98 



Profs. Liveing and Dewar. 



[Jan. 15, 



(1.) A line in the greenish-yellow, wave-length 5710*7. This line 

 is seen not only in the arc but also when sparks are taken from a 

 solution of magnesium chloride. It is a faint line in the latter case, 

 but is seen in carefully purified samples of the chloride, so that we 

 have no doubt that it belongs to magnesium. 



(2.) A line, wave-length 4351*2, very close to a chromium line. It 

 is easily distinguished from the latter by its expansion when fresh 

 magnesium is introduced, while the chromium lines show no expan- 

 sion ; it is moreover slightly more refrangible than the chromium line, 

 and as it dies out it is sharply defined on the less refrangible side, 

 diffuse on the other side. This line frequently appears to be accom- 

 panied by another weaker line at wave-length 4355, but we have not 

 been able to assure ourselves that the latter is due to magnesium alone. 

 It is certainly developed when magnesium is introduced, and we do 

 not know of any other element to which it can be ascribed ; but it 

 behaves very much as some lines of other elements, e.g., the violet lines 

 of lithium mentioned below, coming out brightly when magnesium 

 is freshly put in, and then nearly dying away. 



(3.) A violet line, wave-length 4166, identical in position with the 

 sharp edge of one of the so-called carbon bands in that region. 



(4.) Another violet line, wave-length 4057'3, easily recognised by 

 its position between two strong iron lines. 



These last three lines are conspicuous in Capron's photograph of 

 magnesium in arc. The last (4) may be seen in the spark between 

 magnesium points, and was formerly mistaken by us for the potassium 

 line which is near it. The lines (3) and (4) are so nearly identical 

 in position with two well-known lines of lead, that a suspicion naturally 

 arises that they may be due to the presence of that metal. We can 

 never be sure that the carbons or even the limestone crucibles we 

 have used may not contain traces of lead ; but if that be so, and these 

 lines are the consequence, it is certain that these two lines are 

 strengthened in an extraordinary degree by the presence of mag- 

 nesium, while the other lines of lead are not seen at all. Photographs 

 of the arc taken when a fragment of magnesium is introduced show 

 these lines strongly, while no trace is discernible of other lead lines 

 4062, 4246, or 4386. Until we get some other evidence of the pre- 

 sence of lead, we must ascribe these lines to magnesium. 



There are lines in the "normal solar spectrum" corresponding to 

 all these lines except that at 4166. That at 5710*7 is ascribed to iron 

 by Angstrom, but neither Thalen, Kirchhoff, Huggins, nor Bois- 

 baudran give any iron line at this place ; and we have not succeeded 

 in seeing any such iron line, but there is an air-line very nearly in 

 that position. Rutherford's photograph of the solar spectrum shows 

 a line very nearly at 4166. 



This makes the spectrum of magnesium in the arc consist of lines 



