200 
SIR NORMAN LOCKYER AND OTHERS ON THE 
Hale’s"^ values have been assumed. The values of the lines Her to H;)( are those 
computed from the measurements f)f the photograplis : these agree closely vdth those 
deduced from Balmer’s law. The direct measures of the lines for the use of both 
interpolation curve and formula were made by means of an ivory scale divided to y^th 
of an inch. Tlie positions of the fiducial lines for the interpolation formula were 
checked with a stage micrometer i-eading to xoooodk aii inch, but the majoritv of 
the lines were too broad or too faint to permit tlie use of this instrument with satis¬ 
factory results in all cases. The reduction of the wave-lengths of the lines and the 
assignment of origins were, in the region between 3900 and 1), mainlv done by 
Mr. Baxandall. For the region more refrangible than 3900 Mr. Shackletox is 
chiefly responsible. 
Determination of the Wave-lengths of Coronal Rings. 
The wave-lengths of the coronal rings have been determined by direct comparison 
with a photograph of the solar spectrum on the same scale, utilising the well-known 
lines of the chromosphere and prominences for properly superposing the photographs. 
In the case of the coronal rings, the wave-lengths may he aftected in the same way 
as those of the chromospheric arcs, if all the radiations are not photographed quite 
down to the moon’s edge, and an additional slight error may be introduced by the 
movement of tlie moon during the comparatively long exposures required to photo¬ 
graph some of the rings. One cannot be certain that the inner edge of the ring 
corresponds to the same position of the moon’s edge as that adopted in the case 
of the chromospheric arcs used as references, since the chromospheric arcs are bright 
enough to register their total thickness in view, throughout the exposure, v'hile the 
fainter coronal rings might only be shown in those parts having the maximum 
exposure. Thus, in Photograph 2a of the G-inch series, which was ex^iosed for 
50 seconds, the movement of the moon relatively to the sun during the exposure 
■would be corresponding at Hy to a displacement in wave-length of nearly 
two tenth-metres. The measured wave-length of a line will be different, according as 
the inner edge is taken to represent tlie moon’s position at the beginning or end 
of the exposure, the inner edges of the chromospheric arcs at this stage representing 
essentially the moon’s position at the beginning of the exposure. It is probable that 
the inner edge of a photograplied ling will correspond to the positions of the moon at 
some time intermediate between the beginning and end of the exposure, so that the 
error, even in the case of an exposure as long as 50 seconds, would not amount 
to two tenth-metres. Tlie more useful exposures for coronal rings Avere those of 6 to 
8 seconds, and in these the possible errors due to the moon’s motion will be smaller 
than those Avhich may be introduced by imperfect settings on such indistinct appear¬ 
ances as those presented by most of the coronal rings. Hence no attempt has been 
made to make such corrections of the wave-lengths. 
* ‘ Astro-Phys. Jour.,’ vol. 11 (1892), p. 618 
