TRANSACTIONS OP SECTION A. 



535 



background, and exposures over one hundred times the normal can be given 

 before the background makes itself evident. Photographs of the mercury arc in 

 vacuo with very long exposures showed many new lines, among which the lines 

 forming the continuation of the various series of mercury were strikingly 

 developed. The wave-lengths of the lines of the principal and of the diffuse 

 series have been measured up to the fifteenth, and those of the sharp series up 

 to the thirteenth. The results afford a strict verification of Rydberg's law — that 

 the frequency of the first line of the sharp series is equal to the difference of the 

 limits of the sharp and the principal series. 



6. On Apparatus for the production of Circularly Polarised Light and i 

 a New Form of White Light Half-shade. By A. F. Oxley. 



If we send a beam of plane polarised light through a Fresnel's rhomb — 

 the azimuth of the incident vibration being 45° with respect to a principal 

 section of the rhomb — the emergent beam of circularly polarised light is displaced 

 laterally from the incident beam. This necessitates readjustment of the suc- 

 cessive pieces of apparatus if we are to keep the beam of 

 light in the field of view as the rhomb is rotated. If we 

 use a quarter- wave plate to produce the circularly polarised 

 light no lateral displacement is experienced, but we are 

 limited to the use of monochromatic light in this case. 

 The idea of the present investigation was to devise a piece 

 of apparatus which would produce a beam of circularly 

 polarised white light without involving the inconvenience 

 of the lateral displacement previously referred to. 



Two pieces of apparatus have been devised with this 

 object in view. In the first, two rhombs of glass are 

 arranged as shown in fig. 1, the rhombs being in close 

 contact. A beam of plane polarised light enters at the 

 top and is reflected in turn from the surfaces A, B, C and 

 D, so that the direction of the energent beam is a con- 

 tinuation of that of the incident beam. If the refractive 

 index of the glass be 1"5035 for D (Uviol glass) and the 

 angle of the rhomb be 74° 38''2, the phase difference be- 

 tween the component vibrations polarised parallel and 

 perpendicular to the plane of incidence is equivalent to 

 5, and the emergent light is consequently circularly 

 polarised. If the angle of the rhomb be 42° 34'"8, we still 

 get the same effect ; but since the variation of the phase 

 difference between the components with respect to the wave-length is 

 much smaller for the larger angle, the latter is adopted. In fact, if we 

 choose the larger angle, the ellipticity of the transmitted light for any wave- 

 length is over half the ellipticity for the corresponding wave-length when a 

 Fresnel's rhomb is used, although there are four reflections in the former case 

 and only two in the latter. Hence the present form is more efficient than a 

 Fresnel's rhomb ; but it has the disadvantage of being inconveniently long, the 

 length being 15 cm. for an aperture of 1"1 cm. 



This difficulty has, however, been overcome to a large extent in the second 



Fig. 1. 



Fig. 2. 



form of polariser, in which three reflections are used. Fig. 2 shows the path 

 of a beam of light through the apparatus. The two blocks of glass are pressed 



