110 REroET— 1891. 



joints being luted with plaster of Paris. The tube containing 

 the magnesium was then heated to whiteness in a Fletcher's 

 blast combustion furnace for about half an hour, a moderate 

 current of hydrogen being kept up. On cooling and dis- 

 connecting, most of the magnesium was found in the form 

 of solid lumps, firmly adhering to the walls of the condenser 

 close to the furnace, the rest (only a small amount) being 

 deposited as dust. These lumps were malleable under the 

 hammer. 



It was now decided to use filings. To make them, some 

 ribbon was fused in an iron crucible, under a flux having the 

 composition 3KC1.4MgClo, and containing a little ammonium 

 chloride, as recommended by Matthiessen (Journ. Chem. Soc, 

 viii., p. 107). This flux must not be heated much above the 

 melting-point of magnesium, as it then loses chlorine, and 

 gets gummy. On stirring with an iron rod, the metal ran 

 together into a large globule, after which the crucible was 

 allowed to cool. When the globule had solidified it was taken 

 out with an iron spoon, sufficient flux adhering to it to protect 

 it from oxidation. A curious efi'ervescence was noticed in the 

 flux while the metal was melting. As soon as it was cool the 

 lump of metal was washed, first with water, then with alcohol, 

 and dried at 100*^, after which it was filed up with a rasp. A 

 coarse single- cut file would perhaps have been better. The 

 filings so formed contained traces of iron from the file ; but it 

 was not thought that this w^ould have any deleterious effect on 

 the experiments. 



In this connection I may draw attention to the peculiar 

 action of magnesium on glass. When heated in hydrogen in 

 contact with glass — e.g., in a flask — the glass blackens, where 

 it is hot, just before a red heat is reached, even where there is 

 no metal in contact with it. This phenomenon was also 

 noticed when the magnesium was heated in a platinum boat, 

 placed in a glass tube, through which hydrogen was passing, 

 the glass being blackened above the boat. 



This action aj)pears to take place only in hydrogen, as, in 

 conjunction with Mr. J. B. Kirkland, I have repeated the 

 experiment wdth carbon monoxide, carbon dioxide, sulphur 

 dioxide, air, and hydriodic-acid gas, in all of which the glass 

 was only attacked where it happened to be in contact with 

 the magnesium. Magnesium only acts on these gases (with 

 the exception of air) at a red heat, so that these experiments 

 were strictly comparable with the heating of the metal in 

 hydrogen. 



We were hence led to the belief that magnesium must be 

 volatile in hydrogen below its melting-point ; w^hich appeared 

 to be confirmed by an experiment conducted by Mr. Kirkland. 

 He heated some magnesium-filings in an iron boat, in a roll of 



