252 Investigations on the Spectrum of Magnesium. [May 31, 



Group near s. 



In their list of lines in the spectrum of magnesium ( c Phil. Trans.,' 

 1884, p. 95) Messrs. Hartley and Adeney have given two lines, 

 \3071'6 and \3046'0, which we had not heretofore observed either 

 in the spectrum of the flame, arc, or spark of magnesium ; but in our 

 recent observations we have noticed in many cases a well-marked line 

 which, by interpolation between neighbouring iron lines, appears to 

 have a wave-length about 3073*5, and a pair of narrow bands sharply 

 defined on their less refrangible sides at wave-lengths about 3050*6 

 and 3046*7, and fading away on their more refrangible sides. 



We have little donbt that the lines we have observed are identical 

 with those given by Messrs. Hartley and Adeney, notwithstanding 

 that there is a much greater discrepancy between the wave-lengths 

 assigned by them and by us than there is between the wave-lengths 

 we have respectively found for the iron lines in the same neighbour- 

 hood. 



We have noticed the occurrence of this group in the spectrum of 

 the arc from a Siemens' dynamo between magnesium electrodes in a 

 variety of gases, in all in fact in which we have examined the arc, 

 except in sulphurous acid gas which is opaque to rays of this re- 

 frangibility. Also in the arc from a De Meritens' magneto-electric 

 machine between magnesium electrodes in air, in the flash of a mag- 

 nesium ribbon dissipated by the discharge of a storage battery, in the 

 spark of an induction coil worked in the usual way in air and in 

 hydrogen at atmospheric pressure, and in one instance in the spectrum 

 of an oxygen vacuous tube with magnesium electrodes when a Leyden 

 jar was connected with the secondary wire of the induction coil. 



On the other hand, we do not see this group in the spectrum of 

 other vacuous tubes with magnesium electrodes or with magnesia in 

 the tube, nor in the spark from an induction coil in air or hydrogen 

 at atmospheric pressure when the coil is worked with a De Meritens' 

 machine on the primary wire, nor in the flame of burning magnesium, 

 nor in the oxy hydrogen flame with magnesia or magnesium chloride, 

 nor in the arc between carbon electrodes in a crucible of magnesia. 



The circumstances under which this group is seen and is not seen, 

 do not seem to indicate that its emission is connected with any par- 

 ticular temperatures so much as with the character of the electric 

 discharge, and perhaps also with the density of the magnesium 

 vapour. 



