122 Dr. T. Ewaii. On the 



duced on to the slit/. The tube K can be rotated round the centre 

 of the iron table (represented in the figure by a circle ; it was part of 

 a large Steinheil spectroscope), and thus the slit / can be made to cut 

 out portions of any desired wave-length from the spectrum. The eye 

 placed behind the slit then sees the surface of the photometer prism 

 in the colour of the part of the spectrum which is allowed to pass by 

 the slit. Both slits were provided with micrometer screws, the heads 

 of which were divided into 100 parts, and one whole turn of which 

 opened the slit O5 mm. The axle round which the telescope, K, 

 turns ended below the table in a small mirror, which turned with 

 the telescope, and which served for reading its position by means of 

 the millimetre scale and telescope, MM. The wave-lengths of the light 

 passing through the slit, /, were measured by means of a small spec- 

 troscope with graduated arc, the readings of which were converted 

 into wave-lengths by means of a curve drawn from readings of the 

 positions of the principal Fraunhofer lines. The lamp was enclosed 

 in a blackened box, and all the parts of the apparatus covered as 

 much as possible by blackened screens. In the later measurements 

 the lamp was connected with a gas pressure regulator, as variations 

 in the gas pressure are apt to cause changes in the ratio of the 

 brightness of the two beams of light. 



A difficulty which was encountered at first was due to the ends of 

 the glass tubes containing the solutions not being exactly parallel. 

 They act as prisms of very small angle, and deflect the light which 

 passes through them to some extent ; so that, after adjusting the two 

 images of the slit, c, which are formed by the two lenses, a and b, to 

 fall together at/, and putting in the tube of solution, the two images 

 no longer coincide. The light coming from the lens, 6, is unchanged, 

 while that from a has a colour differing from that which it had at 

 first. It was therefore necessary to have some method of adjusting 

 quickly and accurately the two images to fall together. This was 

 done by allowing the light reflected from the front surface of the 

 prism, G, to fall on a mirror, N, from which it was reflected into the 

 telescope, 0. This was an ordinary telescope, in which the two 

 images of the slit, c, were seen, side by side, as two bright lines, 

 which could easily be brought to coincidence by rotating somewhat 

 the prism, F. A glance at fig. 2 will show that, by a small rotation 

 of this prism, the direction of the light from the lens, 6, is unchanged, 

 while that of the light from a is changed by twice the angle through 

 which the prism is rotated. 



The careful adjustment of the apparatus was also necessary in 

 order to obtain good results. This was done briefly as follows : 



The collimator lenses were placed at their focal distance from the 

 slit, r, and the lamp arranged (the ground glass plate being removed) 

 so that they were both just filled with light. The triangular ground 



