46 
can at least show that the conditions which are most 
favourable to the development of the aurora exist in much 
greater force on the comets than they do on the earth. The 
greatest development of the aurora borealis takes place at 
the equinoxes. There is a cessation in summer, and another 
in winter. Now, the equinoxes are the times when the 
action of the sun on our northern hemisphere is changing 
most rapidly. Hence the condition favourable for the 
aurora is change in the action of the sun. The same thing 
is pointed out by the diurnal variation in the electricity of 
the atmosphere. Now, as has been already shown, the 
change in temperature on the comets is incomparably 
greater than it is on the earth, and its variation corresponds 
with the variation in the splendour of the comet. 
o 
Angstrom has also shown that the light from the aurora, 
the corona, and the zodiacal light, are all of the same 
character, or all give the same bright lines when viewed 
through the spectroscope, and that these lines coi’respond to 
the light from no known substance. This indicates that 
whatever this light may be, the incandescent material is the 
same in all cases; or may we not assume that it is the 
medium which fills space that is illuminated by the electric 
discharges ? This would be supported by the fact that the 
light from the heads of two small comets indicated carbon, 
whereas that from the tails only gave a faint continuous 
spectrum. For an electric discharge would first illuminate 
the atmosphere of the comet, or even carry some of the solid 
material off in a state of vapour, and then pass off to the 
surrounding medium. Thus while the spectrum from the 
head would be that of cometary matter, the tail would be 
due to the incandescent ether. 
I would here suggest that gas, when rendered incandes- 
cent by electricity, may reflect light — it will certainly cast 
a shadow from the eJectric light — and if this be the case, 
part of the light from comets’ tails may after all be reflected 
sunlight. 
