GENERAL DISCUSSION OF SPECTROSCOPIC RESULTS. 
471 
tins portion of the layer remains uncovered by the moon for an appreciable time after 
the sky glare is withdrawn at totality, whilst the lower dense strata immediately in 
contact with the photosphere are instantaneously occulted. 
It might reasonably be assumed, therefore, that tlie intensities of the bright lines 
in the lowest strata differ to some extent from those in the spectra jiliotographed, and 
even more closely approximate to the intensities in the Fraunhofer spectrum. 
But the photographs under discussion portray a graziruj contact, in which the 
motion of the moon was not across hut parallel to the Hash layer. These sjiectra, 
therefore, should more truly rejjresent the radiation from the entire depth of the 
layer, at any rate at points near the apex of the bright arcs, and where the layer is 
sufficiently uncovered, because at such points the very lowest strata would remain 
visible throughout the time the plate was exposed."^ 
A careful comiiarison between the intensities of the lines at jioints near to and far 
from the apex, or centre line of the sjiectra, shows, however, that there are no 
appreciable differences. 
Moreover, the intensities given in Table I., which were estimated at points not far 
from the apex, and where the continuous spectrum of the photosphere was just 
beginning to appear, will be found to he in substantial agreement with the results of 
Lockyeu (1898), Frost (IDOO), and Mitchell (FJOl), all of which were obtained near 
the central line. 
It seems, therefore, that there can be no very striking differences lietween the 
spectra of the higher and lower regions of the flash layer as regards the intensities of 
the lines, unless absorption by the upper regions through which the line of sight passes 
should neutralise such differences. In jiarticular it may be noted that the enhanced 
lines seem to predominate throughout the entire region. 
If it is assumed that the differences between spark and arc spectra are conditioned 
by temperature, the spark being the hotter, it would seem at first sight that the flash 
region must have a higher temperature, and must conseipiently be distinct from the 
absorbing layer, since in the latter the intensities of the lines closely ajijiroximate to 
those in the arc. I think it can be shown, however, that the spark and arc conditions 
may co-cxist at the same altitude aljove the photosphere. 
It IS well known that the outer limit of the chromosphere, as seen in the line of 
hydi'ogen, presents a structure of small filaments like blades of grass coveriim the 
entire surface, and very unlike the diflused, indelinite limit which a true atmospheric 
envelope might be expected to present. 
According to Secciii, “at the base of the chromosphere the hydrogen has the shape 
of small, close filaments which seem to correspond with the granulations of the photo¬ 
sphere.” t 
The terms layer and strata are here used for convenience, but it is nob intended to imply that the 
gases of the chromosphere are in reality stratified, 
t ‘ Popular Astronomy,’ S. Newcomb, p. 275. 
