ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29, 188G. 
327 
scattering of minute particles, only the smaller part can be owing to this cause, as 
otherwise the maximum of actinic intensity in the spectrum would he displaced 
towards the violet and not towards the red end of the spectrum. The faintness of 
the Fraunhofer lines is further evidence in the same direction, and also makes it 
probable that there is not much matter round the Sun sufficiently large to scatter 
Sun light without polarising it. 
If we now turn to the lines shown on the photographs, our attention is at once 
arrested by the H ajid K lines. They form in this, as in previous eclipses, the most 
intense feature of the corona. In the Eclipse of 1882, these lines were so bright that 
the atmosphere scattered enough of their light to make them appear over the disc of 
the Moon and at a considerable distance outside the corona. In the Eclipse of 1883, 
the same lines appeared reversed over the lunar disc; on that occasion, therefore, the 
light scattered by the atmosphere in front of the Moon must have been derived from 
some source showing absorption lines. In our photographs the same lines end sharply 
with the corona, and we must conclude, therefore, that, in spite of the unfavourable 
atmospheric conditions, there was but little light scattered by our own atmosphere in 
the neighbourhood of the Sun. This is confirmed by the photograph of the corona 
itself; for, while on previous occasions the sky light formed a bright background, on 
which the shadow of the wire stretched across the camera showed distinctly through¬ 
out the whole plate, the two needles do not show at all on our plates, even on those 
which had the longest exposure. The analysis of the light seen in front of the Moon, 
or in the neighbourhood of the corona, seems to give us imjDortant information on the 
general light emitted by it, for in three successive eclipses bright lines, dank lines, and 
no lines at all ajapeared on the photograph. 
The prominence lines on the plate present a curious winged appearance towards the 
violet. At first sight it might seem that this is due to some defect in the spectro¬ 
scope, such as a reflection on the inside of the slit; but, after careful consideration, I 
do not think that this is possible. One edge of the wing is perfectly sharp, and 
almost as black as the prominence line itself. The H and K lines in the corona seem 
perfectly sharp along their whole length, and do not show anywhere a trace of fuzzi¬ 
ness such as would be produced by a defective slit. The reference spectrum shows 
the finest lines with perfect definition. The wing appears in all prominence lines, 
and, if it is not due to an instrumental defect, must be owing to a rapid motion of 
the matter forming the prominence. I find that an approaching motion of about 247 
miles per second in part of the prominence would account for the displacement. 
According to Young, such motions have been observed, though seldom. The limb of 
the Sun on which the prominence appears vras receding. 
Another fact of some importance is this, that, while the continuous spectrum on the 
northern hemisphere of the Sun is weaker than on the southern, the H and K lines 
are stronger. The hydrogen lines Hy and A do not appear at all on the southern 
side, while I can, under favourable conditions, trace Hy to some distance on the 
