1888.] Investigations on the Spectrum of Magnesium. 247 



and half its volume of oxygen and burn the gases as they issued from 

 the holder. 



We have not noticed the more refrangible triplet (\3633"7 to 

 36206 about) under other circumstances, but the triplet near M is 

 produced when magnesia is held in the flame of cyanogen burning in 

 oxygen, in the flash of pyroxylin with which magnesium filings have 

 been mixed, or which has been treated with an alcoholic solution of 

 magnesium chloride. 



It is not only very strongly developed, but shows strongly reversed 

 on our photographic plates, in the spectrum of the arc from a Siemens' 

 dynamo taken between electrodes of magnesium in oxygen ; and most 

 of the accompanying ultra-violet bands of the magnesium flame 

 spectrum are at the same time reversed. It is less strongly, but 

 distinctly, reversed in the spectrum of the same arc taken in air, in 

 carbonic acid gas, and in sulphurous acid gas. It appears also if 

 the arc is taken in ordinary nitrogen unless great precautions are 

 taken to exclude all traces of oxygen or carbonic acid, when it com- 

 pletely disappears. It is developed also in the flash produced when a 

 piece of magnesium ribbon is dissipated in air by the discharge through 

 it of the current from 50 cells of a storage battery. Also in the 

 spark in air at atmospheric pressure between magnesium electrodes 

 connected with the secondary wire of an induction coil when the 

 alternating current of a De Meritens' magneto-electric machine is 

 passed through the primary. 



In two cases, but only two, we have found this triplet, or what 

 looks like one or both of the more refrangible of its lines, developed 

 in vacuous tubes. In both tubes the gas was air. One had platinum 

 electrodes and a strip of magnesia from burnt magnesium disposed 

 along the tube ; the other had fragments of the Dhurmsala meteorite 

 attached to the platinum electrodes. The discharge was that of an 

 induction coil worked in the usual way without a Leyden jar. In 

 each case it is only in one photograph of the spectrum that the lines 

 in question appear. In other photographs taken with the same tubes 

 they do not show. 



On the other hand, this triplet does not make its appearance in the 

 arc from a dynamo between magnesium electrodes in hydrogen, coal 

 gas, cyanogen,* chlorine, hydrochloric acid, or ammonia ; nor in the 



* In taking the arc in this way in cyanogen our photographs show the whole of 

 the five bands of cyanogen between K and L well reversed. We have before 

 noticed (' Eoy. Soc. Proc.,' vol. 33, p. 4) the reversal of the more refrangible three 

 of these bands against the bright background of the expanded lines of magnesium 

 when some of that metal was dropped into the arc between carbon electrodes, but in 

 taking the arc between magnesium electrodes in an atmosphere of cyanogen the 

 bright wings of the expanded magnesium lines near L extend beyond the cyanogen 

 bands, and the whole series of the latter are well reversed. — May 23. 



