76 H. J. Grayson: 



machine, it is not i^roposed to enter upon a detailed description, 

 but merely to specify the principal dimensions and essential 

 features. Tlie central hub and the flange for supporting the steel 

 rim -were turned from a single casting of hard jrun metal ; the hub 

 being about 1| in. through and made to fit the corresponding 

 tapered bearing upon the screw. The flange of the hub was about 

 8 inches in diameter and formed the bed and support of the circular 

 st-eel rim. The over-all diameter of the rim was 10 inches and its 

 finished thickness just under ^ inch. This steel rim, which was 

 turned, annealed and afterwards ground and lapped with great 

 care, was accurately fitted and bolted to its gun metal support with- 

 out avoidable stress of any Sort, and with only just sufficient freedom 

 to accommodate for any difference in expansion. 



Two similar heads were thus constructed; in one 360 and 

 in tlie other 540 teeth, were eventually cut. The cutting of these 

 teeth with the necessary accuracy was a formidable undertaking, 

 and occupied a long time. As no milling machine possessing even 

 approximately the accuracy required for cutting the teeth was 

 available, it was decided to grind them out, a method which, 

 •although slow and laboriovis, promised to afford accuracy of a fairly 

 high order. The first requisite for the method proposed to be 

 followed was a well divided circle. Therefore, with a view to 

 securing one sufficiently correct, a number of theodolite and other 

 circles were examined and tested as to their correctnes; but all 

 failed in tliis respect. Hence it was decided to have a circle con- 

 structed and specially graduated. Messrs. Watts and Son, of Lon- 

 don, who had recently built a very accurate circular dividing 

 engine, were communicated with, and they undertook the work 

 of making and graduating a suitable disc of silver. On this circle 

 ■each of the 360° was divided into 10 minvite intervals with gradua- 

 tion lines sufficiently fine to bear a magnification of 100 diameters. 

 The completed circle fully met expectations, its accuracy being well 

 within the limits stipulated for, viz.,± 2.5 seconds. Indeed, after a 

 series of tests we found its accuracy comparable with that of a circle 

 ruled by the same engine under approximately the same conditions 

 and tested by the Imperial Institiite, Charlottenberg. The maximum 

 error of any sort found in the latter were -1-1.4 seconds and —1.7 

 seconds. 



The provision to be made for the actual operation of grinding 

 the teeth was next considered and was mainly worked out as fol- 

 lows : — A number of specially thin dental wheels, 3 inches in 



