OF SPECTRA AT THE SOLAR ECLIPSE OF JANUARY 22, 1898. 
389 
In the hydrogen images A, and the group E, F, G are much fainter than the others, 
and in the ultra-violet the images of these four are very difficult to trace heyond the 
line Hd. A is however very faintly indicated in the three titanium lines above 
mentioned. 
The prominence C is much brighter in hydrogen than the foregoing and can he 
traced to the line H/r. The three titanium lines are also clearly shown. 
The prominences B and D are the strongest of all in hydrogen, and although 
very different in size and general character (as observed in Ha about an hour before 
totality) they give practically the same spectra. 
The prominence B being situated very near the point of second contact (P.A. 56°) 
its spectrum falls almost exactly along the centre of the photograph. It is, therefore, 
in the best position for accurate measures ; and as the images in the ultra-violet are 
well defined circular black spots, much smaller and more definite than the corre¬ 
sponding images of the large prominence D, the wave-lengths given in Table II. are 
deduced from measures of this prominence only. (See p. 411 et seq.) 
Nearly all the lines measured can however be traced also in D ; there are fifty-two 
altogether and all of these are accounted for by the elements H, He, Mg, Al, Ca, Sc, 
Ti, Fe, and Sr. 
A remarkable feature in the spectra of the prominences B and D is well shown in 
this negative. The prominence D shows traces of a continuous spectrum in the 
region less refrangible than H, which is absent in B. Both prominences, however, 
give a strong continuous spectrum in the extreme ultra-violet, beginning abruptly at 
X,3668 near the end of the hydrogen series, and extending as far towards the smaller 
wave-lengths as the impression of the corona spectrum can he traced. 
The actual limit to which the extremely narrow streak due to the smaller promi¬ 
nence B can be traced is at A. 3435 ffi. 
This feature is shown more or less distinctly on all the negatives taken during 
totality. In No. 3 sj^ectrum as well as No. 7 a considerable arc of the chromosphere 
itself shows the ultra-violet continuous spectrum, all the flash spectrum lines more 
refrangible than A. 3668 being immersed in a uniformly shaded band which is absent 
in the less refrangible region. 
(In the reproductions this delicate feature is lost in all the images except fig. 2. 
It is shown more satisfactorily in the enlarged image in Plate 12.) 
Sqoectrum No. 5 (Plate 11, fig. 3).—This was exposed during about 40 seconds near 
the time of mid-eclipse. The exposure began about 10 seconds after the dis¬ 
appearance of the chromosphere on tlie east limb and ended immediately after its 
partial reappearance on the west limb. 
The continuous spectrum of the corona is strongly impressed, and can be traced as far 
as \ 3300 in the ultra-violet. The green corona line is well shown on the east limb ; 
it is distinguished from radiations due to the chromosphere by its diffuse character 
and distinct maxima of intensity corresponding with the brighter regions of tiie 
