192 On Numerical Proportions in Chemical Comlinations, 



It is by these polyhedrons that I have represented the various 

 arrangements of tlie molecules of all bodies. When these bodies 

 contain such substances onlv of which we may measure the vo- 

 lume in the state of gas, we liave immediately the number of the 

 molecules of each species which enter into their composition. 

 When a simple body cannot be obtained in tiie state of gas, we 

 must try in succession various sup]50sitions relative to the num- 

 ber of molecules of this simple body, which are contained in one 

 of the compounds whicii it f(.»ims with a gaseous substance, oxy- 

 gen for instance. Tlie relations in weight make known the 

 number of the molecules of tbe same body which enter into its 

 other compounds ; and the condition which it must satisfy, that 

 all the numbers of molecules whicli are obtained correspond to 

 polyhedrons comprehended in the foregoing table, soon makes 

 known that one of these different suppositions which can agree 

 with the entire whole of the phaenomena: it then becomes easy 

 to calculate the respective weights of the molecules of all the 

 simple bodies; and these weights once determined, it is suffi- 

 cient to have a pretty close analysis of a compound body, and 

 to know how much its particles contain of molecules of each of 

 its elements, and thus correct the inevitable errors of analysis. 



Several chemists have endeavoured to attain the same result 

 bv determiiiing the respective weights of certain proportions of 

 the different simple bodies which always enter an entire number 

 of times into the bodies which are composed of them. These 

 proportions do not lead to results conformable to the experiments, 

 but when they are always multiples or submultiples of the re- 

 spective weights of the molecules : but when we make use of 

 them, nothing can indicate how many of the proportions of a 

 simple body ought to enter into one of those compo\;nds : whereas 

 the consideration of the representative forms shows, in many 

 cases, how much in a compound body there ought to enter of 

 molecules of each of its elements, and even leads us to establish 

 between tlie combinations of two simple bodies with all the 

 others, such a dependence, that, the comliination of one of those 

 bodies bein.g known, we may foresee those of the other. I have 

 found for instance, bv comparhig the combination which oxygen 

 and hydrogen form with different bodies, that, with the exception 

 of chlore and sulphur, the combinations of v\'hich with hydrogen 

 present the j^ropertics of the acids, one same quantity of a body 

 susceptible of being united to hydrogen is combined in such a 

 manner, that there are in general, in each of the particles of the 

 compound, four molecules, of iiydrogen more than there are mo- 

 lecules of oxygen in the corresponding combination of the same 

 body, with this last gas. We may even remark, that when this body 

 forms with oxvgen several combinations, some of whicli are more 



difficult 



