in Chemical Coml'wations. S45 



are \eA, with those of tlie experiment, seem=! to me 'one of the 

 most remarkable proofs of the tlieory therein detailed : but in 

 order to give an. example of t!ie manner in which we may draw 

 from this theorv the determination of the quantity of water which 

 is combined with bodies, either in the state of crystallization or 

 even after thev have undergone the action of a strong heat, I 

 shall cite the determination according to the representative form 

 of the ))articles of potash from the quantity of water v/hich is 

 united with it in these two states. After having established, 

 setting out from the pluenomena which potassium exhibits vvhen 

 we place it in contact with v.ater and ammoniacai gas, that the 

 particles of potash have as their representative fonn an cctohe- 

 dron composed of two molecules of oxygen and four of metal, 

 I find that, in the crystallized hydrate, the quantity of the oxy- 

 gen of the water ought to be double that which is united to 

 the potassium ; but after the hydrate has been fused, those two 

 quantities of oxygen ought to be as 4 :3, because a particle of 

 hydrate in this state has for its representative form a heptaocto- 

 hedron tbrtr.ed by the meeting of a trioctohedron comi)osed of 

 three octohedral particles of potasii and of a tetraoctohedron of 

 four octohedral particles of water. Now, according to the com- 

 position of potash as determined by Messrs. Thenard and Gay 

 Lussa'c, 100 parts of potassium unite with 19"945 parts of oxy- 

 gen to n^.ake lU)-9-45 of potash. It follows from what I havs 

 said, therefore, that this quantity of potash .ought to retain, at 

 every temperature, a quantity of water iu which th.ere is 26*593 

 of oxygen, and which consequently weighs 30- 139 ; that is to 

 iay, nearly -^-jhrth of the weight of tlie potash, as has been found 

 by the most accurate analyses. 



The combinations of oxygen, hydrogen, and chlore, either with 

 themselves or with other bodies, have been successively the sub- 

 jects of researches analogous t>5 those just mentioned. As it is 

 impossible to indicate ;.ll the results here, I shall confine myself 

 to those of the combinations m which all the elements may be 

 obtained in the state of gas, and in which the numbers of the 

 molecules of each of their elements are consequently given im- 

 mediately. 



We have already ascertained the representative forms of the 

 particles of two combinations of azote and of oxygen, the oxide 

 of azote and the nitrous gas ; that of the nitrous acid ouglu to 

 be determined according to the ratio of the volumes of nitrous 

 gas and oxygeri of which it is composed. , Experiments have 

 been made upon this subject, but their results are at variance. 

 According to the analyses of Sir Humphry Davy, this acid is 

 composed of two volumes of nitrous gas, and of a volume of 

 cxygen : each of its particles will then contain two molecules 



of 



