70 THE INDIAN ECLIPSE, 1898. 



and a series of ten with the prismatic camera. The single 

 photograph obtained with the slit spectrograph failed from the 

 above-mentioned cause, the direct sunlight and halation nearly 

 obliterating the faint corona spectrum. The slitless spectro- 

 graph yielded two negatives of the flash spectrum at second and 

 third contacts. They show a large number of bright lines in 

 the region between F and H, but on the whole they do not 

 quite come up to expectation. The most interesting results 

 were those obtained with the prismatic camera. This instru- 

 ment gave images of the spectrum extending from X 6000 in 

 the orange to X 3350 in the ultra-violet and on a scale of 

 33 inch to the moon's diameter, the length of the spectrum 

 between the above limits being nearly three inches. 



All the ten photographs of the series yielded interesting 

 results ; the flash spectrum appears on Nos. 3, 7 and 8, while 

 the long exposures about mid-totality give the corona spectrum, 

 and those taken out of totality show the Fraiinhofer dark line 

 spectrum bordered with bright lines. 



In No. 3 the flash spectrum lines are beautifully denned in 

 the ultra-violet, where they can be traced as far as X 3342. In 

 this region, from H upwards, 218 lines can be counted. Through 

 the kindness of Dr. Eambaut, who supplied me with an excellent 

 measuring apparatus, I have been able to determine the wave- 

 length of every line with a very satisfactory degree of accuracy. 



From the results obtained it appears that a great many of 

 the lines are due to iron, some to calcium, magnesium, etc., 

 and three of the strongest lines in the spectrum have been 

 identified by Mr. Jewell (at the Johns Hopkins University) as 

 due to the comparatively rare element titanium. The vapour 

 of this metal is not, however, confined to the " flash " layer, but 

 extends as high in the chromosphere as hydrogen, and enters 

 into the composition of the prominences as well. 



In this plate (No, 3), there are thirty hydrogen lines shown ; 

 they follow each other with mathematical regularity as indicated 

 by Balmer's well-known formula, and with gradually diminishing 

 intensity. The Greek alphabet is, of course, exhausted in 

 designating them, starting with the C line as a, in accordance 

 with the usual procedure. The wave-lengths actually obtained 

 for these lines agree remarkably accurately with those derived 

 from the formula. A very slight deviation from exact agreement 

 occurs, however, in the lines beyond the w line, but these, on 

 account of their faintness, are subject to a greater uncertainty 

 than the others. 



In the visible portion of the spectrum (H to D), it is not 

 possible to determine the wave-lengths with anything like the 

 accuracy obtained in the ultra-violet ; but for the sake of 

 completeness, every line on the photograph has been measured 

 and its wave-length determined. The result obtained for the 



