406 PROF. F. W. DYSON: DETERMINATIONS OF WAVE-LENGTH FROM SPECTRA 
The measurement of the spectra in which there was no continuous spectrum 
presented no difficulties, and, generally speaking, where there was a continuous 
spectrum the bright lines were sufficiently strong across it and showed tips at the 
edges to make their identification and measurement quite certain. In cases where 
the absorption lines in the solar spectrum are very strong, the corresponding chromo¬ 
spheric lines were to some extent weakened and uncertain in parts of the spectrum 
for which the objective was not in focus. Some of these lines, e.g., 4063'76 Fe 20 
(in O), 407 1 *91 Fe 15 (in ©), are not shown as strongly as in Sir Norman Lockyer's 
and other spectra, where an objective prism was used. 
2. The spectra were measured by comparison with a glass scale whose division 
errors had been previously determined. Wave-lengths were determined from each 
photograph, in some cases by interpolation from a table and in others by Hartmann’s 
formula. It seemed impracticable to give all the results from the separate photo¬ 
graphs, and the means only are given. The number of photographs on which a line 
has been measured is given as well as the intensity. The scale of intensities must be 
considered as applying to lines in neighbouring parts of the spectrum.. 
The following table, extending from the titanium line at 3685 to the end of the 
series of hydrogen lines, is given as a specimen of the results of the separate photo¬ 
graphs. The numbers at the head of the columns refer to the separate photographs 
taken with the quartz spectroscope :—- 
