OBTAINED AT THE TOTAL SOLAR ECLIPSES OF 1900, 1901 AND 1905. 
405 
Quartz Spectroscope. 
(No. 14.) Contains 20 lines from 4101 to 3685. 
(No. 15.) Contains 100 lines from 4179 to 3613. No continuous spectrum. 
(No. 16.) Contains 160 lines from 4179 to 3572. Beyond this point the definition 
is bad and the spectrum not used. The lines are seen on a continuous spectrum in 
which practically no absorption lines are seen. 
(No. 17.) Contains about 130 lines and extends from 4101 to 3474. The chromo¬ 
spheric lines are short tips by the side of a continuous spectrum which shows 
absorption lines at the violet end, but is much over-exposed in the blue. 
The programme of observations in 1901 was arranged to sujqflement as far as 
possible those made in 1900 with the flint glass spectroscope towards the red and 
with the quartz spectroscope in the ultra-violet. A suggestion of Mr. Davidson’s 
was used to give a longer piece of spectrum in focus, two plates inclined at a suitable 
angle being used instead of one. 
1901, May 17, Sumatra. 
Flint Spectroscope .—-Minimum deviation H r 
(No. 5.) Contains 200 lines from 5670 to 4426 and 70 lines from 4401 to 4077. 
This photograph consists entirely of bright lines with no continuous spectrum. 
(No. 4.) Contains 180 lines from 5658 to 4460 and 60 lines from 4401 to 4101. 
The spectrum is measured on the edge of a continuous spectrum. 
Quartz Spectroscope .—Minimum deviation 3400. Objective focussed for 3400. 
(No. 7.) Contains 20 lines from 3970 to 3685 with no continuous spectrum. 
(No. 6.) Contains 160 lines from 4036 to 3641 with no continuous spectrum. 
(No. 5.) Contains 160 lines from 4042 to 3600. The chromospheric spectrum is 
measured on the two edges of the continuous spectrum. 
(No. 3.) About 100 lines are measured on the sides of the continuous spectrum 
from 4035 to 3510. At about this point the band divides into two parts, the 
continuous spectrum disappears, and the chromospheric spectrum is seen as two series 
of dots or very short lines. About 100 lines are shown between these limits. The 
explanation of this spectrum is that the crescent was cut by the slit in two points, 
each of which furnishes a chromospheric spectrum; these spectra are quite clear and 
distinct from 3500 to 3300 (the quartz objective being focussed for wave-length 3400), 
but nearer the blue end of the spectrum the want of focus of the objective introduces 
the solar spectrum as well. 
