Notes on the process of Polishing and Figuring 
18-in. Glass Specula by hand, and experiments 
with Flat Surfaces. 
By H. F. Manpsen, 
[Read before the Royal Society of N.S.W., 7 July, 1886.] 
So much has been written upon the production of glass reflectors 
for use in astronomical ee and so many of these 
being in use at the present time, any improvements in their con- 
struction would be difficult if not impossible to attempt ; still, a 
the method which I have followed is not altogether the astial fae 
and as I am not aware that any mirrors of the size under dis- 
cussion have sib St oa by hand, I have thought it probable 
that certain notes taken down by me during my experiments 
might not be oe void of interest to some of the members 
of this Society. 
It is now more than four years ago since I first began polishing 
specula-flats, &e., with other optical experiments. During this 
period several mirrors from 7-in. to 18-in. diameter have been 
completed with gradually increasing success in the result. As the 
rough castings for the 18-in. mirro , these 
have been refined and repolished several times to gain practical 
information in their construction. They were imported from 
Chance Bros., Birmingham, and when polished were found to 
have been well annealed. 
A piece of plate-glass 10-in. diameter was cemented to the back 
of each mirror to suit its intended cell or mounting, and the bh a 
of the whole speculum when finished was about 70 Ib. (fig. ij. 
In producing these specula the first thing to consider is natu- 
rally the convex tool with which they are ground to the proper 
ane and my first attempt was made by procuring two flat 
dises of glass of the same size, and grinding them together with 
emery and sand, the intended speculum occupying the uppermost 
position until they had attained the desired form; it being well 
known that two flat discs when ground together will form them- 
selves into spherical surfaces, the ‘overhangin g part of the top one 
producing convexity in the one undernea 
My succeeding trial was made by the usual iron tool turned to 
las eusimnate curvature in a lathe; but it was found that two of 
these had to be ground together for a considerable time before 
