192 
ME. 3S T . STOBY-MASKELYNE ON THE 
stopper (fig. 3), or by a funnel of that metal (fig. 4) with a stopper (/) at the top and 
a fine orifice at its lower extremity (h). 
In the side of the retort just below the neck a straight delivery-tube is fixed, which 
again fits into another platinum tube (fig. 5) that, after taking a curve into a vertical 
position, is enlarged into a long cylinder capable of passing nearly to the bottom of a 
test-tube. The test-tube, into which it is fitted by a cork, holds when properly charged 
7-5 cub. centims. = 6-6 grammes of a strong solution of ammonia (of specific gravity 
= 0-88), corresponding to 2-03 of H 3 N, and a glass delivery-tube passes to the bottom 
of another test-tube, also containing a little of that alkali. 
The mineral to be analysed is first powdered extremely carefully in an agate mortar ; 
of this a quantity, that may be from 0-2 to 0-5 gramme, is introduced into the retort 
together with a small platinum ball. The tubulated stopper (fig. 2) is now introduced 
into its place and cemented by the aid of a little gutta-percha varnish, and by winding 
Fig. 5. Fig. 6. 
over its edge and that of the retort a strip of thin sheet india-rubber. The funnel, with 
its top closed, is now put into the tubulure of the stopper and filled with the hydrogen 
fluoride, which has a strength of about 32 per cent, of acid : it contains 1T2 grm. of pure 
hydrogen fluoride, capable of rendering gaseous 0’84 grm. of silicic acid (Si0 2 ), and of 
neutralising 0‘95 of ammonia (H 3 N). The acid is next admitted to the silicate, and the 
funnel removed to give place to the small stopper, which, and the joint of the platinum 
delivery-tubes, are now sealed with gutta-percha varnish. The apparatus, as prepared for 
use, is represented in fig. 6*. Pure dry hydrogen is next allowed slowly to traverse the 
whole apparatus, and the retort is placed in a water-bath at 100° C. for two hours, and 
occasionally shaken to set the ball in motion. During this operation only a minute trace 
of silicium difluoride comes over. 
* The scale of figs. 1 to 5 is one-half the actual size of the apparatus ; that of fig. 6 is one-third the actual 
size. 
