ARTICLES 247 



fine grinding of the spherical surface was commenced. The 

 surface was then polished in order that tests might be made 

 as to the character of the figure at different temperatures. 

 Meanwhile, further attempts were made at St. Gobain to get a 

 perfect disc : a flawless disc of 100 inches aperture and 8 inches 

 thickness was produced, but this was not considered sufficiently 

 thick. The progress of the work on the first disc fortunately 

 soon indicated that it would be suitable for its purpose. 



The working of the mirror was in the very capable hands of 

 Prof. G. W. Ritchey, who had had considerable experience in 

 figuring and testing large mirrors at the Yerkes Observatory, 

 and who had also successfully worked the Mount Wilson 60-inch 

 mirror. Details of the various processes in the construction of 

 a large mirror and of the precautions needed at each stage have 

 been lucidly given by Prof. Ritchey in his paper "On the modern 

 reflecting telescope and the making and testing of optical 

 mirrors " (Smithsonian Contributions, vol. xxxiv.). After rough 

 grinding and edging the mirror was ground to the spherical form 

 with a radius of curvature slightly greater than twice the desired 

 focal length of the finished paraboloid. The mirror was then 

 polished, frequent tests being made during the process to ensure 

 a perfect figure being obtained free from zonal or irregular 

 errors. The parabolising was then done by shortening the radii 

 of curvature of the inner zones, using special polishing tools. 

 The parabolising is a most difficult operation and calls for the 

 utmost skill in working and testing. The delicacy of the opera- 

 tion can be realised from the fact that in the 100-inch mirror 

 the depth of the finished paraboloid differs from that of the 

 spherical surface to which it was brought previous to para- 

 bolising by almost exactly o*ooi inch at the centre, where the 

 difference is a maximum. The work of changing from the 

 spherical surface to a paraboloid therefore involves removing 

 less than one-thousandth of an inch of glass from any part of 

 the surface, yet it occupied over a year, the work being sus- 

 pended for a time during the winter when the artificial heating 

 needed for the optical shop made accurate tests impossible. 

 Tests were made each morning after a day's figuring, special 

 arrangements being provided for lifting and holding the mirror 

 upright for and during testing. 



Two methods of testing were employed : (1) the mirror 

 was tested at the centre of curvature to determine the total 



