36 ROYAL SOCIETY OF CANADA 
In the case of the sulphates which have shown no contraction, I 
give the concentration of the weakest solution examined, and the value 
of the expansion for that solution. 
SULPHATES NOT EXHIBITING THIS CONTRACTION. 






at Solu. Expansion Observer 
tion examined.| 12 CU: CM: 
INN OOVNBYIN iy ss 5055 044008 ‘055 per cent. 00021 Dijken. 
Ammonium-Aluminium ....| 1°50 i ‘00111 Gerlach. 
Ammonium-Iron ...... ..... 2°76 i ‘00513 FF 
Ammonium-Sodium ........ 4°99 He ‘01732 Schiff. 
BLUE RAREMENT RE 3°40 sf 00462 Mendelejefi. 
Errig EPA EME metus cies 10° i "01242 Gerlach. 
Hydrogen-Potassium........ 5° + ‘01454 Kohlrausch. 
ETCOITMIER NE ees ‘48 Ue “00239 Dijken. 
Magnesium-Potassium ..... 3°43 se ‘00407 Schiff. 
PObassiuim, ENT PE PEER cite I< oe ‘00147 Hassenfratz. 
Potassium-Aluminium. ..... 12 2e ‘00041 | MacGregor. 
SOUL ER PRET E rae. sue r ‘13 hf 00018 The Author. 
Potassium-Iront eee 2°85 ae 00416 Gerlach. 



On examination of the last two tables we find (1) that in the 
former table every metal is bivalent, and consequently has only one atom 
in the molecule, while in the last table every metal, except Beryllium, 
has more than one atom in the molecule of its sulphate. Hence so far 
as the available data enable us to form a judgment, this property would 
appear to be peculiar to the sulphates of the bivalent metals. The 
only exception to this statement is Beryllium, and this has only been 
examined in a solution, the concentration of which was 3°4%. The 
accuracy of my work was not sufficient to establish a contraction for 
Manganese, but made it probable. The contraction of Aluminium is 
doubtful. If Aluminium were included in the first table, the results 
would indicate that the sulphates of bivalent metals exhibit contraction, 
those of univalent metals do not, those of trivalent metals may or may 
not, while those of two metals do not. 
