If manufacturers were frank with their clientele, they would 

 acknowledge the difficulty of producing reversion bubbles that are 

 ground to the same sensitiveness on both sides. To get the oppo- 

 site scales to register perfectly also requires ingenuity, and the 

 reversion bubble is worthless as a means for testing collimation 

 (see p. 70) or collar inequalities in wye levels (see p. 14) unless they 

 do. 



We had better reverse the premises. Instead of testing the 

 horizontal collimation with the reversion bubble which, however, 

 s not impossible if the maker is faithful to his obligations, let us 

 collimate as described under V, b-1, p. 68, and test the reversion 

 bubble as follows: 



Level the telescope by the obverse scale and note the particular 

 point in the horizon cut by the horizontal wire. Reverse the telescope 

 and sight the horizon point a second time. If the bubble lies in 

 the center of the reverse scale, the requirements are satisfied; if not, 

 either the bubble axes are not adjusted to parallelism, or the oppo- 

 site scales do not register. Run over the peg test and prove 

 the first condition and finally test both scales against the horizon 

 line so established. If both scales will not check up against this test, 

 center the obverse scale with the adjusting nuts of the vial tube. 

 Invert the telescope and either note the number of divisions con- 

 stant error to apply with the proper sign in future w r ork, or paste 

 on a piece of paper and mark thereon a new normal or central posi- 

 tion for the bubble. 



Our reversion bubbles are guaranteed. Whether both sides 

 are of the same sensitiveness is of no serious concern, but the tan- 

 gential axes of both scales must be parallel. They are covered with 

 a revolvable protection guard, not shown in Fig. 51, that is painted 

 white on the inside to use as a background for either scale. 



The Control Bubble 



The Control Bubble performs an office similar to that of the 

 Latitude Level but for quite a different purpose. When the hori- 

 zontal wire is properly collimated and the telescope bubble adjusted 

 to this sight line, and centered, the index lines of the vertical circle 

 and its verniers must coincide. These various steps are taken up 

 in detail on pp. 68 to 72. 



Whereas the plate bubble lying across the longest dimension 

 of the telescope is most important in preserving ideal conditions in 

 the horizontal axis, the plate bubble running parallel with the tele- 

 scope is the one that is of greatest importance in the matter of 

 correct vertical angles. If the horizontal plates are not exactly 

 horizontal in the direction of the line of sight, no appreciable error 

 will be observable in the horizontal angle, but the vertical angle 

 read will be in doubt. 



94 



