346 Proceedings of Royal Society of Edinburgh. [skss. 
another line focus GH, the distance from GH to the lens being 
equal to the focal length of the curvature on the hack face of the 
lens. Now this lens, if placed behind the prism R in fig. 4, will 
form two pure spectra in the plane normal to the paper through 
the line PQ. Further, if in that figure the strip of metal called 
the trimmer he placed immediately behind the lens of the 
collimator, we can arrange, by properly choosing the radius of 
curvature of the back face of the lens, that the plane in which 
the trimmer is placed shall be conjugate to the plane in which 
the spectra are produced ; and this fulfils our second condition. 
The two spectra so formed from the two beams respectively 
will exhibit a dark gap between them, and therefore, as before, 
the lens is to be cut through the centre in a horizontal plane, 
and then on separating the lens-halves to the required extent the 
two spectra can be moved towards each other till their edges come 
into perfect contact. In fig. 6/3, there are shown two beams of 
homogeneous light, and the resulting lines EM, MF (which are 
two elements of the two spectra that would be formed in the 
general case) are drawn as they would be if brought with their 
ends just to touch each other by an appropriate separation of the 
lens-halves L and L'. 
With a simple lens, on bringing the edges of the two spectra 
near each other, it can be seen that they are not parallel. This 
is due to a mixture of the errors of distortion and of chromatic 
aberration of the lens. To remedy this it would be of no 
avail to make the trimmer wider at one end, as explained on 
page 339 ; for reflection will show that that would merely reduce 
the intensity of the light which forms the adjacent edges of the 
