472 



Messrs. Playfair and Joule oi^\\^ 



salt, but this has been determined by Kopp, who ^ives it at 

 2*037, without however describing the character of the hy- 

 drate which he examined. It is possible, therefore, that it 

 may not be the same as that which we have examined; but , '■ 

 presuming it to be so, the volume of this salt, according to ; 

 I gKc^}! yf^^^ he I 



^■^, ' S'',- :'' = '?" Sp.gr. .imsVi 



■' ! ' NiO, "SO3 + 6HO, vol. of salt 64.-6 ... 2-037. 



The volumes of the magnesian sulphates with 7 atoms of 



r water are obviously less than those which would result were m^, 

 :"'"lhey multiples of the volume 11. But as we have already ' 

 : i ..seen that the water of hydration does not always enter into '^ 

 j combination with the volume 11, but occasionally with thatof Vjj>' 

 1 ^"•19*8, or the volume of ice, the results obtained may be ex- ;' 

 j^{;(plained on this view. Graham*, in his researches on the ,„' 

 ! phosphates, and on the heat of combination, drew attention 

 to the fact, that the atoms of water seem to be attJiched to- 

 gether \xitwos. Mil Ion t more lately has shown that the two 

 last atoms of water in sulphate of magnesia are less firmly 

 attached than the five remaining atoms; tl^t.a magnesian; 

 sulphate in fact may be viewed as \Mlf3' 'i ) fcj; niuaJ 



MgO, SO3 + 5HO + 2HO.'«v# «^flt ni nuDqqn 

 That 5 atoms of water form the natural numbers for the mag- ' 

 nesian sulphates we have evidence in the salts of copper and 

 manganese, both of which possess these 5 atoms of water in 

 combination with a volume of 11, at least. 



CuO, SO3+ 5HO vol. ofsalt 55-4 (P. and J.) 

 MnO, SO3 + 5HO ... 57-6 (Kopp.) 



As then the two additional atoms of water are retained by 

 a less feeble affinity than the remaining five, may we not as- 

 sume that they are present, as in the case of other salts pos- 

 sessing a feeble affinity for water, with the volume of ice, 

 whilst the original 5 atoms possess the higher volume of 11 ? 

 The following table will show that this hypothesis gives re- 

 sults by calculation which do not differ widely from those ob- 

 tained by experiment. . Q,ff^^{ j>, 



* Phil. Trans., "part "l . 1837, page 67. 



f Annales de Chimie, 3 s^rie, t. xii. j). 134. 



■ !9 iid Jonna:* ,qoo '5^ omfia 91 



i\\lU 



.^^Ok.kdi. .O' 



