530 
edge of about 90°, the other two an edge of about 126°, the one 
pair being set symmetrically upon the other. This pointer has been 
patented. On turning it towards any object, and looking into the 
glass, we perceive an inverted image of the object, and on bringing 
the eye to look past the edge of the glass, as when using the camera 
lucida, we see the object also ; and so can collate the image with 
the object. 
By means of the motions in altitude and azimuth we can cause 
the image to agree with the object, and are then certain that the 
line of direction passes through the object. 
I do not wish to occupy the time of the Society with a disqui- 
sition on the theory of the pointer, but rather to indicate the peculiar 
facilities which it gives for astronomical and geodetical operations in 
which a precision to half a minute is sufficient. The instrument is 
inapplicable whenever greater nicety than this is requisite. 
When we point the telescope of a common surveying theodolite to 
a star, we find that the cross-wires have disappeared in the dark- 
ness, and that it is necessary to render the wires visible by making 
an artificial twilight within the tube. To hold the lamp for this 
purpose is so troublesome, unless — as in the case of large instru- 
ments — a special arrangement be made, that very few surveyors 
indeed ever use their theodolites for sidereal observations. 
But, by using the crystal pointer, the star itself is made the point 
of reference, and the illuminating apparatus is unneeded. 
If we bring the catadioptric theodolite into such a position as to 
cause the image of a star to agree with the star seen directly, and 
leave the instrument untouched, the two bright points are seen 
slowly to separate ; while the star moves forward in its diurnal 
circuit the image moves with the same velocity backwards, the rate 
of separation is doubled, and the line joining the two points indicates 
the direction of the motion. In this way it is easy to determine 
whether the star be ascending or descending, or to judge of its being 
at its greatest or least altitude. Thus the determination of the 
latitude from the meridian altitude of a star is very easy; that of 
the sidereal time is not less so. 
We have to note the exact instant at which some star, not far 
from the prime vertical, has a known altitude. For this purpose we 
turn the pointer a little in advance of the star, and bring the image 
to be very nearly above or below the star, according as the observa- 
