AXD A SELECTION EEOM THE OBSEEVATIONS MADE WITH THEM. 
693 
eccentric; small variations in the radius of curvature are thus produced with great 
facility: where, however, there has been a considerable departure from the length 
of stroke necessary to produce a spherical figure, the speculum requires to be ground 
for twelve houi-s, or perhaps much longer, with the proper motions, before it is fit to be 
pohshed. Mr. Whitwoeth is of opinion that greater general accuracy of surface is 
obtainable by scraping than by grinding : the late Mr. A. Ross, as high an authority as 
any one in everything relating to practical optics, held very nearly the same language*. 
He attributed the defective action of the grinding-process to the unequal distribution of 
the grinding-powder, which, accumulating at the centre by capillary attraction, and at 
the edges by mechanical action, unduly shortened the radius of curvature at the centre 
and lengthened it at the edge. He employed the grinding-powder dry in producing 
flat glass surfaces, and believed he thus obtained a better result. It appears to me that 
by cutting up one of the surfaces into minute squares, in the way we have so long prac- 
tised, the causes of unequal action are eliminated. The subject is a very important one, 
as there appears to be no other probable means of working solid materials into accurate 
surfaces for optical purposes than by some modification of the ordinary process of grind- 
ing and polishing. In the ‘ Philosophical Transactions ’ for 1840, there is a sketch of a 
grinding-tool such as we employed in the construction of 3-feet specula ; but I have 
scarcely noticed experiments on grinding, passing at once to the more important subject 
(as it appeared to me at the time), that of polishing. Something useful, however, may 
perhaps be extracted from our record of very numerous experiments on grinding plane 
and curved surfaces. In my very early experiments, the ordinary process for procuring 
a tine plane by grinding three planes in alternate pairs was often repeated. Till we 
adopted the expedient of cutting up two of the surfaces into minute squares, our success 
was very limited. That derice apparently removed all the difficulty, and we were then 
enabled to make large flat mirrors which bore the optical test well in every part, which 
was not the case before. When one of the surfaces is divided into minute portions, with 
sufficiently large and deep intervals, there can be no capillary action such as that 
described by Mr. A. Ross ; neither can there be an accumulation of grinding-powder 
anywhere, because an immediate escape for it is provided ; and if the grinding-powder is 
employed in very small quantities, and no addition is made to it for three or four hours 
before the termination of the process, there will be a high degree of comminution, and 
only just a sufficient number of minute particles to keep up the abrading action, pro- 
bably nowhere more than a single layer. We may form some idea of the accuracy thus 
attainable by examining with a microscope the particles of emery so comminuted. No 
parts can be acted upon strongly except where they deviate from a spherical surface ; 
too violent contact then, and consequently a destructive action, is prevented by the 
moisture interposed. Under such circumstances, with unyielding surfaces, time obviously 
cannot enter as a disturbing element, because there is no abrasion when there is no close 
contact. 
* HoLTZAPFrEL, Meclianical Manipulatioii, vol. iii. p. 1229. * 
