578 
MR. J. NORMAN LOCKYER ON THE 
spectrum, which is of increased width in each case, shows the longitudinal dark 
streaks due to Baily’s heads, and they both show bright chromospheric arcs crossing 
the continuous spectrum, as well as bright arcs at the cusp. In No. 23, however, 
distinct traces of the strongest Fraunhofer lines are seen, especially in the ultra¬ 
violet, and they are still more plainly marked in No. 24. The bright arcs at H and 
K, H v , &c., are not quite coincident with the corresponding dark ones, but lie on the 
outer edges, as might be expected. The number of bright arcs at the cusp in No. 23 
is very much smaller than in No. 22, and in No. 24 it is smaller still. The con¬ 
tinuous spectrum of the corona is faintly impressed at the eastern edge, but 1474 K 
and other coronal radiations are not seen. 
In passing outwards from totality, the bright lines at the cusp become reduced in 
number and intensity, while the Fraunhofer spectrum becomes more and more 
distinct (see Photograph No. 25, Plate 13)* up to a certain limit; after which the 
dark arcs become broad and ill defined. 
V. Description of the Brazilian Photographs. 
Photographs near Beginning of Totality. 
Photograph No. 2, which is reproduced in Plate 12, is as near as possible the 
commencement of totality. It shows rings of chromosphere and prominences in 
positions corresponding to H and K, H /3 , H y; , II 5 , H ? . Near the top of the spectrum 
is a strip much brighter than the other parts of the spectrum ; it corresponds with 
the central part of the chromospheric arc visible when the plate was exposed. 
Projecting out of this strip are many bright arcs, less refrangible than F (H^) and 
more refrangible than K ; between these two wave-lengths, however, in consequence 
of the extreme brightness of the chromosphere the photograph is considerably over¬ 
exposed, and the bright arcs are masked. As the moon advanced and covered the 
chromosphere, other photographs were taken which show that it was obscured at the 
fifth photograph ; hence, knowing the interval of time between these two photo¬ 
graphs (from table = 25 seconds) and knowing the relative motion of the moon 
(•253" per second), this gives 2,850 miles as the depth of the chromosphere, a very 
close agreement with that calculated from the African photographs. Photographs 
3, 4, and 5 are very similar to the corresponding African photographs, but they do 
not show so much detail. 
The spectrum of the corona is represented in Photographs 2, 3, 4, and 5, by 
am apparently continuous spectrum, which by reason of the projection of the broad 
ring is brightest along the top edge of the spectrum. The maximum intensity is 
about X 448, and at this point the continuous spectrum can be traced to a distance 
of 6' - 5 above the moon’s limb. At the bottom edge of the spectrum the intensity is 
* It will be observed that this photograph is solarised, owing to over-exposure. 
