TOTAL ECLIPSE OF THE SUN, JANUAEY 22, 1898. 
199 
which is as 1 to 173-2. The image of the sun in the 6-inch prismatic camera being 
0-85 inch, this layer would be represented at the moment of totality by an arc equal 
in thickness to •0045 inch, or nearly one two-hundredth of an inch. In tlie 9-inch 
prismatic camera its breadth would be '006 inch. In the case of the 6-inch spectra 
a displacement of of an inch corresponds at Hy to a shift of 1-2 tenth-metres, 
while in the case of the 9-inch spectra the corresj^onding shift would be smaller. As 
the majority of the chromosphere vapours do not reach so high as 1000 miles above 
the photosphere, as indicated by the lengths of the arcs, the possible displacements due 
to the above cause are generally too small to be taken into account. Distinct displace¬ 
ments, however, have been noted in some of the arcs ol helium and carbon, and in 
these cases the known wave-lengths have been adopted. 
Another possibility of error is introduced in the case ol the brightest arcs by 
irradiation. The depth of the hydrogen stratum, for example, as indicated by the 
thickness of F(II^) arc is in some cases nearly three times as great as that indicated 
by the length of the arc. Accordingly if the arc be supposed to have expanded 
equally in both directions, the true position of the moon’s edge will be nearly one-third 
the thickness of the arc from its inner edge. For this reason the hydrogen arcs 
HyS, Hy, HS, He, as well as the K line of calcium, have not been used as fiducial lines 
in reducing the chromospheric spectra. The ultra-violet series of hydrogen lines, 
however, are much finer, and have been used for plotting the curve for the reduction 
of that part of the spectrum. 
The actual procedure was as follows 
For the purpose of reduction the spectrum has been divided into two parts, one 
extending from Ha in the red to X 3900 in the violet, and the other from X 3900 to the 
limit of the photographic impressions in the ultra-violet, about X 3663. 
In the first part the wave-lengths were determined by direct comparison with a 
photograph of the solar spectrum taken with the same instrument provided with a slit 
and collimator, the latter being a 4-inch Cooke triplet of 6 feet focal length. 
For the ultra-violet part of the spectrum, this direct method could not be used, as 
in consequence of the absorption of ultra-violet rays by the dense prisms, satisfactory 
photographs of the solar spectrum beyond X 3900 could not be obtained. Some 
of the photographs of metallic arc spectra, however, show a considerable number of 
lines in this part of the spectrum, and these have been used to confirm wave-lengths 
determined in other ways. In the first instance, the wave-lengths were determined 
by the use of an interpolation curve, adopting well-known lines of hydrogen, titanium, 
magnesium, and iron as fiducial points. Besides this. Dr. Hartmann’s interpolation 
formula for prismatic sjDectra"^ lias been employed for the reduction of the ultra-violet 
part of the spectrum. The wave-lengths adopted for the ultra-violet hydrogen 
lines H^ to Hy are those of Ames,! from H^ to Hp. 
‘ Publicatioueii des Astrophysikalischen Observatoriums zu Potsdam,’ No. 42, vol. 12, 1898. 
t ‘Phil. Mag.,’ vol. 30, 1890, p. 33. 
