971 
Fe,Cl, 12 aq. (melting point 37°), and over that of FeCl, 7 aq. (melting- 
point 32.5°) there is the greatest chance that in this neighbourhood 
the homogeneous solution will contain considerable hydrate concen- 
trations. For at higher temperatures the hydrates will dissociate 
more strongly as a rule, hence their concentrations will decrease. 
The times of outflow found at 40° are recorded in the table of 
the preceding page. 
When these results are graphically represented, as has been done 
in Fig.1, we get a curve of viscosity exhibiting a very pronounced 
maximum and minimum, lying on the left and the right of the 
concentration of the hydrate with 12 aq. 
This peculiar shape of the viscosity curve must in my opinion 
be interpreted in the following way. 
When no hydrates were formed in the solution, the viscosity of 
the solution with the Fe,Cl,-concentration would increase in an 
ever greater degree, and in the end very high values would occur, 
because the viscosity of supercooled liquid Fe,Cl, at 40° will be 
exceedingly great. 
At 6 a decrease of the viscosity is now found here, which in my 
opinion must be ascribed to the increasing hydrate concentration. 
When only the hydrate Fe,Cl, 12 aq conld exist in solution, it was 
to be expected that the minimum would lie near the concentration 
of this hydrate. 
At 40°, however, we pass not only over the top of the melting- 
point line of Fe,Cl, 12 H,O, but also, though not at such a small 
distance, over that of the melting point line of Fe,Cl, 7 H‚O. 
At the descent from 6 to c an appreciable increase of the con- 
centration of the hydrate with 7 H,O will, therefore, also take place, 
and when this gives also rise to a decrease in viscosity, the result 
will be that the viscosity curve, which must finally ascend again in 
consequence of the increase of the Fe,Cl, concentration, presents a 
minimum, lying on the righthand side of the concentration of the 
hydrate with 12 aq. 
After this result had been obtained, and the plan had been formed 
to examine also the system HO,—SO,, because this seemed to be 
particularly suitable for this purpose, it appeared that KyretscH'), 
who studied this system from different points of view, also gives a 
viscosity line which presents a close resemblance with the curve 
discussed above and of which up to now no notice had been 
taken. 
1) Ber. 34, 4102 (1901). 
