261 
1912-13.] Some New Types of Spectrophotometer. 
collimator and the motor anywhere near, on the table, etc. Also, I 
find that it is an excellent plan to get rid of the momentarily discarded 
beam at the earliest possible place in the optical train, because this gets 
rid of diffuse accidental reflections later on, which are often a source of 
serious trouble. To aid this, I have replaced the usual narrow collimator 
tube by a large blackened rectangular box, in which the momentarily 
discarded beam is effectually lost. The lens wheel just described was only 
tried after various other arrangements had been discarded, the objection 
to it being its apparent lack of symmetry. A symmetrical arrangement 
of two similar glass wedges set opposite ways and given a reciprocating 
motion had to be abandoned because of its mechanical awkwardness. 
This was succeeded by the device shown in fig. 5, where A and B are halves 
of the same convergent lens, and, turning on axles at C and D geared 
together by two toothed wheels, come alternately in front of the collimator 
slit E. The symmetry, however, is not perfect, because, in order that they 
may clear each other, one of the halves must be nearer the slit than the 
other ; and besides, there is the objection that toothed wheels always give 
rise to a certain amount of noise. It is therefore fortunate that it turns 
out on examination that the asymmetry of the lens wheel which would 
otherwise be so satisfactory does not cause any alteration in the relative 
intensity of the two beams, or at all events does not do so if the beams 
are regarded as made up of sensibly parallel rays. The proof of this 
statement is simple but a little tedious, and hence is omitted. 
(2) After the preceding description, the type of instrument employing 
reflection instead of refraction will be very readily understood. In this 
a “ mirror wheel ” consisting of a semicircular disc of mirror glass H, fig. 6, 
takes the place of the lens wheel of fig. 3, and when the mirror wheel covers 
the collimator slit the beam Y is reflected at 45° into the slit, and then 
when the slit is uncovered the other beam, X, passes into it. As before, 
the (imaginary) prolongation of the axis of the half disc intersects the 
(imaginary) prolongation of the slit either above or below the latter, but 
in this case the end of the collimator must be levelled away as shown, in 
order that the half disc may come sufficiently near. There is no difficulty 
about this, as the movable jaw of the slit, with its accompanying screw, etc., 
can be put on the other side, where there is plenty of room. Complete 
symmetry is secured by reflecting X as shown at a mirror placed either 
as M x or as M 2 . It will be observed that in the case of the mirror wheel 
the two beams are at the same horizontal height ; in the case of the lens 
wheel they are one above the other ; and hence it is convenient to have 
the instrument provided with a wheel of each sort to be employed as 
