TRANSACTIONS OF SECTION B. 



577 



fluorides ; and carbnnic iicid breaks up at high temperatures into a mixture of 

 ■carbonic oxide and oxygen. 



Amongst illustrations of the greater atomic values which elements assume by 

 combining with both chlorous and basylous atoms than with atoms of the one 

 kind only we may take the following cases : Platinum is a metal of which the 

 atom has been supposed to be always tetravalent, because it has not been found 

 capable of combining with more than four atoms of chlorine. The common solution 

 formed by aqua regia contains the compound Hj Pt Clg, a perfectly definite and 

 crystallisable hydrogen salt. Chemists are constantly making and using the potas- 

 sium and ammonium salts, &c., corresponding to it, yet they conceal from them- 

 selves the fact that the atom of platinum is directly combined with eight monads 

 by calling the compounds double salts. The atom of silicon in the silico-fluorides 

 such as H.J Si Fg or Ko Si F^. is combined with twice as many monads as it can 

 take up of one kind ; so boron in the crystalline salt Na B F^ has a higher atomic 

 Talue than in its fluoride, owing to the presence of the atom of sodium. 



In like manner the atom of gold in the well-known salt NaAuCl^ has a 

 higher value than it can assume with chlorine alone. 



Sulphur of which the atom does not combine with more than 2 atoms of 

 hydrogen, forms with 3 atoms of methyle, or ethyle, and one atom of iodine, or 

 chlorine, &c., the well-known compounds like ISMeg; and iodine, which is con- 

 sidered a monad, forms the crystalline and stable periodate 0I(0H)5 and the various 

 metallic derivatives, such as 



01 ONa 01 (ONa)„ 01 (OAg), 



(OH), (OH)f 



The crystalline compound of the perchlorate with water (HC10^2HjO) has 

 probably a similar constitution. Chemical journals abound with descriptions of 

 definite and well-characterised compounds, which have, like the above, been put 

 aside by the atomicity theory, as mere molecular compounds. The following 

 formulae are taken almost at random, in illustration of the generality of atomic 

 -values far beyond those acknowledged by the theory of atomicity. 



KsAgl, 



I have for convenience written in the middle of each of these formula the 

 symbol ot the atom which I assume to act as connecting element. If we consider 

 the atomic values usually found iu these elements together with those represented 

 by the above list, we see that their atomic values vary according to the numbers 

 given in a line with them respectively in the following table. It has yet to be 

 proved that the atom of platinum is tetravalent iu any known compound, for there 

 is no sufficient evidence to show that platinic chloride has a molecular weight corre- 

 sponding to the formula PtCl^, instead of one corresponding to Pt^Clg, each atom 

 of platinum being partly combined with the other, partly with chlorine. 

 Atomic Sj-mbols 



c 



s 



1881. 



Pt 



Si 

 Sn 

 Cu 



Hg 



Mg 



Ag 



Atomic Values 

 2,4 

 2,4 

 4(?),8 

 4,8 

 4,8 

 2,6 

 2,4,6 

 2,4,6 

 ],5 



p p 



