GrRUBB — On an Improved Equatorial Tele&cope. 495 



pushed round by hand, yet when the tuhe is set in position, as it 

 always finally should be, by the button-head % on the shaft between 

 the two wheels, the error will be the difference and not the sum of 

 the " lo&8 " in teeth of each pair of wheels and pinions, and this diffe- 

 rence can be reduced to such a very small amount as to become 

 practically inappreciable. 



Also, with regard to the velocity-ratio, care is taken that when 

 the leaf of one pinion is in any particular phase as regards its 

 wheel, the leaf of the corresponding pinion at the other end of the 

 shaft is in the same phase as regards its wheel ; consequently the 

 error due to this cause does not take effect. It is found practically 

 that these circles work with quite sufficient accuracy for all ordi- 

 nary setting purposes with small instruments. The whole instru- 

 ment is mounted upon a frame which can be any moment changed 

 from an equatorial to an alt-azimuth, or vice verm. When used as 

 an alt-azimuth the circles read altitudes and azimuths instead of 

 ascensions and declinations. 



