JULY 8, 1920] 



NATURE 



585 



scattered light is blue. The transmitted beam is 

 robbed of its bluer constituents, and tends to 

 become yellower, as you may see on the screen. 



The light scattered laterally is to be compared 

 to the blue sky; the yellow transmitted light to 

 the direct light of the setting sun when it has 

 traversed a great thickness of air. 



As the precipitation goes on, the transmitted 

 light becomes orange, and even red. But the 

 particles of sulphur eventually get bigger, and 

 then give a less pure blue in the lateral direction. 

 We shall have more than enough to occupy us if 

 we confine our attention to the earlier stages, 

 when the particles are small compared with the 

 waves of light. 



A very important property of the scattered light 

 is its polarisation. The vibrations of the scattered 

 light as you have seen it, viewed laterally in the 

 horizontal plane, are almost wholly up and down. 

 No light is emitted which vibrates in the horizontal 

 plane. It is easy for individual observers to verify 

 this with a Nicol's prism held to the eye, but 

 this direct method unfortunately does not lend 

 itself to public demonstration. 



We may, however, use polarised light to begin 

 with, and you can then observe that if the polar- 

 ising Nicol is set so as to transmit up and down 

 vibrations, these are abundantly scattered towards 

 you by the small particles. As I turn the polar- 

 ising Nicol through, a right angle, you will see 

 that the light scattered towards you is extin- 

 guished. 



The polarisation of light scattered by the 

 sulphur particles is one of the most conclusive 

 reasons for considering it to be an analogue of 

 the blue light of the sky, for the latter shows a 

 polarisation of exactly the same kind when exam- 

 ined at right angles to the sun. 



A cloud of small particles of any kind is capable 

 of producing these effects, the essential condition 

 being that the individual particles should be of 

 small dimensions compared with the wave-length 

 of light, so that at a given moment the vibration 

 at a given particle may be regarded as having a 

 definite phase. In this case it was shown by my 

 father that the shorter (blue) waves are of neces- 

 sity more scattered than the longer ones (red) ; 

 thus the scattered light is bluer than the original. 

 This conclusion can be justified in detail whether 

 we adopt the elastic solid theory, or the electro- 

 magnetic theory of the nature of light, but it is 

 also deducible from the general theory of dimen- 

 sions, without erttering upon any details of the 

 nature of light beyond its characterisation by the 

 wave-length. 



An alternative theory which still sometimes 

 shows its head attributes the colour of the sky to 

 a blueness of the air, regarded as an absorptive 

 medium. Such blueness is referred to the presence 

 of ozone, and appeal is made to the undoubted 

 fact that a sufficiently thick layer of ozone shows 

 a blue colour by absorption. This theory gives 

 no account of why the sky light is polarised, or 

 indeed of why there is any light in the clear sky 

 at all. Further, its fundamental postulate that the 

 NO. 2645, VOL. 105] 



air is blue by transmission is contrary to observa- 

 tion. The setting sun is seen through a greater 

 thickness of air than the midday sun. According 

 to the theory under discussion, the setting sun 

 ought to be the bluer of the two, which everyone 

 knows it is not. No doubt the presence of ozone 

 tends to make the air blue by transmission. But 

 this effect is more than compensated by the lateral 

 leakage (scattering) of blue light from the beam, 

 which makes the transmitted light yellow. 



Dusty Air and Pure Air. 



If it be conceded that the blue sky is due to 

 scattering by small particles, we are confronted 

 with the question : Of what nature are these par- 

 ticles? At the time of my father's early investiga- 

 tions (1871) this was left open, though they were 

 regarded as extraneous to the air itself. In 1899 

 he returned to the subject, and considered the 

 matter from the point of view of what was lost by 

 the original beam by lateral leakage (scattering), 

 which simulates the effect of absorption. He then 

 found that the air itself, regarded as an assem- 

 blage of small particles (molecules of oxygen and 

 nitrogen), would have an apparent absorbing 

 power not much less than that actually deduced 

 by observations of the sun at different altitudes. 

 The inference was that the air itself was capable 

 of accounting for much, if not all, of the scatter- 

 ing which is observed in the blue sky ; in fact, that 

 the molecules of air are the small particles in 

 question. 



When a beam of sunlight enters a room through 

 a small aperture in the shutter, its course is 

 readily traced by the brightly illuminated motes 

 in the air. Prof. Tyndall, working in this institu- 

 tion, devoted much attention to the nature of these 

 motes, and the methods by which they may be 

 got rid of. His results may be consulted in his 

 fascinating essay on "Floating Matter." One 

 way of getting rid of the motes is to filter the air 

 through cotton-wool. We have here one of 

 Tyndall 's own experimental tubes. The electric 

 beam passes axially along it, and is concentrated 

 to a focus about the middle of its length. Its 

 track is conspicuous. If now we displace the air 

 originally in the tube by filtered air, you see that 

 the cone of light fades into invisibility. 



Another of Tyndall's experiments was merely 

 to place a spirit lamp or Bunsen burner under the 

 beam. Since most of the dust particles are com- 

 bustible, the gases rising from the flame are free 

 from them. As you now see, dark rifts appear in 

 the beam where the uprising stream of dust-free 

 gases traverses it. 



Tyndall. on the strength of these experiments, 

 stated without qualification that dust-free air does 

 not scatter light, but my father's views and theory 

 lead clearly to the conclusion that it does. But 

 when I asked him what he thought about the feasi- 

 bility of detecting it by a laboratory experiment, 

 he was not very sanguine of success. It seemed 

 worth while, however, to make the attempt, and 

 I came to the conclusion that the difficulty was not 



