730 Proceedings of Royal Society of Edinburgh. [sess.. 
probably has the constitution 
OH 
CrZc 2 0 4 -K 
\C 2 0 4 or C 2 0 4 : (H0.Cr.C 2 0 4 .K) 2 . 
C ' C 2 0 4 - K 
X OH 
Probably by the application of heat an anhydrous compound may 
be obtained, exactly as in the preceding cases. It is interesting to 
notice that in his first paper on the chromoxalates * Rosenheim, 
in putting forward a somewhat similar classification of doubled 
constitutions derived from the sesquioxide Cr 2 0 3 , has predicted the 
existence of this compound, though he failed in his efforts to 
obtain it. 
This type will be styled B. 3. 
The remaining series, C., is that in which but one hydroxyl 
group is replaced. The chromoxalate of this class and its 
anhydrous derivative would have the constitutions 
OH 
Cr^OH and 0 : Cr.C 2 0 4 .K 
x C 2 0 4 -K 
respectively, but these compounds have not been prepared. 
Members of this series are up to the present rare, but a potassium 
antimony oxalate 0 :Sb.C 2 0 4 .K, and tartar emetic 0 :Sb.C 4 H 4 0 6 .K, f 
may be taken as instances. 
Nearly all the complex derivatives of the sesquioxides can be 
referred to one or other of these types. Some of the important 
exceptions will be mentioned later. 
Most of the investigators in this field of research have recognised 
to some extent that analogies exist among these complex com- 
pounds. Thus E. A. Werner has extended his observations from 
the chromoxalates to the chromium compounds of other organic 
acids, and Rosenheim and his collaborators have investigated 
the derivatives of oxalic acid and a number of trivalent and 
quadrivalent metals. Again, Christensen | has compared the 
* Zeit. fur anorg. Ghem 11 , pp. 212 et seq. 
f The constitution of this salt is doubtful. 
t J. pralct. Ghem., ii., 34, p. 44, and 35, p. 172. 
