( 889 ) 
the region of their proper periods; but, strictly speaking, co-vibrating 
takes place in some degree throughout the spectrum, though with 
decreasing intensity as one recedes from the line corresponding to 
the free vibration. Now, according to our hypothesis, the same cannot 
be said with regard to absorption, this process being confined to a 
narrow part of the region of resonance. 
Let us consider the share which scattering has in the act of damping. 
Within the region of strongest absorption it keeps in the back-ground ; 
rye pe 
: 7 ‘ . A spe 6 —vi) 
indeed, for sodium light the factor ——- amounts to only 2 x 10 ~, 
OC 
while the value of the entire damping parameter in sodium vapour 
BA 
was found to be: h=mr'=7 X Les Sc tl 2 Se 10° —7.8 «10 ? 
But on the other hand, we have no ground to doubt, as in the case 
of absorption, the proportionality of this damping effect to the velocity, 
however small it may be®). Scattering extends all over the spectrum, 
wherever the proper frequencies of the electrons may be found. 
Indeed, apart from the theory of absorption and dispersion RayLuicu 
has proved that a beam of light of intensity -/, and wave-length 2, 
after having travelled a distance er through a mass of gas whose 
index of refraction is n, and which contains MN scattering molecules 
per unit volume, will have sunk to the intensity 
32 x3 (n—1)? 
S= ae SMN ni EEA MO rents 140 (9) 
: aan (n— il)? 
The quantity s= Semele) 
32!.N 
is called the coefficient of scattering. 
Rito | acai 
Let A be the density of the gas, then any =f may be considered 
constant for any given wave-length. As we may put A proportional 
- Nt sr el RN 
to N, say A=/.N, the expression — rae =f. k is also a constant 
a 
for any definite kind of light. Introducing it into the coefficient of 
scattering, we find 
820° Nf? R? 320° Af. R? 
Pa ak en dae aa 
where # represents the refraction constant of the medium. 
The coefficient of scattering is thus inversely proportional to the 
Ss 
(10) 
') From observations by Harro, concerning the magnetic rotation of the plane 
of polarisation in sodium vapour, Vorer calculated : v'= 11,2 > 102. Cf. Vorar, 
Kc: p: 142. 
*) If we admit Prarck’s new radiation hypothesis, this statement will have to 
be corrected. 
58* 
