61 
chloride, the gas was analysed, according to Bunsen’s 
method, and proved to be a mixture of carbonic oxide, 
marsh gas, di-methyl, with a small amount of olefines, the 
following being the result : — 
Carbonic Oxide 4 - 29 
Olefines 0 - 63 
C 2 H 6 18*57 
CH 4 76*51 
100*00 
This analysis clearly shows that Schiitzenberger’s descrip- 
tion of the decomposition is incorrect, the greater part of 
the gas consisting of hydride of methyl or marsh gas. 
II. Preparation of Di-methyl by Frankland’s Method. 
Schiitzenberger’s method not giving the required gas, this 
process was tried. I took advantage of the directions given 
by Schoyen for the preparation of pure di-ethyl. 
Into stout glass tubes closed at one end, metallic zinc, 
made rough on its surface, was introduced; the open end 
was then softened before the blowpipe, thickened, and drawn 
out into a strong capillary, which was bent twice at right 
angles, as described by Frankland. Through this the methyl 
iodide was introduced, and afterwards the ether, equal in 
volume to that of the iodide. The air was expelled by 
boiling the ether and closing the capillary. 
Tubes thus prepared were heated to 130° C. until all the 
zinc was dissolved, then opened before the blowpipe when 
cold, to allow any marsh gas to escape which had been 
formed by the presence of moisture, again closed and heated 
to 150° C. for several hours. The tubes when cold were 
then immersed in a mixture of salt and ice; the end 
of a narrow tube of thick caoutchouc drawn over the 
capillary, the other end being attached to a gas-holder con- 
taining a saturated solution of common salt, and the end of 
the capillary was then broken off. The gas rushed into the 
