1888.] Investigations on the Spectrum of Magnesium. 245 



In fact, when chemical changes are occurring in a flame it cannot 

 be taken for granted that the temperatures of the molecules are all 

 alike, or that the vibrations which they assume are the result of heat 

 alone. On the other hand, the temperature of the metal separated 

 from magnesia by the oxyhydrogen flame cannot, we suppose, be at a 

 temperature higher than that of the hottest part of the flame. We are 

 therefore inclined to think that the metallic lines (6) are manifested at 

 a lower temperature than the bands of the oxide ; and the appearance 

 of a line in the position of the first band without any trace of the 

 second band (which is nearly as bright as the first), and without any 

 trace of the b group, is quite sufficient to create a suspicion of mis- 

 taken identity when Mr. Lockyer ascribes the sharp green line in the 

 spectrum of nebulae to this band of magnesia. This suspicion will be 

 strengthened when it is noticed that the line in question is usually in 

 the nebulas associated with the F line of hydrogen, if it be borne in 

 mind that the spark of magnesium in hydrogen does not give the 

 bands, and that the oxyhydrogen flame hardly produces them from 

 magnesia when the hydrogen is in excess. 



In Mr. Lockyer's map of the spectrum of the nebula in Orion 

 (loc. cit., p. 134), he has represented three lines in the position of the 

 edges of the first three of these bands. If these three lines were 

 really seen in the nebula, there would be less room to doubt the 

 identity of the spectra; but the authorities quoted for the map (loc. 

 cit., p. 142) mention only a single line in this position. 



When the flame of burning magnesium is viewed with a high 

 dispersion these bands are resolved into series of fine, closely set 

 lines. Seven such series may be counted, beginning at the approxi- 

 mate wave-lengths 5006'4, 4995"6, 4985'4, 4973-6, 4961-6, 4948'6, 

 4934'4, respectively. When a condensed spark is taken between 

 magnesium electrodes in oxygen mixed with a little air, the pair of 

 strong nitrogen lines may be seen simultaneously with the bands, and 

 lying within the first band, the bright edge of the band being some- 

 what less refrangible than the less refrangible of the two nitrogen 

 lines. 



When the bands are produced by the spark discharge between 

 magnesium electrodes in oxygen or other gas, we have not been able 

 to resolve them into lines, but the whole amount of light from the 

 spark is small compared with that from the flame, and besides it is 

 possible that the several lines forming the shading may be expanded 

 in the spark, and thus obliterate the darker spaces between them. 



Triplet near M and adjacent Bands. 



Our former account of the spectrum of the flame of burning magne- 

 sium included a description of a triplet near the solar line M, and a 

 series of bands extending from it beyond the well-known triplet near 



