80 PROCESS OF POLISHING AND FIGURING 18-IN. GLASS 
they became of a true spherical form ; and knowing that the co- 
efficient of expansion of iron was 0.000012, that of glass only 
0.000008, and that in both cases the surfaces were very sensitive 
to small variations in temperature (as illustrated hereafter by my flat 
surfaces), I formed the opinion that a truer surface could be pro- 
duced by having the tool made of the same material as the intended 
reflector, and therefore in producing the two larger specula under 
consideration I reverted to my former process. 
Three plates of rough glass l-in, thick and 18}-in. diameter 
were ground together to fit one another, and then cemented so as 
to form a firm and solid block. 
To produce the proper convexity in so large a size in the usual 
manner would occupy a considerable time, and require a great 
amount of labour ; yet such a form would be easily given to it by 
the machinery used by plate-glass grinders. I had the tool made 
‘of about one quarter more convexity than the required concavity of 
the speculum, and the latter having being partly hollowed out by 
the use of a leaden weight and rough emery, the two (tool and 
glass) having been ground together soon formed themselves into 
perfect spherical surfaces, and very nearly of the proper curvature. 
Where many specula are to be produced upon the same tool, 
ron may be preferable on account of its surface becoming extremely 
hard ; but I have with one of glass ground or rather refined the 
speculum during three hours, without producing so much as 1-in. 
alteration in the focal length. 
During the course of my earlier attempts with smaller glasses, 
much time was lost in the polishing, the smoothing process no 
having been sufficiently carried out with the finest grade of emery, 
but with the two under consideration to-night, this was guarded 
against, and no polishing was begun until the mirror was bright 
‘enough to reflect the image of the sun at an almost perpendicular 
incidence. 
succeed, but with 5m. polishing it reflected the light from a very 
small pinhole, and when thus examined I was delighted to find 
that it presented a true and regular surface, slightly inclining from 
the spherical towards the spheroidal form, and that irregularities 
of more than ss}55 part of an inch (probably much less) had 
‘no existence. 
_ This proved beyond doubt that the class of tool I had used left 
nothing to be desired. The speculum having, then, at this 
he process an absolutely true curve, the polishing was pro- 
ceeded with in the usual manner with rouge upon pitch (the 
speculum being uppermost), so well described by the late Dr. 
