1903 - 4 .] Mr Milne on a New Form of Spectrophotometer. 339 
needed, provided that the light rays from any the same point of 
ah and cd are all in a horizontal plane through that point, and this 
condition, as will be seen later on, is actually fulfilled. 
The edges a and d' of the two spectra formed by the objective of 
the telescope are necessarily somewhat hazy and ill-defined, whether 
the gap between the spectra has been produced by the menisci of 
a liquid or by the edge of a solid. To remedy this, a strip of 
metal ad is placed so as to cut off the extreme edges of the two 
spectra, and by this means the edges which are afterwards brought 
into contact in the plane SS' are beforehand made perfectly 
straight, and are sharply delimited. This strip of metal or 
“ trimmer ” ad (fig. 3) really consists of two similar pieces, which 
by means of a slow motion screw can be arranged in such a way 
that the compound strip is slightly wider at one end than at the 
other. This arrangement the author has found to be necessary, 
as in his model instrument, for reasons of economy, the divided 
lens is a simple one, and so the neighbouring edges of the images 
of the two spectra formed by the divided lens are not parallel 
to each other. This difficulty is perfectly overcome, however, by 
making the trimmer slightly wider at one end or the other as 
may be required. The point is mentioned because even with a 
more perfect lens the device might be necessary to obtain the 
most exact results. 
It inevitably happens that the two beams of light falling on 
the trimmer ad in fig. 3 suffer marked diffraction, and if (say) the 
lower beam be stopped off, obvious diffraction bands at the lower 
edge of the remaining spectrum may in general be seen on looking 
through an ordinary telescope eyepiece, placed behind the divided 
lens at a distance of about twice the focal length of the latter. If, 
however, the eyepiece, originally somewhat too far off to focus 
objects in the plane SS', be slowly pushed nearer that plane, the 
diffraction bands, which in this case are dark and are .situated 
upon the bright strip, are observed to begin closing in towards the 
edge of the image, and when the eyepiece is exactly focussing the 
plane SS' no bands are to be seen at all. On continuing to move 
the eyepiece towards the plane SS' the bands reappear, being now 
bright lines- situated outside the bright strip, and they continue to 
move out from its edge with the motion of the eyepiece. These 
