458 
MR. J. EVERSHED ON THE SOLAR ECLIPSE OF 1900, MAY 28. 
spectrum arcs being impressed during quite half a minute at the time of greatest 
obscuration (see Plate 2). 
In all the images the continuous spectrum extends from X3500 to X5100, and 
throughout this long range the focus appears to he almost perfect, a striking 
testimony to the good qualities of the reflector as comjDared with a lens. 
Some of the stronger arcs show a difiuseness on the violet side, a defect which has 
been traced to a want of homogeneity in the glass at the base of the 60"^ prism. In 
the ultra-violet region this shading becomes scarcely noticeable, and the deflnition 
here is very fine ; this is no doubt owing to the almost complete absorption of the 
ultra-violet rays in traversing the thickest part of the prisms. 
The four negatives selected for special study, and which are reproduced in Plate 2, 
are from the exposures numbered 9, 10, 11, and 13. 
No. 9 was exposed for 2 seconds, beginning 15 seconds before mid-eclipse. The 
flash spectrum is impressed in a rather narrow rift in the continuous spectrum, 
extending from position angle 140° to 148°, and including a region between 70° and 
77° south latitude. The^ bright arcs crossing the rift are exceedingly narrow thread¬ 
like lines, well defined throughout the spectrum, and are therefore well adapted for 
accurate wave-length determinations. Although the arcs are inclined about 30° 
from the normal to the direction of dispersion, this was found in making the measures 
to detract but very little from the accuracy of a setting. 
In the ultra-violet the Fraunhofer lines are particularly well-defined in this image 
right up to the end of the plate on the continuous spectrum, but between and 
they are obliterated by over-exposure.* The stronger dark lines, and many of the 
weaker ones, are continuous with and run into bright lines in the rift, and in several 
instances the density of the silver deposit is the same in the bright line as it is in the 
dark line, giving the impression that the change from dark to bright is entirely one of 
contrast resulting from the withdrawal of the bright background of continuous 
spectrum. Some of the more intense lines, such as those of titanium at XX 3685, 
3759, and 3761, do not become dark lines on the continuous spectrum, but, being 
more intense than the latter, appear bright even upon the bright background. 
This may be accounted for by the great altitude to which the titanium vapour 
extends, not to its being intrinsically brighter than the photosphere at the limb. For 
the continuous spectrum is produced by what is virtually a slit of extreme fineness, 
defined by the limbs of the sun and moon, and subtending an angle of less than 1", 
whilst the titanium spectrum comes from a stratum, or virtual slit, of 7" or 8" in 
angular width. 
No. 10 was exposed also for 2 seconds at about 10 seconds before mid-eclipse. The 
limb of the moon, advancing eastwards, has covered up the lowest strata of the flash in 
the position where the lines are so well developed in No. 9. There is, however, a 
* In the reproductions the Fraunhofer lines are almost invisible except at the extreme ultra-violet end 
in No. 13. 
