164 EECENT PROGRESS IN ASTRONOMICAL RESEARCH. 



tions of wave length of numerous lines made by Messrs. Perot and 

 Fabry with their special interferometer. 



To the general reader it might seem a matter of indifference that 

 defects amounting to only one part in 250,000 should exist in the 

 relative accuracy of the wave lengths of different solar and metallic 

 spectrum lines, but important conclusions in at least three different 

 lines of research depend on these small differences. First may be 

 mentioned the determination of the motions of heavenly bodies in 

 the line of sight of the observer, which depend on the measure- 

 ment of the shifting of the lines of their spectra. In this work an 

 error of 0.02 Angstrom unit in the wave lengths corresponds to a 

 difference of velocity of more than a mile a second. As another 

 research demanding accurate standards of Avave length may be men- 

 tioned the comparison of metallic and solar or other celestial spectra, 

 for the purpose of determining the constitution of the heavenly 

 bodies. It has often occurred heretofore that erroneous theories of 

 the broadest scope have been based on a supposed coincidence of spec- 

 tral lines, which truly were of very slightly differing wave length. 

 Thirdly, the investigations of regularities in the distribution of the 

 different spectral lines of an element, on which may rest important 

 theories of the inner constitution of matter, would be vitiated by 

 errors of the wave lengths such as have been noted. 



It has been decided by the Union that a neAv sA^stem of Avave 

 lengths is necessary; that it shall be based on the Avave length of a 

 suitable spectrum line, to be fixed permanently, in units differing 

 as little as possible from jo^o Ao^o o niillimeter, and to be called Ang- 

 strom units; that secondary standards shall be determined by an 

 interferometer method at distances not greater than 50 Angstrom 

 units apart, and the source of light for the determination shall be 

 produced in the electric arc of to 10 amperes; that tertiary stand- 

 ards shall be selected at distances of 5 to 10 Angstrom units Avhose 

 waA^e lengths shall be determined from the secondary standards by 

 the aid of gratings. This Avork Avas placed in charge of a committee 

 of expert spectroscopists, and progress is reported from several 

 laboratories in making the measurements required. 



^. /Solar radiation. — The amount of solar radiation may be ex- 

 pressed in terms of the rise of temperature in degrees centigrade 

 Avhich a solar beam of a square centimeter cross section, shining for 

 a second of time, will produce if entirely absorbed and transformed 

 into heat Avithin a mass of one gram of water at maximum density. 

 Such units of measurement are called calories (or therms) per sec- 

 ond. Inasmuch as the amounts usually measured of solar radiation 

 are small, it is customary for the sake of convenience to use as a unit 

 of solar radiation a calory per minute, for the quantity usually ob- 

 served is between 1 and 2 calories per minute. 



