of the combinations of phosphorus. 323 
M. Dulong’s two methods of ascertaining the quantity of 
chlorine in the sublimate, appear to me at least as objection- 
able as his process for determining the composition of phos- 
phoric acid, and liable to great errors : the first from the 
uncertainty of the absolute quantity of chlorine admitted; and 
the second, from the loss arising from the vapour of the sub- 
limate, which must be carried off by the current of chlorine. 
How great a deficiency may originate from the last circum- 
stance, is shown by the following experiment : 5 grains of 
phosphorus were converted into sublimate by chlorine in 
great excess, the remaining chlorine was displaced by pa ssing 
common air through the vessel for some time, till not the 
slightest smell of chlorine could be perceived ; the reton was 
then weighed, and a current of air passed through it. Though 
this current could hardly have replaced the air contained in 
the retort, yet the loss of weight was 1.7 grains, and copious 
vapours were produced in the atmosphere. In a second trial 
of the same kind, there was a greater loss of weight, and by 
barely exhausting the retort, and then again admitting air, 
there was a loss of ^ of a grain. 
When chlorine is made to act upon phosphorus over mer- 
cury not carefully dried, some muriatic acid gas is always 
formed ; but when the mercury has been recentl boiled, no 
effect of this kind is produced, and the vapour in the gas» 
forms a minute quantity of a liquid hydrate of the perchloride 
which by more water, is converted into muriatic and phos- 
phoric acids, as I proved by some very delicate experiments; 
so that there is certainly no hydrogen denoted in phosphorus 
by the action of chlorine, and in their mutual action a mere 
binary compound of the two substances is formed. 
Uu 
MDCCCXVIII. 
