A s i i. v !: i; i: D fj T, A s s r K r, K s r o i- K. 



It is well, on Commencing to polish with a tool made in this way. to warm tlic 

 i;lass as well as the tool in water (page -I) before bringing the two in contact. If 

 this is not done the polishing will not go on kindly, a good adaptation not boin^ 

 secured tor a length of time, and the glass surface heinj injured at the outset. The 

 rosin on a polisher put away for a day or two suffers an internal change, a species 

 of irregular swelling, and does not retain it-- original form. Heating, too, has a 

 good effect in preventing disturbance by local variations of temperature in the glass. 



The description of Local Polishers" will be given under 



d. 



iij l:'s<i iiu'ii i 



' 



1 liave been in the habit of testing mirrors exclusively at the centre of curvature, 

 not putting them in the telescope tube until nearly parabolic or finished. The 

 means of trial are so excellent, the indications obtained so precise, and the freedom 

 from atmospheric disturbances so complete, that the greatest facilities are often d 

 for ascertaining the nature of a surface. In addition the observer is entirely inde- 

 IH-ndent of day or ni^lit. and of the weather. I do not think that anything more 

 is learned of the telescope, even under favorable circumstances, than in the work- 

 shop. For the improvement of these methods of observation. Science is largely 

 indebted to M. Foncault. whose third test the second in the next paragraph is 

 sufficient to ali'ord by itself a large part of the information required in correcting 

 a concave surface. 



There are two distinct modes of examination : 1st, observing with an eye-piece 

 the image of an illuminated pin-hole at the focus, and the cone of rays inside and 

 outside that plane; ',M. receiving the entire pencil of light coming from the mirror 

 through the pupil on the retina, and noticing the distribution of light and shade, 

 and the appearances in relief on the face of the mirror. 



The arrangements for these tests are as follows: Around the flame of a lamp (a, 



Fig. 8. 



Testing a Concave at the Outre of Curvature. 



1 i^'. H) a sheet of tin is bent so as to form a cylindrical sen-en. Through it at the 

 height of the brightest part of the flame, as at /<, two holes are bored, a quarter of an 

 inch apart, one ,'., of an inch in diameter, the other as small as the point of the finest 

 needle will make perhaps ^^ of an inch. This apparatus is to be set at the centre 



