The direction of the Magnetic Meridian, as indicated by that of a freely suspended magnetized needle will generally be erroneous, unless the magnetic axis of the needle is parallel to its axis of figure; and hence, in order to obtain an accurate value of the magnetic declination, it becomes necessary to take the mean of two observations of the needle, first suspended in its usual position and next inverted. If, however, the inversion of the needle is not accomplished with perfect accuracy, the correction, for want of parallelism between the magnetic axis of the needle, and its axis of figure, will not be complete; and the value of the magnetic declination obtained from the mean of two observations of the needle, first in its usual position, and then inverted, will be affected with a residual error due to imperfect inversion of the needle. The present investigation refers chiefly to that form of declinometer magnet, in which the magnet is converted into a collimator by attaching to it a lens and cross fibres or a divided glass scale, in the principal focus of the lens.