124 
PHYSICS: C. BARUS 
Data similar to the above may now be inserted; viz., for a small cylinder of 
water (to be used in the experiments below) 
R = S cm.; /z = 1.33; co = 628; c = 3 X lO^o 
whence the path difference 1.82 X 10~^ cm. results. 
This, curiously enough, is about the same value which was obtained in 
case of equation (11), so that identical deductions apply. The conditions are 
somewhat more favorable for larger values of ju. Thus in the hmiting case 
ix^ = 3, the path difference would be about doubled. 
5. Experiments. — To carry out these experiments at the present time is of 
course out of the question; but a number of contributory observations may be 
made with advantage. The case of figure 2 is similar to figure 1, where the 
dispersion of the cyhnder G in the former case is simulated by the prism P 
and the auxihary mirrors w, m' , of the latter. If the slit of the collimator at 
L is not too coarse, two reversed spectra will be seen in the telescope at T, 
which on being superposed by rotating w or TV on a vertical axis, will show a 
vivid linear phenomenon in the line of symmetry of the two superposed 
spectra. On rotating w or TV on a horizontal axis, the distance apart of the 
fringe dots along this line may be given any reasonable value, at pleasure. 
These displacements are at once referred to the definite wave length in which 
the linear phenomenon is put. The dispersion of the prism has no bearing 
on the clearness of the phenomenon: 30° and 60° prism were tested with like 
results. 
To obtain the achromatics and increased luminosity in the spectrum fringes 
(now to be horizontal bands throughout from red to blue), the rays of the spec- 
trum will have to be reassembled and that may be done by inserting a second 
prism say P', between L and N' , in a way to counteract the effect of the first. 
If the achromatics are to be obtained, the glass paths of the two rays in P 
and P\ respectively, must be coincident. Hence, the axis of the collimator 
at L must be inclined to accommodate the angle of minimum deviation of the 
identical prisms P, P'; and while TV and m are parallel, N' and m' normal to 
each other, L and T have their axes symmetric to N. The adjustment is not 
difficult as they need not be perfect to secure good achromatics; but if it is 
not made the fringes are numerous, colored, and unsatisfactory. 
The experiments, figure 2, differ from the case figure 1, because the rays 
are parallel in the former case and condensed to a caustic by the eccentric 
refraction of the cylinder in the latter. Hence with these a short range tele- 
scope with strong objective is necessary; but as has been stated, the lines of 
the solar spectrum nevertheless come out clearly. Experiments were there- 
fore made by simulating the glass cylinder GG by a thin cylindrical glass 
shell, closed below and above and containing a solution of mercury potassic 
iodide with an index at pleasure between 1.5 and 1.7. It was not difficult to 
meet the conditions of figure 1 so far as mere refraction is concerned, and 
certain incidental results obtained in this work have been given elsewhere. 
