GENERAL DISCUSSION OF SPECTROSCOPIC RESULTS. 
45i) 
good but narrow image of the flash at about position angle 137°, or latitude — 63° to 
— 66° E.; and on the west side also there is a fine thread of faint continuous spectrum 
in latitude — 56° W., upon which the flash lines appear as minute dots, like beads on 
a string. About twelve of these dots may be counted between H and K. 
No. 11 spectrum was exposed during 10 seconds near the time of mid-eclipse. 
Judging by the symmetrical distribution of the bands of continuous spectrum on each 
side of the central line of the image, the middle of the exposure must have been 
timed almost at the moment of greatest phase, which appears to have coincided with 
the computed time of mid-eclipse. 
The continuous spectrum in this negative is reduced to nine or ten narrow bands, 
due to indentations in the moon’s limb, and the flash spectrum appears in the form of 
long arcs crossing the bands and extending over the whole of the south-polar region 
of the sun. Most of the arcs cover 80° degrees of the limb, extending from latitude 
— 75° on the east side to latitude — 28° on the west. The sharpest definition is 
along a band at position angle 212° in latitude — 41° on the west side, and this 
portion of the image was selected for measures of wave-length and estimates of 
intensity. 
The bright lines on this negative are more strongly impressed, and can be traced 
further towards the more refrangible end of the spectrum than in any of the other 
images. Some of the Fraunhofer dark lines can still be traced near the end of the 
image in the ultra-violet, crossing the narrow strips of continuous spectrum. These 
lines therefore do not wholly disappear before the last remnants of continuous 
spectrum vanish, hut they become exceedingly faint, and are easily obliterated by 
over-exposure. 
No. 13 spectrum, exposed for 2 seconds about 14 seconds after mid-phase, shows a 
considerable arc of the photosphere uncovered over the south-west limb, and the 
negative is somewhat fogged from the increasing sky illumination. There is, 
however, a good image of the flash spectrum near the middle line of the image in the 
south-east quadrant. The lines are here very short, the flasii layer being exposed in 
a narrow depression of the moon’s limb, but they are well adapted for measurement. 
Methods of Measurement. 
The photographs numbered 9, 11, and 13 were measured with a micrometer 
microscope, lent to me for this purpose by Major Hills, R.E. This instrument has a 
screw of 1 millim. pitch and about 200 mlllims. in length. The head of the screw 
being divided to 100 parts, readings can be made to ’01 millim., and by estimation to 
'001 millim. 
In practice it was found that '01 millim. was about the limit of accuracy attainable 
with the best defined lines. A preliminary set of measures of the sharpest lines over 
the whole length of spectrum photographed was made, to test the accuracy of the 
screw over long runs, a duplicate series of measures being made over the same 
3 N 2 
