MINERAL CONSTITUENTS OP METEORITES. 
199 
available being too minute for any quantitative results to be expected from it. More- 
over it was found that these little crystals resisted the action of the acids when in their 
integrity, and when crushed their minute size rendered the manipulation most difficult 
and the results uncertain. Boiled for a long while in nitric acid they were unchanged ; 
and even hydrogen fluoride had no apparent action on them. They passed unscathed 
through a fusion of a small amount of Oldhamite with potassium-sodium carbonate ; 
though when fused with potassium chlorate, a crystal of Osbornite entirely disappeared — 
perhaps from its escaping notice. 
The crystals are very brittle, and after being crushed the powder retains the beautiful 
golden colour of the surface, which is therefore intrinsic and not due to a tarnish ; to 
which cause, however, a ruddy hue in two of the crystals may have been due. 
A few of the crystals placed in a glass tube and ignited in a current of dry oxygen, 
underwent only an external oxidation. Similarly treated in a slow current of chlorine 
they were decomposed ; but they sank into the glass before the decomposition was com- 
plete. An experiment was therefore made in which the crystals were supported in a 
cavity scooped into a small fragment of the thinnest Japanese porcelain. The chlorine 
was passed through the apparatus and the tube ignited previously to the introduction 
of the octahedra, in order to determine that the chlorine was without action on any 
part of the apparatus. This apparatus consisted of a small tube in which the porcelain 
splinter was placed; one end of it was drawn out so as to form a U tube of considerable 
length, surrounded at the bend by a freezing-mixture, and dipping ultimately into a tube 
of pure water. 
The whole of the available material, except a few crystals reserved for the gonio- 
meter, was placed in the cavities of the porcelain splinter, and dry chlorine allowed 
slowly to traverse the tube. On the application of heat shortly below redness, a glow 
was seen to commence among the minute crystals, which, extending itself through the 
whole, lasted for a few seconds. 
The crystals appeared to have somewhat increased in bulk ; they still retained their 
forms, but their metallic lustre had left them, and their colour became of a pale honey- 
yellow. 
The tube had become slightly iridescent in front of the assayed mineral; the drawn- 
out portion of it contained a small amount of a white sublimate, and a slight fuming 
came with each chlorine bubble through the water. 
The altered crystals on being exposed to the air soon began to deliquesce and assumed 
a pasty consistence ; treated with water, they only partially dissolved and gave the solu- 
tion an alkaline reaction. This solution gave no precipitate with ammonia, but yielded 
one to ammonium oxalate. The residue undissolved by this water dissolved, though 
neither with ease nor quite completely, in hydrogen chloride. 
This acid solution gave with ammonia a slight precipitate, which was redissolved in 
acid and reprecipitated : it seemed to contain a little iron. Filtered from this, the solu- 
tion further gave a very distinct precipitate on being treated with ammonium oxalate and 
2 d2 
