TOTAL ECLIPSE OE THE SUN", APRIL 16, 1S93. 
605 
1871, bat he does not seem to have attached any special importance to it. It may 
be that the feeble Fraunhofer spectrum is under these conditions lost in the bright 
continuous spectrum of the lower corona ; or, since the bright arcs due to hydrogen 
and calcium are strong enough to be seen superposed on the continuous spectrum of 
the chromosphere, it is conceivable that the Fraunhofer lines are masked by a com 
plicated radiation spectrum from the base of the ^chromosphere such as would be 
furnished by the supposed reversing layer. 
To produce the latter result, however, it is not necessary that every dark arc 
should have its counterpart among the bright ones due to the chromosphere, for the 
reason that during the exposure of this photograph the chromosphere was broader 
than the photosphere. The breadth of the chromosphere was about 8", while 
that of the photosphere in its widest part varied from l" to 3" ; thus some of the 
chromospheric arcs would be broad enough to mask several of the close dark ones 
in their vicinities, and the absence of the Fraunhofer lines in this photograph there¬ 
fore does not necessarily favour the existence of a “ reversing layer.” 
A consideration of the photographs disposes of any suggestion which may be made 
as to the possible production of an apparently continuous spectrum by the crowding 
together of a great number of short bright arcs, in the spectrum of the cusp, pro¬ 
ceeding from the very lowest part of the chromosphere. The actual appearance of 
the African Photograph No. 22, Plate 13, did at first suggest that such was the case, 
the continuous spectrum of the photosphere having an excess of brightness 
immediately at the cusp where it joins the bright arcs. In the succeeding photo¬ 
graphs, however, there is no such brighter band of continuous spectrum at the cusp, 
and the band which appears in Photograph 22 is seen to retain the same position as 
referred to the images of the prominences. 
It seems probable, therefore, that the bright continuous band at the cusp, in 
No. 22, is due to an indentation of the moon’s limb at the point where it was 
produced and that its coincidence with the cusp in the photograph is accidental. 
In fig. 16, the bright arcs recorded in the African Photograph No. 22 are drawn 
in juxtaposition with a photographic map of a portion of the solar spectrum with 
lengths proportional to their intensities. 
It becomes evident at once that the radiation spectrum is most distinctly not 
identical with the Fraunhofer spectrum ; the most important point is that some of 
the strongest bright lines do not appear among the dark ones in the solar spectrum, 
while some of the strongest dark lines are not seen bright in the spectrum of the 
stratum of vapours in immediate contact with the photosphere. The region covered 
by the diagram lies between wave-lengths 4100 and 4300, but similar results follow 
when other regions are included in the inquiry. 
These positive conclusions are not weakened by the consideration that the resolving 
power of the prismatic cameras employed is not sufficiently great to show all the 
lines of the Fraunhofer spectrum, which is used as a term of comparison ; in fact, 
