Physical Discovery. 91 
rected to the north, at a certain deviation, which, for Paris and its 
environs, is known to be about twenty degrees. The cap of the 
small needle, is then taken out, which leaves a hole in its middle, 
and thus admits of its being placed on the cap of the large needle and 
pressed down to the latter. ‘The cap therefore should not have the 
exterior form of common compass needles, but that of a small ver- 
tical cylinder, whose height should not exceed that of the needles 
generally used. ‘The small needle is then slipped down upon this 
little cylinder, like a ring upon the finger, so as nearly to touch the 
large needle ; as for reasons, which are too long here to state, it is 
necessary to prevent the immediate contact of the two needles; and 
to effect this, it is expedient to place a very thin leaf of copper be- 
tween them. By this means, the two needles are suspended upon 
one and the same cap, but imperfectly crossed, in such a manner as 
to represent the cross of St. Andrew, or a pair of scissors more or 
less opened. When satisfied how great must be the separation, it is 
necessary to establish, the small needle must be secured in its posi- 
tion by a little sealing wax, or in any other manner, so that, as the 
combined needles turn upon the pivot, they be not deranged, by the 
air, a jar, or any other cause. As to the exact extent of the separa- 
tion of the points of the needles, it is impossible to determine or adjust 
it, by an invariable theory, because this depends upon the size and 
relative magnetic powers of the two associated needles; it is there- 
fore, necessary to ascertain, for each set of associated needles, what 
the extent of this separation of the needles should be, according to 
their relative powers, which will require an experiment of a few min- 
utes duration. ‘T’o perform this, the needles are to be arranged on 
the cap, more or less crossed, and the distance between their points, 
is to be enlarged or diminished, until it is perceived that the large 
needle, in moving freely, at last is directed accurately to the true 
north. Now this is easily accomplished, after having tried and ex- 
amined, for a minute or two, various distances, until the requisite de- 
gree of separation has been discovered. 
It must not be forgotten, that the essential conditions of the opera- 
tion require, that the two needles be crossed in such a manner, that 
the north point of the small needle, be directed to the left of the north 
point of the large one, between the west and the north, and the other 
one between the east and the south; for if the north point of the small 
needle, from the nature of the separation, cross the large needle in 
two opposite directions, this erroneous disposition, augments the ac- 
tual deviation of the magnet, instead of rectifying it. 
