326 Dr Brewster on a peculiarity in Glauberiie, 



they seemed to be, to a single axis of double refraction, could 

 be calculated in the most rigorous manner, by supposing the 

 crystal to have three positive axes at right angles to each other, 

 each of which exercises a different dispersive action upon the 

 differently coloured rays. This result, which is of considerable 

 importance in the theory of double refraction, is strikingly 

 confirmed by the phenomena of Glauberite, while these at the 

 same time present us with a new and still less equivocal case of 

 the composition of axes. 



In the case of Glauberite, observation exhibits to us one ne- 

 gative axis A, which is the single axis for the violet light, and 

 the principal axis for the red and the other less refrangible 

 rays; and, at the same time, it presents to us a second axis B, 

 which may be either negative or positive , but which must be 

 90° distant from A. If it is negative, it must be in a plane 

 perpendicular to the plane passing through the two resultant 

 axes for red light ; and it must bear to A the ratio of the 

 square of the sine of 2|° (half the inclination of the resultant 

 axes) to unity. * If it is positive, it must lie in the plane pas- 

 sing through the resultant axes, and it must bear to A the 

 ratio of the square of the sine, to the square of the cosine of 

 2J°. But whether it be positive or negative, it exercises no 

 action whatever upon violet light, a supposition so absurd, 

 that it cannot for a moment be received. Since the combina- 

 tion of axes, therefore, indicated by experiment for the 

 single system of rings in violet light, and for the double 

 system in the other rays, involves a physical absurdity, we 

 must seek for a new combination, not liable to such an ob- 

 jection. 



If we suppose that the axis A for violet light is the result- 

 ant of other axes, and that these other axes are two positive 

 axes B and C at right angles to each other, and also to the ap- 

 parent axis A, we shall obtain an explanation of all the pheno- 

 mena. If the axes B, C, exercise the same action on the violet 

 rays, they will produce a single negative axis at A for violet 

 light, as given by observation ; and if the relative intensities 

 of their action upon red light are in the ratio of the square of 



• See VhHos<>]>hical Transactions, 1818, p. 237, &c. 



