198 
SIR NORMAN LOCKYER AND OTHERS ON THE 
enlargement in Plate 8. Two other spectra taken with longer exposure nearer mid¬ 
eclipse are reproduced in Plate 7, namely, photographs '2a and Ah. 
PiEDUCTIOX OF THE PHOTOGRAPHS. 
The Determination of Wave-lenyths of Chromosj^heric Lines. 
When we come to determine the wave-lengths of lines in a spectrum of the 
eclipsed sun, taken Ijy means of a prismatic camera, it is found that the problem is not 
c[uite the same as in the case of the slit spectroscope, in which care must always be 
taken oidy to utilise the centre of each of the lines for micrometric measurements, no 
matter how broad they may be. 
In the prismatic camera pictures, in which arcs take the place of lines, not oidy 
must great care be taken to measure the distance between them in the direction of 
dispersion, hut these measurements must always be made from the side representing 
the limb of the moon. Now the c[uestion has arisen, does the inner portion of these 
arcs in all cases represent the limb of the dark moon '■ If it does, then all measure¬ 
ments for wave-length made from it would be correct. 
I have shown previously"^ that there may exist arcs which would be brighter some 
distance from the moon’s limb, and these would decrease in intensity as this limb 
was approached, such an arc being due to a shell of vapour concentric with the photo¬ 
sphere, but some distance from it. It is quite easy to understand that a photograph 
of such an arc, especially if it were under-exposed, would give us a record of the 
brightest part of the arc in the line of sight, but this fainter portion extending down 
to tlie dark moon would be missing. 
In such a case the measurements for wave-length made from the inner edge would 
not then be justifiable, and if the arc in question were of unknown origin, its deduced 
wave-length woidd be, and the suggested origin might Ije, erroneous. Theoretically 
all arcs which represent layers high up above the photosphere should be more intense 
where the line of sight of the observer meets the layer in t[uestion tangentially. Thus 
arcs that represent layers which are lowest should be brightest nearer the moon’s 
limb, while the opposite should be the case with those at higher levels. 
It appears, however, from the photographs taken during the last three eclipses that 
such a gradation of light has not been recorded, but this is owing probably to the 
small size of the solar image used. 
An idea of the sizes in the 6-inch and 9-inch prismatic cameras, and the thickness 
of some of the arcs, may be of interest in this place. In these two cameras the solar 
images were 0'85 and 1'12 inches in diameter respectively. Considering a layer at 
the sun to extend 5000 miles above the photosphere, this is equivalent at the mean 
distance of the sun to IIT seconds of arc ; taking the sun’s diameter as subteiiding 
an aimle of 32' 3"'6, then tlie ratio of this laver to the diameter is as ll'l to 1923‘6, 
* ‘Roy. Soe. Proc.,’ vol. 187, p. 587. 
