110 ON THE REFRACTIVE INDEX OF FLUIDS. 
subsequent calculation could not be neglected. An ordinary 
thin glass covering-plate may be conveniently used. 
2d. It is necessary to provide an object, the diameter of 
which is very exactly known—the most suitable is a strip of 
metal coloured white. This strip should be placed beneath, and 
parallel with, the stage of the microscope, and should be so 
arranged on a proper holder that its middle point shall coin- 
cide precisely with the prolongation of the optical axis of the 
instrument. 
3d. It is necessary to know, as nearly as possible, the dis- 
tance between the upper surface of the object and the air-b>ll. 
In my investigations, I prefer a fixed distance of 100 milli- 
metres, on account of the convenience of this round number 
in calculation. The construction of most microscopes also 
renders this aconvenient distance. Between the distance and 
the diameter of the object a certain ratio should be observed. 
If the latter be more than one fifth of the former a correction 
of the final result becomes necessary, for in consequence of 
the excessive obliquity of the rays proceeding from the 
margins of the object, the difference between their angles of 
incidence and of refraction becomes too sensible to be neg- 
lected. 
4th. The microscope being so arranged that the object is 
brought distinctly into view, the diameter of the air-bell and 
of the image of the object below it are to be successively 
measured; and in doing so it will be of course necessary to 
alter the focus of the instrument slightly, the margins of the 
air-bell and the image lying in different planes. 
As the accuracy of the result in great measure depends 
upon these two measurements, it 1s scarcely possible to bestow 
too much care in taking them. For the methods to be 
followed in this stage of the observation, I refer to the 
chapter on Micrometry.* I must not neglect to add that 
these measurements should be made by reflected light—if 
transmitted light be used, the influence of diffraction causes 
the results to be somewhat too small. It is also advisable 
that the strip of metal used as an object should be of a white 
colour. 
It is essential that the successive measurements of air-bell 
and image should be made rapidly, both in order to obviate 
the influence of changes of temperature, and because the 
gradual absorption of air by most fluids, and especially by 
* In the ‘Monthly Journal of Medical Science,’ May, 1852, p. 453, a 
very full abstract of this chapter will be found. The most exact methods 
are those in which the screw-micrometer eye-piece or the plan of “ double 
vision” are used. (T.) 
