28o 



NA rURE 



[January 19, 1899 



graphs taken by Mr. Kowler during the eclipse of 1893 

 (Phil. Trans., A, vol. 187, p. 593). Among the brightest of 

 these rings, which is common to all three sets of photographs, 

 is one about wave-length 4231, which probably is identical with 

 the corona line photographed by Schuster in 1886, and staled to 

 have a wave-length of 4232 -a on Angstrom's scale (4233 4 

 Rowland). Schuster stated that this line was " probably the 

 same line as 4233 'O often observed by Young in the chromo- 

 sphere" (Phil. Trans., A, vol. 180, p. 341). The chromo- 

 spheric line at this wave-length has since been identified as an 

 enhanced line of iron, of which the precise wave-length is 

 4233'3- Captain Hills photographed this corona line with a 

 slit spectroscope in the last eclipse, and he gives its wave- 

 length as 4233-5 (J\'oy. Soi. Proc, vol. 64, p. 54), which within 

 the limits of error might be considered coincident with the 

 enhanced line of iron. 



The later researches on the spectrum of iron have shown that 

 the iron line which I observed in 1869 to be coincident with 

 the bright chromospheric line at 1474K (531679 Row- 

 land) is also an enhanced line, agreeing absolutely with 

 Young's latest determination of the wave-length of the 1474 

 chromospheric line (Scheiner's " Astronomical Spectroscopy," 



With regard to the ring in the green, the lack of sufficient 

 photographs on isochtomatic plates in 1893 does not permit of a 

 final determination of wave-length. Important data, however, 

 were obtained, both in 1896 and 1898. A measurement of the 

 position of the chief ring in the green, as shown in these photo- 

 graphs, comparing the ring with the spectrum of the chromo- 

 sphere and a solar and iron spectrum taken by the same 

 prisms, shows beyond all question that the wave-length is very 

 different from that generally accepted. The mean result of 

 measurements of different parts of the ring made by Messrs. 

 Fowler and Shackleton and Dr. Lockyer is 53037, or about I J 

 tenth-metres more refrangible than 1474K (531679). 



Although the new wave-length is not to be regarded as final, 

 for the reason that the conditions under which the photographs 

 were taken necessitate certain small corrections which have not 

 yet been fully workevl out, it is not likely that it can be in error 

 by so much as I tenth-metre. 



The examination of the photographs, which has been under- 

 taken in the first instance by Mr. Fowler, indicates that other 

 important conclusions are to be drawn from the admirable series- 

 obtained by him, among them the possible existence of one or 

 more new gases, some of the lines of which, as gathered from (he 





t the chief hne m the >peclrum ol the corona (.\) «ilh ihe enhanced hne 

 spectrum of the chromosphere (b) and in the ordinary solar spectrum C. 



Frost's translation, p. 425), with which, according to 

 his eclipse observations, the green line of the corona is 

 coincident. 



According to these results then, two of the chief lines in the 

 spectrum of the corona would be coincident with enhanced 

 lines of iron. The remaining corona lines which have so far 

 been measured, are not, however, coincident with enhanced 

 tines. It did not seem possible, therefore, that two of the 

 enhaiiced line.-, of iron should be present without the others, 

 even if it be admitted that the corona may have a temperature 

 high enough to produce any enhanced lines. 



It appeared then, either that the coincidences of the chromo- 

 spheric and coronal lines about 423 and 531 were accidental, 

 or that they were not real coincidences at all. A careful 

 examination of the eclipse photographs of 1896, taken by Mr. 

 Shackleton, and those of 1S98, taken by Mr. Fowler, has 

 therefore been undertaken, with special reference to this 

 point. 



The wave-lcnglh of the coronal ring at 4231, already pub- 

 lished in case of the 1893 photographs, has been confirmed. 



The 1896 and 1898 photographs further indicate that the 

 corona line near 4231 is not coincident with the chromospheric 

 line to which reference has been made, and show that while the 

 chromospheric line is coincident with the enhanced line of iron 

 at A4233'3, the corona line has a wave-length of 4231 •3. 



NO. 1525, VOL. 59] 



dispersions as yet available, appearing also in the spectra of 

 some stars and planetary nebula;. 



The photograph which accompanies this paper has been 

 prepared by .Mr. Fowler. 



NIGH VACUA PRODUCED BY LIi^UID 

 HYDROGEN.-' 

 A S an illustration of the extraordinary power of the new 

 ■'"'■ cooling agent— liquid hydrogen, the extreme rapidity with 

 which high vacuo can be produced by its use is, perhaps, one of 

 the most striking. The absolute boiling points of hydrogen, 

 oxygen, and chlorine are respectively 35 , 90 and 240", in 

 other words oxygen boils at a tem|x;rature two and a half times 

 higher than lii|uid hydrogen, and liquid chlorine similarly at 

 two and a half times that of liquid oxygen. From this we 

 infer that liquid hydrogen .is a cooling agent ought to be relative 

 to liquid air as elTectivc .is the latter is c<mipared to that of 

 liquid chlorine. Now chlorine at the temperature of boiling 

 oxygen is a hard solid, some 80° below its melting point, and 

 in this condition has an excessively feeble vapour pressure. 



' " Application of Liquid Hydrogen to the Production of High Vacua, 

 together u-itli their Spectroscopic Examination. " Caper read at the Royal 

 Society on Deccmtier 15, 1898, by James Uew.-u, F.R.S. 



