of the combinations of phosphorus . 321 
according to the second, of 100 to 137.2; and according to 
the third, of ico to 131.3 : the mean will be 100 to 134.5. 
The light of the phosphorus burning in vapour in these 
experiments was excessively bright ; yet the top of the retort 
never became softened ; and the phosphoric acid, which in- 
creased the weight of the tube, principally combined with the 
glass at the aperture where it was red hot. I cannot but 
consider this process of burning phosphorus in the gaseous 
state in a great excess of oxygen, as the most accurate mode 
that has yet been devised for ascertaining the composition of 
phosphoric acid. In this instance no phosphorous acid, as I 
ascertained by direct trials, is formed from the vapour ; and 
no substances are concerned except those that actually com- 
bine. M„ Dulong’s method of ascertaining the composition 
of phosphoric acid, appears to me much too complicated to 
afford any results approaching to accuracy. He first com- 
bines copper wire with phosphorus, by passing phosphorus 
over it by means of a stream of hydrogen gas ; he then 
dissolves his phosphuret of copper in nitric acid, and deter- 
mines the quantity of phosphoric acid formed by precipitation; 
in all of which processes sources of error may exist. 
M. Berzelius's methods of ascertaining the composition of 
phosphoric acid, that of reviving gold from its oxide by 
means of phosphorus, and that of determining the quantities 
of phosphate and muriate of silver formed from perphos- 
phorane, or the perchloride of phosphorus, appear to me 
still more exceptionable ; yet his results on the quantity of 
oxygen approach nearer to mine than those of M. Dulong. 
The facts which I endeavoured to establish respecting 
chlorine, in a paper published in the Philosophical Transac- 
