IN CONNEXION WITH THE SPECTRUM OF THE SUN. 
5G3 
II. REFRACTION MAP. 
A. Instruments, &c. 
The same general arrangement of the spectroscope, electric lamp, and lenses figured 
in the third of the present series of memoirs has been employed for obtaining the 
photographs used in the construction of the refraction map. 
The spectroscope used is one constructed on the model of Bunsen and Kirchhoff, 
by Schmidt and Hensch, of Berlin. It is provided witli a train of four flint glass 
prisms, viz., three of 45° and one of 60°. 
The prisms are levelled by Kirchhoff’s method, and adjusted as nearly as possible 
for the minimum deviation of the centre of the section being worked upon. 
The camera employed is provided with a simple quartz lens of 5-feet focus. The 
image of the spectrum formed by this lens falls on a sensitized glass plate, lG'Gx 5'8 
centims. The dispersion obtained by this arrangement gives a photographic impres¬ 
sion, in which the distance from G to K (using Cornu’s nomenclature) is about 
11 centims. and from H to K about 1 '3 centims. 
The focus is first fixed approximately by viewing the image of the spectrum on 
the ground glass focussing screen by means of a positive eye piece, and the final 
adjustment is determined by a series of trial plates, using the solar spectrum when 
possible, or, in the absence of the sun, the spectrum of some metal such as cerium, 
which contains an immense number of lines throughout its entire length. For fine 
definition it is requisite that if the beam of light which falls on the slit is not parallel, 
the light of the wave-length to be photographed should be brought to focus on it. 
B. Method of Mapping. 
a. Construction of Interpolation Curves. 
The relative positions of all the most prominent lines visible in an enlargement of 
the photographs employed were first laid down on a strip of paper and then transferred 
to the horizontal line ruled at the base of some curve paper to furnish the ordinates. 
The lines thus selected were referred to Cornu’s map, and their wavedengths fur¬ 
nished the abscissae. A wave-length scale was marked off on the vertical line, 4 millims. 
of which represented 1 millim. of Cornu’s map. In the case of lines assigned to 
particular metals by former observers, a photograph of the metallic and solar spectra 
confronted was found useful in this work of identification. The wave-lengths of the 
selected lines having been found, vertical lines are carried up from them, and horizontal 
lines marked across from the corresponding wave-lengths. Where the vertical and 
horizontal lines meet, a series of points is obtained, through which is drawn a curve 
with as much regularity as possible. The drawing of this curve at once reveals errors 
of identification or of wave-length in the map used. 
4 d 2 
