When the instrument is reversible, the axis is first to be set truly upright. Suppose the instrument in its meridian position nearly, and face east, read off the division bisected by the plumb-line, or the two ends of each level. Now turn half round, read off again, and bring, by the adjusting screws, the plumb-line or the levels half-way to the first readings, and finally adjust each level by its own screw to read each cud alike. If this be carefully done, when the instrument is restored to its first position, the plumb-line or levels will remain undisturbed by the last reversal. Now turn the axis one quarter round, and correct whatever change is thereby caused, by the east and west screws of the axis. The axis is now vertical, or by a repetition of the process may be made so. The next adjustment is to make the line of sight describe a great circle. This is the collimation error of the transit. This may be done as described above, from knowing the true time ; or by observing one star or two stars near the zenith in reversed positions, when the disagreement between the observed and computed difference will give the quantity and direction of the alteration required. In a modern instrument this adjustment would be by antagonist screws carrying the wire-plate. If the instrument be simply a zenith-tube, make a
star run along the declination-wire, and the adjustment is finished. With a sector place the instrument nearly in the meridian, observe the transit of a zenith star, which gives the time. Then by turning round the axis, make an extreme star pass at the proper time and clamp the , axis. In the present ordnance sector the instrument rests on a tray which it adjusted as to meridian by strong screws on the stand, acting against the sides of the tray. Finally, twist the wire-cell till a star i runs along the declination-wire. A comparison of the zenith distances of the same stars observed in reversed positions of the instrument, will give the error of collimation, and this may be corrected if the observer wishes, but It is better to leave it untouched, and to consider the sum of two observations, Face East and Face West, as a double zenith distance.
(For plates and descriptions of some of the constructions here referred to, and others which we have omitted, see Pearson's Practical Astronomy, vol. ii., pp. 531, 554, plates xii., xiii., xxvi., xxvii.)