Iii Kitj. 3, let OMB lepiesent the geometrical axis of the tube, 

 as determined by the collars, aiui let it In- assumed that the objec- 

 ti\e is eccentrically mounted as indicated. 



If such an objective were collimated on a Aery distant object, 

 the incident beam \vould enter the objective in nearly parallel rays and, 

 whether it were right-side-up or up-side-down, the collimated cioss 

 wires \\ould incur in the optical axis at the principal focus at A; the 

 line of collimation, NA, would be parallel to the collar axis. Ml',. 

 and this would still be true if, in addition to its eccentricity, the 

 objective were slightly inclined to the axis of the tube. The only 

 error induced, therefore, while leveling on distant objects, would be 

 a rather imperceptible plus or minus quantity equal to a. 



When focused for collimation on a comparatively near object, 

 as at O, the image will be formed on the line joining O with the 

 nodal point of the objective, N, as at C, where the cross wires 

 would have to be located. If the telescope were inverted, the image 

 would occur at Ci but the cross wires would also occur at this point 

 and no error would be noticed except that the field and cross wires 

 would appear to move around together in a small circle. 



For example, let it be assumed that OM D = 50 feet, that 

 MN 1 mm and that angular error e. For very small angles 



a 

 tan t nn . 



Substituting these values in the formula given above, 13. .V, 

 or .006 feet per 100 ft. 



The Principal Bubble may have a sensibility of 10 sec. , 5 sec. , 

 or even more if desired. It is mounted at the top of the telescope 

 in a brass box encasement and otherwise protected from dust and 

 sudden temperature changes by a crystal glass window. 



The cover of the box contains a reading mirror* which snaps 

 down in place when not in use. The cover is hinged on universal 

 pivots to absorb a shock from any direction and reduce the 

 probability of breakage. 



To Adjust the Precision Bubble turn the mirror back- 

 ward and off to one side; remove the retainer-plate and slide back 

 the glass cover only enough to expose the front adjusting screw. 

 A preliminary test may be conducted by end-for-end reversals in 

 the wyes but it should be verified by the pegs, particularly if there 

 is a suspicion of unequal wear of the collars. (See page l l >). 



* 1C. E. U'hittifr in Eng, & Cont. June 12, 1912 says, "Theory. 

 level is inaccurate for precise work and 4 4 it is not equipped with a 

 4 4 the advantages of which 4, 4 cannot be (ix> highly recommended. 



13 



