384 
DIFFUSE MATTER IN SPACE 
This optical method of finding distances assumes that there is no 
appreciable loss of light by absorption or scattering during the journey 
across space. The direct evidence for this assumption is very limited. 
Some check is afforded by the general agreement of spectroscopic and 
trigonometrical parallaxes, but this gives an upper limit to the ab 
sorption too high to be of much service. The evidence which is usually 
quoted as indicating an almost perfect transparency of interstellar space 
is H. Shapley’s demonstration that the light from the stars in globular 
clusters (at a distance of the order 10,000 parsecs) shows no appreciable 
reddening; it is assumed that absorption without reddening is unlikely. 
According to modern ideas Shapley’s result does not really carry us 
much farther. We shall presently show that the reddening by diffuse 
matter is small in proportion to the absorption so that the absence of 
detectable reddening is no proof of absence of dimming. 
The weakness of the position is apparent when we consider the possible 
causes of dimming of a distant object. These are— 
1. Obstruction by particles large compared with the wave-length of 
light (meteoric matter). 
2. “Rayleigh scattering” by atoms, ions or particles comparable with 
the wave-length of light. 
3. Scattering by free electrons. 
4. Continuous absorption by gaseous material. 
Line absorption, discussed in §§ 258, 259, has insignificant effect on 
the general brightness of the star and cannot occur without betraying 
itself in the spectrum. 
Of these only the Rayleigh scattering definitely causes reddening. It 
is proportional to the inverse fourth power of the wave-length, so that 
blue light is eliminated from the transmitted beam faster than red. But 
we might almost have rejected Rayleigh scattering without observational 
test for reddening; to produce appreciable effect a vast quantity of inter 
stellar material is required which would be irreconcilable with dynamical 
studies of stellar velocities. 
The causes (1) and (3) act independently of wave-length and would not 
produce reddening. No. (4) is likely to be selective but whether it would 
make the light redder or bluer cannot be foretold. 
We have seen reason to believe that most elements lose their valency 
electrons in interstellar space; and in any case the fixed calcium lines 
constitute observational evidence that calcium is ionised. Hence the 
material contains free electrons. We shall consider whether the electron 
scattering can produce appreciable dimming of the stars. If there were 
500 free electrons per cu. cm., a column of 1 sq. cm. section and length