400 Mr. J. E. Kcrln. "i, tl,,> [Mar. in. 



order of accuracy to prove that the small observed interval betwc 

 the nebular line and the magnesium fluting is real, and not due 

 errors of observation. 



A detailed account of all the tests to which the apparatus was sub- 

 jected cannot be given here. Nothing that suggested itself wi 

 omitted. The best tests, however, both for constant and 

 accidental errors, are afforded by observations of the motion in 

 line of sight of bodies whose motion is already known. As 

 example of such observations, I may refer to the measures of 

 motion of Venus in the line of sight given in the table on p. 21 

 1 Publications of the Astronomical Society of the Pacific,' No. 11, ii 

 which the greatest error is one English mile per second. Simil 

 measures of the displacement of lines in the lunar spectrum we 

 seldom in error by more than two miles, and measures of the moti< 

 of a-Tauri and a-Orionis, usually made on the same nights that 

 nebula was observed, were of the same order of accuracy, as det 

 mined by their agreement with each other, and with the photograpl 

 results of Professor Vogel. 



In work of this character the periodic shifting of lines in 

 .spectra of the stars and nebula? due to the earth's annual motion is 

 of a magnitude not to be neglected, and it should appear in the cor 

 parison of observations made at different seasons. So faithfully ii 

 the orbital motion of the earth reflected in my observations on 

 nebula of Orion, that I would with some confidence undertake 

 determine the month of the year, by measuring the distance of 

 principal line from the lead line used in the comparison spectrum. 



With these remarks on the degree of accuracy which character 

 the observations, 1 give below the results which have been obtaii 

 up to the present time, for the nebula of Orion. 



From sixteen complete measures, made on eleven different night 

 (two of which were in the winter of 1889-90), the wave-length 

 the principal line, corrected for orbital motion of the earth, 

 X 5006'22 + 0'014, the probable error corresponding to an unc 

 tainty of U'5 mile per second in the line of sight. When t\ 

 measures were made on the same night, they were always in differ 

 spectra of the grating. 



Ten comparisons of the third nebular line with terrestrial hydrogen 

 were made on seven nights in 1890-91, showing, when corrected for 

 the orbital motion of the earth, a displacement of the nebular line 

 toward the red of 0'28 + 0*026 tenth-metres. This corresponds to a 

 motion of recession of the nebula from the sun of 10*7 + I'O miles 

 per second. 



In recent comparisons of hydrogen with the third nebular line, I 

 have not been able to attain the small probable error of 1^ miles per 

 second for a single evening's comparison, given in my letter to the 



