598 
CHEMISTRY: HARKINS, HALL AND ROBERTS 
(Equations Lineaires) introduce an arbitrary parameter p and write 
II 2 II =-^21 j as an instance. 
1 H. B. Fine, these Proceedings, 2, 546 (1916). 
2 J. Kurschak, /. Math., Berlin, 142, 211-253 (1913). 
THE COBALTAMMINES 
By William D. Harkins, R. E. Hall and W. A. Roberts 
KENT CHEMICAL LABORATORY, UNIVERSITY OF CHICAGO 
Received by the Academy, August 22, 1 91 6 
The metallic elements whose properties of aggregation are such that 
they have low atomic volumes, or high ^cohesion/ form a series of com- 
plex ammonia compounds whose stability seems to increase with the 
cohesion, or decrease with the atomic volume of the metallic aggre- 
gate. These complex ammonia compounds, or ammines are of peculiar 
interest because, like the hydrates, their structure has not yet been 
explained in a satisfactory way from the standpoint of ordinary valence 
theories. Werner, however, has developed a special theory of valence 
which seems to fit very well the chemical behavior of these particular 
compounds, whether or not it is in agreement with the general behavior 
of other chemical substances. 
Now perhaps the most important point to be established with re- 
spect to a series of salts such as these ammines, is the type of each salt 
with reference to its ionization in solution. When we turn to the work 
of Werner and of other investigators in this field, it is found hat no 
w^ork has been done which determines directly the type for these salts, 
although it might be considered that their molecular conductance as 
determined by Werner, together with their chemical action, makes it 
seem probable that they belong to the types according to which he has 
classified them. 
The freezing-point method should give the most easy and certain 
method for the determination of the type of such salts as these, but it 
is just here that not only the work of other investigators, but even that 
of Werner^ himself, does not agree with his theory. Since in this labora- 
tory there is a double combination potentiometer system specially 
designed for us by Dr. W. P. White, to give with a fifty-couple copper- 
constantan thermocouple a reading to one twenty-thousandth of a 
degree, it seemed advisable to make for the first time a series of accurate 
freezing-point measurements upon the special series of cobaltammines 
which was most used by Werner in his work. The analyses were made 
by a Haber-Zeiss water interferometer, loaned to us by the Geophysical 
