220 W. A. Norton on Molecular Physics. 
phere. Accordingly the red rays will proceed from the lower 
portions of the atmosphere, and the violet from more elevated 
portions. 
If the electric atmospheres diminish in density by insensible 
egrees from bottom to top, there should be no break in the con- 
tinuity of the rays between the two extremes. But we know, 
from the existence of bright bands in the spectra obtained from 
colored flames, and from the highly heated vapors of metals and 
other substances, that the rates of vibration of the luminous 
iven off by incandescent vapors, seldom, if ever, vary by 
insensible degrees from one extreme to the other. e m 
conclude therefore that the electric atmospheres of highly heated 
molecules are made up of distinct layers of unequal density. 
henomena, attending the propagation of light. The absorp- 
tion of light by a medium may be regarded as the taking up of 
the ethereal pulses by the electric atmospheres of the medium. 
In order that a ray may be completely absorbed it must encoun- 
ter a layer of the electric atmosphere of a molecule, which puls- 
ates naturally in unison with the wave-pulsation of the ray. 
n 
the other hand, a medium would intercept rays that should 
penetrate to atmospheric layers that are in unison with the rays. 
Accordingly, if an incandescent vapor should emit rays of cer 
tain colors, as shown by bright bands in its spectrum, those col- 
ors, if transmitted through the vapors, should be absorbed, and 
the spectrum given by erriadiitted hi ght should show dark lines 
answering to the bright lines of the other spectruam—which is 
e well known discovery of Kirchhoff and Bunsen. 
wa Hssiteoe in the directions of the molecular axes. 
refraction, and that this property is attributed to a supposed } 
equality of density, or of elasticity, of the ether in the direction 
lar axes. A mechanical pressure exerted alon 
also develops the property of double 
