252 On ths Cause of Chemical Proportions. [ApWI, 



know that the analysis of ytterite, made by the late Mr. Ekeberg, 

 is one of the most exact that mineralogy possesses. In it we find 

 28 parts of silica combined with 55"5 of yttria and 16*o of 

 oxide of iron, such as analysis gives it, and equivalent to 13*42 of 

 pure black oxide of iron. From experiments which I shall men'- 

 tion hereafter, 55"5 yttria contain 10-3 of oxygen ; the 2S of 

 silica, from the preceding data, should contain 10*9 of oxygen, 

 and the 15-42 of black oxide of iron contain 3-5 ; now 35 X 3 =s 

 10'5. This coincidence is an additional proof, that my determina* 

 tion of the composition of silica is very near the truth. I have not 

 spoken of the 4 parts of glucina contained in ytterite, because 1 dd 

 not know the composition of that earth. But it is probable that 

 they will not form an exception to the general rule. 



The great quantity of oxygen in silica renders it probable, that 

 it contains more than one volume, and as the composition of the 

 triple fluate of silica an 3 ammonia, shows that it cannot contain 

 S volumes, it is probable that it contains 2. In that case the 

 volume of silicium ought to weigh 216. 



11. O.smium. (Os). — Unknown. 



12. Iridiimi. (I). — Unknown. 



13. Bhodium. (R). — Dr. Wollaston, to whom we are indebted 

 for our knowledge of the existence of this metal, has been so good 

 as to furnish me with the quantity necessary for determining the ca- 

 pacity of saturation of this rare, and difficult to be procured, metal. 



My first essay was to reduce it from its triple muriate by means of 

 mercur)', in order to find in that manner the quantity of oxygen 

 which it contained, by means of t!ie quantity of mercury necessary 

 to reduce it. But this experiment did not succeed. Rhodium can- 

 not be reduced in that manner. We obtain a black powder com- 

 posed of amalgam of rhodium, muriate of mercury (calomel), and 

 an insoluble muriate of rhodium, of which I shall give a descrip- 

 tion afterwards. This experiment proves, that the affinity of rho- 

 dium for mercury is too strong to enable us to form, at the expense 

 of its oxide, a pennuriate of mercury (corrosive sublimate). 



I next endeavoured to combine rhodium with sulphur. I re- 

 duced the metal to powder in a steel mortar, and I then mixed it 

 with an equal weight of sulphur. When I heated this mixture, 

 the sulphur sublimed without combining with the metal; but 

 towards the end of the process, when scarcely any thing remained 

 but the yellow gaseous vapour of sulphur in the retort, the metal 

 took fire, and produced a sulphuret, which was not, however, satu- 

 rated with sulphur. 



In another experiment, I distilled concentrated nitromuriatic 

 aeid oft' the metal in the state of a very fine powder. The rho- 

 dium was not sensibly attacked. After I had in vain evaporated 

 considerable quantities of acid from the metal, I distilled it off at 

 last. The acid had acquired a reddish shade, but had dissolved 

 very little of the rhodium. Dr. Wollaston has informed us that 

 rhodium, in order to be dissolved ia acidsj must be alloyed with 



