Prof. S. P. Thompson on a New Polarizing Prism. 349 



originally were each bisected by the axis of y, have then come 

 to occupy the two remaining opposite quadrants. Thus the 

 one set of curves of constant induction merges into the other, 

 and the forms of both sets, which may be described as ovals, 

 depend upon the angle between the axes of the primary and 

 secondary coils. 



We see, then, that the second set of curves, although found 

 experimentally, are not entirely independent of the first set, 

 and that being positive and negative in alternate quadrants, 

 they have no corresponding surfaces of revolution, that the 

 only curves of constant induction which have corresponding 

 surfaces of revolution are those which are obtained when the 

 axes of the primary and secondary coils are parallel, and that 

 all curves of constant induction which are due to the coils 

 when their axes make any other angle with each other are 

 merely modified forms of them. 



The arrangement of apparatus employed in this investigation, 

 namely two similar coils in the primary and two other similar 

 coils in the secondary circuit, was employed by Dove in his 

 researches, and was called by him the " Differential Inductor" 

 (Annates de Chimie et de Physique, tome iv. 1842). 



The same arrangement was employed by Felici, a short 

 account of whose researches is given in Maxwell's ' Electricity 

 and Magnetism,' vol. ii. pp. 169-172. 



In conclusion, my special thanks are due to Prof. G. C. 

 Foster, for the general interest which he has taken in this 

 subject, and for many important suggestions made during the 

 progress of the investigation. 



XLIV. On a New Polarizing Prism. 

 By Professor Silvanus P. Thompson, B.A., D.Sc* 



1. 1VTEITHEE the polarizing prism of Nicol nor that of 

 li Foucault can be regarded as perfect. The latter 

 especially has so small an angular aperture available, as to be 

 very inconvenient for any but narrow beams of parallel light. 

 The author has sought to improve upon the existing forms of 

 polarizing prism ; and his investigations into the cause of 

 their defects have led him to produce prisms having a consi- 

 derably wider effective angular aperture. 



2. In the text-books it is usual to tell students that in the 

 Nicol prism the ordinary ray is suppressed by total reflexion 

 because the ordinary index of refraction is greater than that 



* Communicated by the Author, having been read before Section A of 

 the British Association, September 7. 1881. 



