342 Messrs. Trowbridge and Sabine on Wave-lengths 



glass, thus inferred, are so remarkably similar to the acoustic 

 properties of the monochord. 



20. I conclude, for the present, with several of the simpler 

 inferences that may be drawn from the preceding specification 

 of wave-surface. 



(1) Consider Fresnel's experiment with compressed prisms. 

 The light in that experiment is refracted in a plane perpen- 

 dicular to the line of strain. We see therefore, by reference 

 to the diagram of art. 18, that one of the polarized pencils (the 

 ordinary) is deviated by the strain, and the other not. 1 have 

 not attempted to verify this inference directly by experiment. 



(2) Consider a plate which is strained in one direction 

 parallel to its faces, and a polarized ray which is incident 

 upon it in the principal section, that is, in a plane parallel to 

 the line of strain. By calculation based on art. 18 or 19, I 

 find that the difference of retardations of the two transmitted 

 rays varies directly as the cosine of the (primitive) angle of 

 refraction. By exact measurements with a quartz compen- 

 sator, through a considerable range of incidence (0° to 70°), 

 I have succeeded in verifying this inference perfectly. 



(3) Consider a polarized ray, which is incident on the same 

 plate in a plane perpendicular to the line of strain. Both 

 refractions are ordinary, and the difference of retardations of 

 the two transmitted rays varies inversely as the cosine of the 

 angle of refraction. 



(4) Inferences (2) and (3) explain the following propo- 

 sition of Brewster's : — " If a plate of glass in a state of com- 

 pression or dilatation is inclined to the polarized ray in a 

 plane parallel to the axis of dilatation on compression, the 

 tints will descend in the scale ; but if it, is inclined in a plane 

 at right angles to these axes, the tints will ascend." But 

 here it should be noticed, that Brewster's order of ascending 

 tints is the same as the order of Newton's rings from central 

 black outwards, that is, the order of increasing difference of 

 retardations. 



Glasgow, August 25, 1888. 



XXXVII. Wave-lengths of Metallic Spectra in the Ultra- Violet. 

 By John Teowbkidge and W. C. Sabine*. 



Introduction. 



THE Catalogue of Metallic Spectra, revised by a Committee 

 of the British Association, and published in its volumes 



* From an advance proof from the i Proceedings of the American 

 Academy.' Communicated by the Authors. 



