( 284 ) . 
spectrum of the limb a widening of narrow lines, especially on their 
red side, is to be expected. 
Indeed, rays coming from the limb have, as a rule, accomplished 
a longer distance through the solar gases than rays coming from the 
centre, and, therefore, were more subject to loss of intensity by the 
process of incurvation towards the photosphere. The amount of the 
irregular ray-curving depends on the absolute magnitude of R m A, fl , 
which (near the weaker lines of the solar spectrum) is sensibly 
greater with R-light than with V-light. So the lines must chiefly 
widen at their red-facing side, in proportion as the opportunity for 
losing light increases. 
That the character of the limb-spectrum resembles that of the 
spot-spectrum in many respects, is immediately explained by the 
refraction theory. Inded, the light which we receive from the inner 
zones of the penumbra, has followed paths, along which the condi¬ 
tions of limb-rays prevail, (cf. fig. 3). 
On the Other hand, there are numerous peculiarities in the beha¬ 
viour of various Fraunhofer lines, which we cannot so easily explain 
by s i m ply applying the laws of refraction — either because entirely 
different processes are involved, or perhaps because the consequences 
of refraction have to be more deeply scrutinized. (We remind of the 
weakening of the wings of the stronger lines in the limb-spectrum; 
of the weakening of several lines in the spot-spectrum; of the different 
way in which the hydrogen lines behave in the spectra of the limb and 
the spots, etc.). All this requires an ample study of details, in which it is 
necessary to dispose of an extensive collection of observational data. 
Finally I wish to call attention to an inference deduced from our 
conception of the Fraunhofer spectrum, which, to my knowledge, is 
not yet supported by any direct observation. 
Among the various kinds of V-light of each line, one kind must 
exist, for which R m — 0 ( at c ' i n fig. 8). Rays of this and of 
neighbouring wave-lengths suffer less refraction than rays, belonging 
to blank parts of the spectrum; so they must produce strips of 
greater brightness (pseudo-emission lines) in the spectrum of the 
centre of the disk. With strong Fraunhofer lines the point c' lies 
ar away, on a part of the dispersion curve so slightly inclined, that 
the bright band must be vague and inappreciable; but with narrower 
lines the bright companion will also be narrow, and relatively more 
conspicuous. Its distance from the absorption line is always smaller 
than the width of the Fraunhofer line. By applying the extremely 
sensible method of Fabry and Butsson the bright lines may perhaps 
(October 28 , 1909 ). 
