64 MR ADIE ON THE USE OF METALLIC REFLECTORS FOR 
does not lie so long in the water as to allow the rust to commence, if it be care- 
fully wiped dry, when taken out, there is no danger of spoiling afterwards. Now, 
this is a very important matter, and more particularly to surveyors, than any 
other class of nautical men; for such accidents frequently occur in boat-sound- 
ings.” 
I come now briefly to notice the second subject announced in the title of 
this paper, viz., direct methods of determining the errors in the mirrors and sun 
shades used in reflecting instruments. I am not aware of any method adopted 
by practical men for the discovery of such errors, except that of a careful process 
of what is termed parallel grinding, and testing the glasses in the instrument when 
fitted up. This is effected by observing known angles, and noting that a contact 
of the sun’s limb, by reflection, does not vary on changing the sun shades inter- 
posedbetween the direct and reflected images seen in the telescope. This varia- 
tion is noticed by Mr M‘Kay, in his work on Determining Longitude, and he re- 
commends that it should be observed and applied as an index error affecting the 
several shades. Errors may, however, exist in the reflectors and shades, which, 
from the particular position they have in their settings, are not discovered by such 
trials. The methods I have adopted are as follows :— 
First, For the mirrors I place a mirror at about an angle of 45° before the 
object-glass of a telescope, mounted on a divided circle, capable of reading an 
angle of 10’; I have, besides, a moveable micrometer wire at the stop holding the 
cross wires of the telescope, by which an angle of one second may be observed ; 
the mirror placed before the object-glass rests on three smooth studs, to which it 
is pressed up by a light spring at the back. In this position we turn the tele- 
scope and mirror in azimuth till we obtain the reflection of a well-defined distant 
object, which is brought to the intersection of the cross wires of the telescope. If 
we now turn round, or reverse the mirror on these studs, and find the same object 
in the intersection of the cross wires, we know that in that line there is no want 
of parallelism; and if we have the same result on repeating the trial at right angles 
to the first direction, the mirror is said to be perfect in respect to the parallelism 
of its surfaces. If, however, we find on turning the mirror that the reflected 
object is not intersected by the cross wires, then, the glass is not parallel; and 
half the amount of error read by the circle or micrometer head, is the error which 
would arise if such a mirror were applied to a sextant. 
Second, For the sun shades I have a telescope whose object-glass is divided, 
and the one-half moved over the other by means of micrometer screws, having a 
divided head. The value of the divisions of the micrometer head is obtained by 
measuring the sun’s diameter. In that which I use, two divisions on the head are 
equal to one second. 
Before one-half of the object-glass is placed a fixed sun shade ; and before 
