166 M.v.¥ridesu\x on the mean Atomic Weights of Simple Bodies. 



Then 3-9 phos. + 5' oxyg. will give 8*9 for the atom of phos- 

 phoric acid. But this quantity saturates two atoms of base, 

 and therefore probably contains two atoms of phosphorus, 



according to Berzelius's formula (Ph 9 ). Thomson makes 9 

 phosphoric acid saturate 2 atoms of base; Berzelius 892*3 1. 

 His number for phosphorus 196* 155, being very nearly half 

 of that found above, and the mean of the three (calculating- 

 Thomson's by the same formula) is adopted, neglecting the 

 low decimals. 



(n.) Rhodium, &c. — Berzelius has given in the 40th volume 

 of the Ann. Ch. some elaborate Recherches sur les Metanx qui 

 accompagnent le Platine, in which their atomic weights were 

 ascertained by a very unexceptionable process, — reduction of 

 the triple chloride, placed in a glass tube, over a spirit-lamp, 

 by a current of hydrogen gas. The loss gives the chlorine ; 

 and the alkaline chloride being washed away, the metal is 

 weighed. Soda-chloride of rhodium was used ; and potash- 

 chloride of the other four. His numbers will hardly require 

 correction for the very different ones of Thomson, obtained by 

 less satisfactory operations ; but being calculated from chlo- 

 rine, estimated at 442*65, they give (when adjusted to 4*46, 

 the mean weight of that substance,) the numbers in the third 

 column. 



(o.) Silicon. — Berzelius's number is dependent on his 

 Canon, before quoted; and the deductions of Thomson (First 

 Prin., article Silicon) are so satisfactory, and his number is so 

 convenient in application, as to claim preference over any 

 result of hypothetical considerations. 



(p.) Tungsten. — The difference between these numbers is 

 so great, that one of them must in all probability involve some 

 unobserved cause of error; but the results of the experiments 

 (First Prin. ii. 62; and Ann. Phil. iii. 245) are so nearly 

 equal, that it is not easy to discover with whom it lies. It 

 seems better to go between the two, than to run the hazard 

 of taking the wrong; and the mean is adopted from expe- 

 dience merely. 



(q.) Yttrium and Zirconium. — Yttrium is subject, in a less 

 degree, to similar observations with tungsten; and zirconium 

 also, with this difference, that the difficulty of choice consists 

 in the experiments of Thomson being indirect, and those of 

 Berzelius not given. 



(r.) Zinc. — The difference here also is great ; but Thomson's 

 experiments are so comprehensive, and appear to have been 

 so attentively conducted, that, notwithstanding the exceptions 

 taken to them, his number seems entitled to more confidence 

 than that of Berzelius, drawn only from the composition of the 



oxide 



