Then, reading the declination circle, one can set off the co-atitude or latitude as the case may be and, leaving the telescope clamped in declination, lower or raise the polar axis until the tube levels to the horizontal. When the mount does not permit wide adjustment and has no latitude scale one is driven to laying out a latitude templet and, placing a straight edge under the equatorial head, or suspending a plumb line from the axis itself, setting it mechanically to latitude. Now suppose we are dealing with the same instrument, but are planning to plant it permanently in position on its pillar mount. It is now worth while to make the adjustments quite exactly, and to spend some time about it. The pillar is commonly assembled by well set bolts on a brick or concrete pier. The preliminary steps are as already described. The pillar is levelled across the top, the equatorial head, which turns upon it in azimuth, is levelled as before, the adjustment being made by metal wedges under the pillar or by levelling screws in the mount if there are any. Then the latitude is set off by the scale, or by the declination circle, and the polar axis turned to the approximate meridian as already described. There is likely to be an outstanding error of a few minutes of arc which should in a permanent mount be reduced as far as practicable. At the start adjust the declination of the optical axis of the telescope to that of the polar axis. This is done in the manner suggested by Fig. 172. Here _p_ is the polar axis and _d_ the declination axis. Now if one sights, using the cross wires, through the telescope a star near the meridian, i.e., one that is changing in declination quite slowly, starting from the position _A_ with the telescope _E_. of the polar axes, and turns it over 180° into the position _B_, _W_. of the polar axis, the prolongation of the line of sight, _b_, will fall below _a_, when as here the telescope points too high in the _A_ position. [Illustration: FIG. 172.—Aligning the Optical Axis.] In other words the apparent altitude of the star will change by twice the angle between _A_ and _p_. Read both altitudes on the declination circle and split the difference with the slow motion as precisely as the graduation of the declination circle permits. The telescope will probably not now point exactly at the star, but as the tube is swung from the _A_ to the _B_ position and back the visible stars will describe arcs of circles which should be nearly concentric with the field as defined by the stop in the eyepiece. If not, a very slight touch on the declination slow motion one way or the other will make them do so to a sufficient exactness, especially if a rather high power eyepiece is used.