522 Messrs. Play fair «w«? Joule ow vv 



one of the family of which potassium stands as the type. In 

 calomel and chloride of ammonium we have a direct case in 

 point, and the similarity of volumes is very striking. 



Diff. 

 Chloride of ammonium, NH4CI . . . 34*0^ . q 

 Calomel, Hg^Cl 33-2 J^^'" 



Tn this case we have taken chloride of ammonium, because 

 KCl assumes the volume of four atoms of ice. 



Subchloride of copper, like NH4CI, possesses three volumes, 

 according to Karsten's experiments and our own, but these 

 three volumes are multiples of 9'8, and not of ll'O. 



By experiment. Bv calculation. 



. ^ , , " , 



Subchloride of copper, vol. 29*2 3'376 9*8x3 = 29-4 3-363 



Another illustration is furnished in sulphate of protoxide 

 of mercury and sulphate of potash. 



DifF. 

 Protosulphate of mercury, vol. 33"20\„ , _ 

 Sulphate of potash .... 3305 j 



These are instances in which two atoms of a magnesian 

 metal are at once shown to be equivalent to one of a metal of 

 the potash family ; but it does not thereby preclude the pos- 

 sibility of two atoms of a magnesian oa^ide being equivalent 

 to one atom of potash. For example, a magnesian sulphate, 

 MgO, SOg, affects a volume 22, or 11x2, while the same 

 salt united to an atom of constitutional water has the volume 

 33, or MgO, SO3HO becomes equal to KO, SO3, which 

 also possesses a volume of 33. The most striking case, how- 

 ever, is seen when crystallized subnitrate of lead is compared 

 with nitrate of potash. 



Nitrate of potash. KO, NO5, vol 49'0 



Subnitrate of lead, PbO, NO5 + PbO, vol. . 49-0 



The fact that two atoms of a magnesian oxide are equiva- 

 lent to one of potash, appears to find its explanation in the 

 circumstance that we uniformly find the salt of potash assu- 

 ming one volume greater than the corresponding salt of mag- 

 nesia. Honce, as the volume of the oxides corresponding to 

 the latter body is equal to unity, the equivalency of two of 

 their atoms to one of potash becomes a matter of necessity. 



To sum up these remarks, we conceive (1.) that Graham 

 has taken the correct view in supposing subsalts to represent 

 hydrated salts, in which water has been replaced by a me- 

 tallic oxide ; and (2.) that the volume of two atoms of a metal 

 of the magnesian family, in which Me include hydrogen, is 

 equal in volume to one of the potassium group ; or two atoms 



