or Astrophysics

Change in Physical Conditions.—The early idea that the position of a line of a spec trum was the result of chance is still further modified by the change of the wave-length of a line resulting from a change in the physical conditions. Jewell noticed while measuring solar spectrum photographs that the arc lines of the comparison spectrum did not in many cases exactly correspond with the lines in the sun. This led to the investigations of Jewell, Humphreys and Mohler (Astrophysical Jour nal, VI, 169, 1897), on the spectra of an arc under pressure of from one to 15 atmospheres. From measurements of the spectra of 53 ele ments, it was shown that the lines are shifted by pressure toward the red end of the spec trum, the amount of the shift being directly proportional to the increase of pressure, but being independent of the temperature. For a given element, the shifts of similar lines are proportional to their wave-length, but lines of different series principal first and second sub ordinate, are shifted in the ratios of 1:2:4.

The appearances of lines in a spectrum are greatly altered by other physical conditions. Eder and Valenta have found that argon gives three distinct spectra under different electrical conditions, Schenck that the spark line of Mg at A 4481, which has so often been considered as a sure sign of a high temperature, vanishes if the electrodes become so hot that they glow and begin to melt. Lockyer has made a great number of investigations on 'enhanced" lines, or those which are brighter in the spark than in the arc. He explains their meaning on the assumption that the spark is hotter than the arc, an assumption which is hard to reconcile with other observed phenomena. In this con nection, Kayser, in his excellent

Solar Spectrum.—The infrared solar spectrum has been investigated by Becquerel and Lommel to 2< 9500 by using phosphorescent screens, by Abney to R 27,000, who photo graphed with a special emulsion of silver bromide and collodion, and by Langley by the use of the bolometer to 2< 53,386. Rowland's (Baltimore 1888) made with a powerful concave grating extending from A 3000 to R 6950, and his 'Table of Solar Spectrum Wave-Lengths> (Chicago 1898), which contain accurate measures of some 20,000 lines, give the most accurate information we possess of the solar spectrum, and have been accepted by all astrophysicists as the common standard of reference. Rowland's determinations of relative wave-lengths leave little to be desired in ac curacy. Measures were made by the method of coincidences which is rendered possible by the use of concave gratings which permit two over lapping spectra to be photographed on the same plate without change of focus. However, the standard lines of the spectrum have not had their wavelengths determined with equal ac curacy, and they have not been quite properly spaced throughout the spectrum, with the re sult that at the Saint Louis conference, 1905, a redetermination of Rowland's standards was re garded as one of the present greatest needs in astrophysics. Considerable work has already been done toward obtaining the wavelengths of 200 standard lines to be reduced to the so called 'international system." This work has made evident the difficulty of obtaining satis factory wavelengths for the standard lines on account of a slight change in wavelength due to a change in physical conditions. Although the systematic deviations of Rowland's wave lengths from the international system amounts to 0.02 Angstroms or more, the accidental er rors are much less than 0.01 Angstrom. As present-day investigations of solar problems need the third decimal place in the wave lengths, St. John concludes (