NATURAL PHILOSOPHY. 151 



arising from certain inconsistencies in the present theory, as to the 

 relation of the direction of vibration in polarized light to the plane of 

 polarization. The author then proposes what he calls the hypothesis 

 of oscillations, which consists mainly in conceiving that the luminifer- 

 ous medium consists of detached atoms or nuclei distributed through- 

 out all space, more or less loaded with atmospheres of ordinary matter, 

 and endowed with a species of polarity, in virtue of which three 

 orthogonal axes in each atom tend to place themselves parallel respec- 

 tively" to the three corresponding axes in every other atom ; and that 

 plane-polarized light consists in a small oscillatory movement of each 

 atom round an axis transverse to the direction of propagation and 

 perpendicular to the plane of polarization. The square of the velocity 

 of propagation of such a movement would be proportional directly to 

 a co-efficient depending on the rotative force or polarity of the par- 

 ticles in a given space, and inversely to a co-efficient denoting the 

 sum of the moments of inertia of the luminiferous atoms in a given 

 space together, with their loads of atmosphere, round the axes of 

 oscillation. The author shows that it is necessary to suppose that the 

 co-efficient of polarity for transverse axes of oscillation is the same in 

 all substances, and for all directions ; and that the variations in the 

 velocity of light depend wholly on the variations of the moments of 

 inertia of the luminiferous atoms with their loads, in different sub- 

 stances, and round different axes. The co-efficient of polarity for 

 longitudinal axes of oscilliation must be supposed to be very great 

 compared with that for transverse axes. How powerful soever the 

 polarity may be, which is here ascribed to the luminiferous atoms, it 

 is a species of force which must necessarily be wholly destitute of 

 effect in producing resistance to compression or distortion, so that it is 

 no longer necessary to suppose the luminiferous medium to have the 

 properties of an elastic solid. The author deduces from this hypothe- 

 sis the known mathematical laAvs of the wave-surface, of the intensity 

 and phase of reflected and refracted light, and its plane, circular, and 

 elliptic polarization, and of all other phenomena to which the existing 

 theory has been applied, the equations being identical in form. 



BADIATION OF LUMINOUS BODIES. 



On looking at a very brilliant light, it sometimes appears to be sur- 

 rounded by brilliant luminous rays, clear, free from cloudiness, and 

 which must not be confounded with those caused by the eye-lashes 

 when the eyes are partially closed. These rays may be observed 

 most distinctly by looking at an image of the sun reflected upon the 

 surface of a convex glass, or still better upon a lens having consider- 

 able curvature. They may be most easily observed by looking at an 

 image of the sun formed in the focus of a lens placed at the extremi- 

 ty of a tube blackened in the interior. If the observer place him- 

 self in a room into which the light penetrates only through a narrow 

 opening, the phenomenon appears with great splendor, and it may 

 even be said with extraordinary magnificence. The rays are either 



14* 



