326 Mr, mudge on the Confiruclion 
infomuch, that half or two-thirds of it would be com- 
pletely polifhed when the circumference of it was fcarcely 
touched by the tool. Obferving this in fome of the firft 
metals, and not confidering that this way of polifhing 
was in faff a fpecies of grinding, and as perfect as that 
upon the hones, I went on reluctantly with the work, 
almoft defpairing of being able to produce a good 
figure. However, I always found myfelf agreeably 
deceived; for when the polifh was extended to the 
edge, or within the tenth of an inch of it, I almoft 
conftantly found the figure good, and the performance 
of the metal very diftinCt. But this fame cficumltance of 
apparent defeCt in the metals, was in faCt that to which 
their perfection was owing; for they all, contrary to my 
expectation, turned out parabolic. However I did not 
for a great while know any certain way of giving that 
degree of parabolic tendency which was juft neceflary, 
and which will be defcribed hereafter. It was a long 
time before I got rid of my prejudice againft this appa- 
rent imperfection in the procefs,. or could reconcile my- 
felf to the irregular manner in which the polifh pro- 
ceeded ; for I looked upon it as a certain fource of error, 
and notwithftanding I faw it eventually fucceed, yet 
whenever I chanced to find that a metal, when firft ap- 
plied to the polifher, took the polifh equally all over, 
and 
