion from below, would 

 rupture the cloud into ribs or bars at right angles to the current. 

 If the mass of the cloud were stationary or moving slowly, 

 prominent parts might be drawn out into " mares'-tails. " 



FURTHER EXPERIMENTS ON FLAME 

 T N my former paper, published in Nature, vol. xxxi. p. 272, 1 

 showed that there are two classes of continuous spectra, 

 viz. those due to an incandescent precipitate, in which case the 

 flame has the power of reflecting and polarising light ; and, 

 secondly, flames that possess no reflecting power, but give a soft 

 continuous spectrum without maxima or minima. 



Of this second class is carbonic oxide, which gives, at normal 

 pressures, a fairly bright, and at increased pressure, according to 

 Dr. Frankland, a very bright, continuous spectrum. I have 

 observed its spectrum recently under reduced pressure, using an 

 apparatus similar to that described by Dr. Frankland in his 

 "Experimental Researches," p. %%\et scij. 



I had considerable difficulty at first in keeping the flame 

 alight at anything like low pressures, and finally adopted a glass 

 jet, of a trumpet shape, increasing very gradually from t milli- 

 metre to 3 millimetres in diameter, the flame being farther 

 shielded from draughts by a wide disk of cork 10 millimetres 

 below the mouth of the jet. 



ExpeHmcnt I. — Carbonic oxide was burnt in oxygen. The 

 flame was densest close to the jet, and diminished in brightness 



Flame of carbonic oxide burning in oxygen at 60 mm. pressure, with spectrum 

 showing maxima. The continuous spectrum at the bottom is given by 

 the red-hot top of the glass jet. 



to the tip, without any definite separation into mantles with a 

 space between. At normal pressure every part of it gave a 

 continuous spectrum. 



■- "At about 260 millimetres there began to be a noticeable con- 

 centration of the light in the violet and the green in the position 

 of the principal bands of the carbon spectrum. At 120 milli- 

 metres the concentration was unmistakable, but the spectrum 

 was still continuous. At 60 millimetres it presented the appear- 

 ance shown in the sketch. There appeared to be a second 

 maximum in the green — not, however, at all well defined — but 

 the principal maximum was continued upwards into a faint green 

 cloud corresponding to the very faint tip of the flame ; this 

 cloud was perfectly isolated, but, unlike the carbon bands, was 

 brightest in the middle.' I failed to see a similar cloud over the 

 maximum in the violet, but this might be owing to insufficient 

 light, my pumps being only able to maintain so high a vacuum 

 against a very small flame. Mr. T. Legge, of Trinity, who was 

 with me, observed that the comparative absence of the blue was 

 very remarkable. 



My supply of oxygen becoming exhausted, I had to use air. 

 The flame became less bright, and the maxima less marked. By 

 turning it very low, we brought the gauge down to 40 milli- 

 metres. The flame still burnt steadily. 



Finally, at 60 millimetres pressure, I adjusted the flame to a 

 height of three-quarters of an inch, opened the air-taps, and 

 checked the pumps. The flame increased in brightness and 

 decreased in size to rather more than a quarter of an inch at 

 normal pressure, the spectrum becoming again perfectly con- 

 tinuous. 



^ It is impossible in a woodcut to give a true idea of the extreme faintness 

 of this isolated cloud. It is only visible when the brighter part of the spec- 

 tnim is hidden from the eye, and the room is perfectly dark. 



Experiment 2. — Having the apparatus ready, I repeated Dr. 

 Frankland's experiment of burning coal-gas in air under re- 

 duced pressure. He says that " finally, at 6 inches pressure, the 

 last trace of yellow disappears from the summit of the flame, 

 leaving the latter an almost perfect globe of a peculiar greenish- 

 blue tint." 



He used a jet contracted at the mouth to I -5 millimetres. 

 With my much wider trumpet-shaped jet, by turning on more 

 gas I could produce smoke at 160 millimetres so as to blacken 

 the glass chimney. At 120 millimetres the light was noticeably 

 less vivid, the flame having a diluted appearance, but the spec- 

 trum showed the usual carbon lines much more sharply defined, 

 the mantles being very much thicker than at normal pressure. 

 With this exception there was no difference caused by the reduc- 

 tion of the pressure to 60 millimetres, and even then, on turning 

 up the gas a little, the ellipsoidal flame became pointed, and the 

 yellow light, giving the incandescence spectrum, re-appeared in 

 the tip of it. It is evident that the trumpet-shaped jet allows 

 carbon to be precipitated in the flame at much lower pressures 

 than the contracted jet. In the same way alcohol heated in a 

 bulb tube burns from the mouth of it with a bright and even 

 smoky flame, whereas it burns from a wick with a blue one. 



One phenomenon observed by Dr. Frankland I was dis- 

 appointed not to see. He says : "Just before the disappearance 

 of the yellow portion of the flame there comes into view a 

 splendid halo of pinkish light forming a shell half an inch thick 

 around the blue-green nucleus; . . . the colour of this luminous 

 shell closely resembles that first noticed by Gassiot in the strati- 

 fied electrical discharge passing through a nearly vacuous tube 

 containing a trace of nitrogen." He does not speak of having 

 used the spectroscope to determine the nature of this pink glow. 



I went considerably below the lowest pressure mentioned in 

 his paper, viz. 4"6 inches, but entirely failed to reproduce it. 

 But I have noticed that very small flames from capillary tubes, 

 observed under a power of 100 in the microscope, are sometimes 

 tinged with rose-colour in the outer mantle, from a very faint 

 trace of sodium orange light mingling with the blue of the soft 

 outer mantle ; and I think that the jet he used or the glass 

 chimney may have been sufficiently heated to give a rosy tinge 

 to the iiame. 



One other point I would call attention to. The appear- 

 ance of the gas-flame at low pressures is precisely like 

 that of a very small gas-flame under the microscope. The 

 inner mantle appears to be bordered with bright green 

 light, due to the principal green band of the carbon spectrum 

 extending slightly beyond the others. Beyond this, again, comes 

 a zone of violet light due to the band in the violet, and in most 

 cases this extends nearly, if not quite, to the outer mantle. At 

 ordinary pressures this can only be seen with a magnifying-glass, 

 except with a special burner ; but the in vacuo flame is, as it 

 were, magnified as to its structure, which is thus visible to the 

 naked eye. This fact suggests that flames may in a sense obey 

 Boyle's law, i.e. that the space required for complete combustion 

 under given conditions varies inversely as the pressure. I am 

 continuing my experiments. George J. Burch 



SOCIETIES AND ACADEMIES 

 London 



Royal Society, November i8. — " The Coefficient of Vis- 

 cosity of Air. Appendix." By Herbert Tomlinson, B.A. 

 Communicated by Prof. G. G. Stokes, P.R.S. 



In the previous experiments by the author on this sub'ect, the 

 coefficient of viscosity of air was determined from observations 

 of the logarithmic decrement of amplitude of a torsionally 

 vibrating wire, the lower extremity of which was soldered to the 

 centre of a horizontal bar. From the bar were suspended verti- 

 cally and at equal distances from the wire a pair of cylinders, or 

 a pair of spheres. The distances of the cylinders or spheres 

 from the wire were such that the main part of the loss of energy 

 resulting from the friction of the air may be characterised as 

 being due to \he pushing of the air. 



Acting on a suggestion of Prof. Stokes, the author proceeded 

 to determine the coefficient of viscosity of air by suspending a 

 hollow paper cylinder about 2 feet in length and half a foot in 

 diameter, so that its axis should coincide as to its direction with 

 the axis of rotation. The cylinder was supported by a light 

 hollow horizontal bar, about 7 inches in length, to the centre of 

 which the vertically suspended wire was soldered. The wire 



