160 Dr Brewstor^s DescfvptiorK^u D^ihle hnage MicTQmeieri 
images^ we may preserve on two sides the natural surface of the 
cleavage j which I have often found to exceed the finest polish 
that can be given by art. 
When the prism is constructed, it may be made to form part 
of the lens of the eye-piece next the eye, by cementing it to the 
outer surface of that lens; or it may be placed between that 
lens and the eye, as was done by Rochon and Dr Pearson, 
where it will give a double image of all objects seen through 
the telescope. The angular distance of these images being in- 
variable for any given position of the eye-piece, they may be 
brought into contact by a motion of* the eye-piece, to or from 
the great mirror, according as they were previously overlapped 
or separated. 
Although we have proposed to follow Rochon in producing 
the double image by a doubly refracting crystal, yet there are 
other means of accomplishing this, which are well worth the at- 
tention of the practical astronomer. 
1. The double image may be produced by a small bisected 
plane speculum, placed between the eye-lens and the eye, and 
one of the halves may be made to move by a screw, not for the 
purpose of bringing the images in contact, but in order to vary 
the constant angle, according as it is wanted for large or small 
discs. 
% The duplication of the image may be effected by bisecting 
the eye-lens, or by placing a bisected lens between the eye-lens 
and the eye. We have now before us a bisected achromatic 
lens, madeby Tulley, which produces the two images with very 
unusual distinctness. 
3. The two images may be formed by means of a slightly in- 
clined face, ground upon a highly polished and parallel plate of 
Fluor-spar ; one image being seen by half of the pupil, through 
the parallel plate, and the other through the inclined face. 
Fluor-spar is recommended, as producing the least dispersion 
under a given angle of deviation ; but though the colour is per- 
ceptible even at small deviations, yet, we are persuaded, from 
experiment, that it will not occasion any sensible error in the re- 
sults ; and even this imperfection may be removed by the ordi- 
uai’Y means. 
