66 SCLENCE. 
leaving a space of from 4 to 1 mm, between 
them. The strips should be of about the 
same width as the clear spaces. They are to 
be mounted in two groups on opposite sides 
of the lamp, and the rays passing between 
them produce the polar 
streamers. The proper 
number, width and dis- 
tribution of the strips 
necessary to produce the 
most realistic effect can 
be easily determined by 
experiment.* A circular 
disk of metal, a trifle 
larger than the lamp, should be fastened to 
the tip of the lamp with sealing- wax, or any 
soft, water-resisting cement; this cuts off 
the direct light of the lamp and rep- 
[N.S. Vou. XIII. No. 315. 
rheostat in circuit with the lamp to regulate 
the intensity of the illumination. On turn- 
ing on the current and seating ourselves 
in front of the tank we shall see a most 
beautiful corona, caused by the scattering 
of the light of the lamp by the small par- 
ticles of mastic suspended in the water. 
If we look at it through a Nicol prism we 
shall find that it is radially polarized, a 
dark area appearing on each side of the 
lamp, which turns as we turn the Nicol. 
The illumination is not uniform around the 
lamp, owing to unsymmetrical distribution 
of the candle power, and this heightens the 
effect. Ifthe polar streamers are found to 
be too sharply defined or too wide, the de- 
fect can be easily remedied by altering the 
tin-foil strips. The eclipse is not yet per- 
Fia. 3. 
resents the dark disk of the moon. The 
whole is to be immersed in the tank, with 
the lamp in a horizontal position, and the 
metal disk close against the front glass 
plate. (Fig.2.) Itisa good plan to havea 
* Probably black paint could be used instead of 
the tin-foil strips. 
fect, however, the illumination of the sky 
background being too white and too bril- 
liant in comparison. By adding a solution 
of some bluish-green aniline dye (I used 
malachite green) the sky can be given its 
weird color, and the corona brought out 
much more distinctly. If the proper amount 
