1814.] Elements of Organic Nature are combined. 407 



analysis the water obtained from 690 grains did not exceed 0*3 grain 

 in weight. 



The small glass containing the potash remained usually 24 hours 

 in the gas, though five or six hours are sufficient. I then allowed 

 the mercury to fall by opening the stop-cock, and I removed the 

 jar to prevent any accident in withdrawing the small glass through 

 the mercury. I removed the leather very quickly, and corked up the 

 glass with the same cork that had been formerly employed. I then 

 wiped its exterior surface clean, and weighed it. In several cases I 

 examined the residual oxygen gas by burning it along with hydrogen, 

 but never could discover the smallest trace of carbonic oxide. 



1 must observe that in these experiments a small quantity of car- 

 bonic acid remains combined with the soda in the burnt mass. The 

 oxide of lead decomposes the muriate of soda, aud forms a submu- 

 riate of lead, setting at liberty a corresponding quantity of soda. 

 The free alkali seizes upon the carbonic acid requisite for its satura- 

 tion, and of course this quantity remains in the burnt mass. It 

 would be difficult to determine by a direct experiment the quantity 

 of carbonic acid retained by the soda, because the greatest part of 

 this carbonic acid may remain dissolved in the water requisite for the 

 solution of the salts. But we may easily calculate that quantity; 

 for in the submuriate of lead the acid is combined with four times 

 as much base as in the neutral muriate ; and consequently the oxide 

 disengages a quantity of soda which requires, in order to become a 

 subcarbonate, one-fourth as much carbonic acid as would be 

 requisite to carbonate the whole oxide of lead present ; that is to 

 say, almost exactly one-twentieth of the weight of the oxide of lead. 

 We must then in each experiment add this quantity of carbonic acid 

 to what we obtain in the state of gas. But there is one exception, 

 and that is in the analysis of oxalic acid, where a portion of the 

 oxide of lead is always converted into peroxide, and this portion 

 does not decompose the muriate of soda. 



There are in general three methods of determining whether the 

 experiment has given an accurate result or not. These are the 

 following : — 



1. The repetition of it with the same result. This proves that the 

 experiment has been well executed. But it does not decide whether 

 we have a method of operating capable of giving a correct solution 

 of the problem. For this purpose we must 



2. Examine if the loss, which represents the oxygen of the sub- 

 stance analysed, be a multiple by a whole number of the oxygen in 

 the oxide of lead with which it was united. If this be the case, 

 we estimate the number of volumes of oxygen in the pubstance 

 analysed the same as the multiple which it i^ of the oxygen in the 

 oxide of lead. Then we examine if the carbon and hydrogen 

 constitute a certain number of entire volumes. In that case the 

 experiment has every appearance of being exact. But still another 

 cluck is requisite, without which we may still be deceived. We 



inuit 



