362 Proceedings of Royal Society of Edinburgh. [sess. 
reflecting surface depends the perfectness of the joining of the 
beams. 
It is. to be noted that the use of polarisation for regulating and 
measuring the light intensities is not prohibited by the adoption 
of this appliance, even when the light is polarised before being 
passed through the latter. No change of polarisation or produc- 
tion of polarisation can be caused by the entrance to or exit from 
the glass, for that only takes place normally to the various faces* 
If the two beams are plane polarised vertically and horizontally 
before entrance, with a view to the adjustment of their relative 
intensities later on by means of a Nicol prism, then because the 
plane of polarisation in each case is either in or normal to the 
plane of incidence on the silver surfaces no change of polarisation 
can occur. On the other hand, if the two beams have their 
respective planes of polarisation inclined to the vertical and to the 
horizontal, these beams, because they are each twice reflected at 
parallel silver surfaces, will emerge plane polarised still, though 
the plane of polarisation of each has been rotated to some extent. 
Hence in both cases the analysing Nicol can be used as before- 
for the purpose of measuring the light intensity, although the zero- 
will have been permanently displaced through a definite angle. 
A suggested application of the juxtapositor described above will 
be readily understood by anyone conversant with the construction 
of Laurent’s “half-shade” polarimeter. In that instrument two- 
parallel beams of light polarised in planes at an angle to one 
another are passed through a substance whose rotative power it is 
desired to measure, and are then analysed by means of a Nicol 
prism. By properly adjusting the position of the latter, the two 
half-circles of light seen in the eyepiece of the instrument, due to 
the two beams, can be made equally bright. It is found that in 
this way a much more accurate setting of the rotating Nicol can be 
obtained than when, as in the ordinary case, only one beam of 
light is employed and the Nicol is set to extinction. But the 
accuracy of the measurement in Laurent’s improved form of instru- 
ment turns on the degree of precision with which the eye is able to- 
determine when the two halves of the circle seen in the eyepiece- 
are equally bright. Now, these two halves are separated by a 
dark line, and accordingly, as explained above in the case of the 
