ASTRONOMY: ADAMS AND BURWELL 
129 
is also mdicated by the great strength of the green carbon fluting, and 
by the appearance of a large number of double reversals among the bright 
lines observed. 
This phenomenon of double reversal appears to be essentiall)^ a 
universal characteristic of the lines in the flash spectrum. In appear- 
ance the reversals resemble closely those of the hydrogen lines and 
of the H and K lines of calcium, and like them they are always approxi- 
mately symmetrical. There seems to be no sufficient reason for con- 
cluding that these reversals are not due, as in the laboratory, to the 
radiation of a dense mass of gas through a cooler outer layer of lower 
density. On this basis the presence of numerous double reversals 
in the flash spectrum would indicate the existence of comparatively 
dense gas at the point under observation, and hence a low level in the 
solar atmosphere. 
The wave-lengths of the bright lines in the spectrum of the flash 
taken without an eclipse have been determined with reference to those 
of the dark lines of the Hmb spectrum which are present on the same 
photographs. Accordingly it is possible to judge of the existence 
of any systematic displacement of the bright lines, a result which can- 
not be obtained from eclipse spectra which contain bright lines alone 
and no standards of reference. A comparison of the wave-lengths of 512 
identified lines gives a value: bright lines — dark lines at limb = — 0,002 
angstrom. The preponderance of the negative sign in the case of the 
individual elements makes it fairly probable that this small quantity is 
to be regarded as real. If so, it finds a ready explanation in the slightly 
higher level of the point under observation in the case of the flash spec- 
trum. At this higher level the cause which produces the displace- 
ments of the solar lines at the Hmb toward the red, which is probably 
a pressure effect, in part, at least, would act less strongly, and the bright 
lines would have a position to the violet of the dark lines. 
A comparison of these measures with those by Mitchell on his eclipse 
spectra indicates a marked gain in accuracy for the Mount Wilson 
results. The average deviation of all of the lines common to the two 
sets of observations between the limits X 4800 and X 5880 is as follows: 
Eclipse Results, 0.030 angstrom; Mount Wilson Results, 0.012 angstrom. 
In this comparison blends and double reversals have been omitted. It 
appears, accordingly, that the greater linear scale of the Mount Wilson 
photographs contributes notably to the accuracy of the measurements. 
It is well known that the lines of certain of the elements of high 
atomic weight are extraordinarily prominent in the flash spectrum. 
