New Eixjvne for Rulhig. 71 



and advanced past a small tine carborundum grinding wheel, run- 

 ning on the dead centres of the lathe, at about 5000 revs, per min- 

 ute. This plan of " truing " the bearings ensured a very close 

 agreement between the thread and bearing axes. ^ The two bear- 

 ings thus ground were then available for correcting the centre 

 holes on which the threads were first cut, so that eventually, 

 centres, bearings and threads were found to be in such close agree- 

 ment that the usual mechanical tests applied to them indicated no 

 aj^preciable error. 



As we are not here concerned with the work intermediate be- 

 tween what has just been described, and certain further correc- 

 tions made at a later i^eriod, and which resulted in the practical 

 elimination of all measurable irregularities between screw, nut and 

 bearings, we may at once proceed to describe how this further 

 imj^rovement was accomplished. 



III. — Final inetJiod of adjustiiig and testing the screw and its 

 bearings. 



It lias already been stated that the ruling macliine was completed 

 up to a certain stage, and a number of trial ruling made under 

 unavoidably adverse conditions. As was anticipated, the results 

 from these early trials were imperfect but valuable in that they 

 served to bring under notice defects of various kinds. Among these 

 one was such as could have been caused either by a bent screw or by 

 a screw whose axis was eccentric with its bearings by amounts too 

 small for the rough method already described to detect. 



After some experiments had been made, the following device 

 for the detection of small errors in the screw or its bearings was 

 designed and constructed, and proved both convenient and effec- 

 tive. An inspection of Fig. 2, Plate XIV., wliich is a plan of the 

 apparatus, will aid the verbal description here given. 



As a sufficiently sensitive test coidd not be directly applied to tlie 

 screw when in position upon tlie ruling machine, both screw and 

 nut were detached and placed in polished steel V-shaped bearings, 

 secured to the carefully worked surface of a heavy slab or surface 

 plate of glass about 1| inches thick. The strict alignment of the 

 V bearings with respect to each other was obtained by first placing 

 upon them a straight round bar of steel accurtely ground to the 

 same diameter as the bearing surface upon the screw. The use of 

 this bar ensured the parallelism of the V block surfaces. 



1 This method of correcting- the liearings assumes the str.iig-htness of tlie screw ; an assumption 

 which afterwards was found to he incorrect. 



