GENERAL DISCUSSION OF SPECTROSCOPIC RESULTS. 
463 
General Discussion of Results. 
The identification of the bright lines in these spectra with the dark lines of the 
Fraunhofer spectrum presents very little difficulty in the case of the strong, or well- 
defined flash lines, and it appears to be generally true that the more reliable the 
values of wave-length obtained in photographs of the flash, the more closely do they 
correspond with Eowland’s values of the dark lines. Thus many of the lines 
measured on small-scale photographs obtained in 1898 show apparent displacements 
considerably greater than the accuracy of the measures seemed to warrant, and which 
rendered many of the identifications doubtful. This is particularly the case in the 
region between 3700 and 3900, where the iron lines especially seemed to be systemati¬ 
cally of smaller wave-length than the corresponding dark lines, whilst the hydrogen 
lines in the same region agreed very closely indeed with their theoretical positions. 
In the present measures, however, in which the scale of the plates is nearly four times 
greater, these displacements are not confirmed, and the same lines are found to agree 
with Eowland’s values within '04 tenth-metre. 
As regards the fainter ill-defined lines and groups there is, of course, considerable 
uncertainty in assigning the particular dark lines of which they are supposed to be 
the reversals, or which lines in a group of dark lines are reversed in the flash. 
It is, however, abundantly clear, from an examination of Table I., that every well- 
defined bright line of the flash (excluding hydrogen and helium lines and the line at 
46857) can be assigned to a dark line of Eowi-and’s table of an intensity exceeding 
2 of his scale. There are no bright lines of even medium strength which occur in 
blank spaces of the solar spectrum where tlie lines are weaker than 0, and only a few 
of the very weakest lines in the table coincide with solar lines vfith an intensity less 
than 2. 
As a corollary to this, it may be stated that in general the greater the intensity of 
a dark line in the solar spectrum, the more probable is its presence as a bright line in 
any given image of the flash, and in the long range of spectrum covered by the 
spectra under discussion, X3500 to X5000, the dark lines of intensities exceeding 7 
are all present as bright lines, except in two or three instances where they are 
obviously obscured by strong hydrogen or calcium lines. 
In the tables of flash-spectrum lines published by Frost and by Mitchell, the 
same general fact is apparent in the large number of identifications made with 
prominent Fraunhofer lines. Professor Frost concludes that “at least 60 per cent, 
(and probably many more) of the stronger dark lines of the solar spectrum are found 
bright in a stratum not exceeding 1" in height above the photosphere.”^ 
It will probably be generally admitted, therefore, that the flash spectrum as 
photographed hitherto is a reversal of the more prominent of the Fraunhofer lines, 
* ‘ Astrophysical Journal,’ vol. XIL, p. 345. 
