1903 - 4 .] Mr J. R. Milne on a New Form of Juxtapositor. 361 
will, in general, be less than it would be were such a' junction as 
BG (fig. 2/3) arranged in the path of the upper beam. Moreover, 
in such cases to make the junction OD optically more homo- 
geneous, such a liquid as the well-known a-monobromonapthaline 
might be used between the faces of the glass blocks instead of 
the ordinary cement. 
The following experiment was undertaken with the view of 
approximately ascertaining the amount of light reflected by such 
a cemented junction OD as occurs in this juxtapositor. The face 
DE of the latter was blocked up by an opaque screen, so that no 
light could pass through. The apparatus was then brought near 
a window, and the image of the latter produced by reflections 
at the silvered part CO of the interface, and at the silvering on 
the face HG was observed by looking into the face HC. No 
image whatever could be observed caused by a reflection from the 
unsilvered part OD of the interface, and not even an increased 
darkness could be seen corresponding to the places where the 
images of the window bars would fall. As a still more stringent 
test, the juxtapositor, with the lower face DE blocked up as 
before, was brought quite close to an incandescent electric lamp. 
In this case an image caused by reflection from the unsilvered part 
OD of the interface could be seen, but the image did not show 
the glass or brass fittings of the lamp, but only the glowing 
filament itself. Accordingly, it is clear that while there must be 
some difference between the refractive indices of the glass and of 
the cement used in the juxtapositor, which gives rise to reflection 
of light at the cemented surface, the fraction of the total light so 
reflected is very small indeed. It was noted also that the colour 
of that part of the glowing filament which was reflected by the 
unsilvered part OD of the interface appeared to be unchanged, 
which indicates that the small difference in the refractive indices 
of the glass and of the cement must be at least approximately 
constant for different wave-lengths. 
The edge of the silvered part of the face PQRS (fig. 1) of the 
upper block of glass is cut off very trim and sharp by means of an 
ivory chisel and nitric acid. This is a most important point in the 
construction, because it is at this place that the two beams of 
light unite, and on the abruptness of the termination of the 
