The adjustment for parallelism should now be tested again, for 

 these adjustments are inter-related and both are dependent upon 

 the perfection of the collars. 



, The procedure thus far does not profess to test the bubble-axis 

 for parallelism by direct comparison with the line of sight. The 

 collar axis is the intermediary basis of the test. The only direc t 

 means of comparison available in the field is the "peg method", 

 and with worn-out collars this is the only method by which the 

 Y-level can be adjusted. 



Fig. 3 



Theoretically we now have the optical axis, DO, which is 

 understood to be coincident with the collar axis, CCi and the bubble 

 axis, BA, arranged in parallel lines, which for convenience of opera- 

 tion, must be adjusted to right angles with the vertical axis, VX. 



Although very desirable this adjustment is not vital. The ad- 

 justments might be stopped at this point so far as accuracy is con 

 cerned, but it is highly desirable to facilitate the work that the 

 sight-line shall describe a true horizon during a revolution on the 

 vertical axis. 



The Wye Axis is the axis of an imaginary cylinder, of which 

 the contact points with the collars are longitudinal elements. It 

 is immediately related to the vertical axis and is to be tested by 

 leveling up carefully again over either set of leveling screws. 



Re\olve the instrument on the vertical axis 180 as closely as 

 may be gauged by the eye; correct half the bubble displacement in 

 the leveling base and the rest by raising or lowering one of the v\ \ e 

 supports as the case may require. Make the test across the other 

 set of leveling screws and continue the operation until the bubble 

 remains centered during an entire revolution. 



This operation brings the axis of the bubble perpendicular to 

 the vertical axis and incidently adjusts the vertical axis to perfect 

 vertically. It can always be perfected \\itluuit respect to con- 

 ditions in the collars or others portions of the instrument. The 

 chief precaution in making the test is to look for lost motion in the 

 vertical axis. 



