68 
POPULAR SCIENCE REVIEW. 
no light in that particular part of the spectrum. If the beam 
be broad, the deficient part is spread over too wide a space, 
overlapping and being overlapped by the contiguous rays ; but 
if the beam be narrow, the deficient part is restricted to its 
true dimensions. Hence the necessity of a narrow beam 
passing between two straight edges. However, since, on 
reducing the width, the amount of fight is reduced, there 
is in all cases a particular width which gives the best 
general result ; and therefore it is always desirable to 
have the means of varying the width of this narrow opening 
or slit by a screw movement. Now, with such a compound 
prism mounted in a tube as I have described, all that is abso- 
lutely requisite, to obtain a good spectrum, is to have such a 
slit which can be inserted into an ordinary eye-piece like a 
micrometer, so that it may be in the exact focus. Then, on 
placing the cap holding the prisms over the upper lens, a very 
good spectrum is seen. There is, however, this disadvantage 
in what would otherwise be a very simple arrangement — that 
the upper lens is not achromatic, and therefore the whole spec- 
trum is not in focus at the same time ; and moreover, since it 
is often most useful to be able to compare two spectra side by 
side, I consider that it is well worth while to have a special 
•eye-piece, as shown in fig*. 1, which can be adapted to any 
microscope, and comprises, with the prism, all that is requisite 
for the majority of objects. The upper lens (a) is compound 
and achromatic, and is mounted so that the focus may be 
adjusted by turning round the milled head b b. The slit is 
supposed to be represented by the opening c d, and a small 
right-angled prism ( e ) is fixed half over it, so that fight passing 
through an opening at / is reflected through the half of the 
Hit, on the side d ; whilst fight coming up the body of the 
microscope through the field- glass g, passes through the other 
half on the side c. Therefore, when the analyzing prism is placed 
over the eye-piece, we see side by side the spectrum of the 
fight passing from /, and that passing from g ; and if the 
reflecting prism i-s properly adjusted, the two spectra are quite 
continuous. In order to hold objects in front of the opening 
at /, a stage is attached as shown at h i, with appropriate 
springs to hold flat pieces of glass, or test tubes, which fit 
into Y-shaped notches at h and i, and. fie flat on the surface 
of the stage. Since the amount of fight passing up the body 
of the microscope is much less than that passing through f, 
on account of being spread over a larger surface by the 
magnifying power of the object glass, it is requisite to have 
a second vertical slit at /, which can be made wider or narrower 
till the two spectra are of equal brilliancy. 
In using this eye-piece with a simple microscope, the 
