TRANSACTIONS OF SECTION 4, An” 
However, these values, especially for the swinging-disc method, were very con- 
siderably altered— 
(i) By dialysing the liquid for three or four days: e.g., not dialysed (7.c. by me), 
X=0:2018 ; after first dialysis, \=0°2760; after second dialysis, = 0°5162. 
(ii) After allowing disc to remain in solution for some time: eg., last- 
mentioned value had become = 0:9753 after 2880 minutes, 
(iii) By removing disc, washing it with water, and replacing it, caused large 
decreases in value: e.g., as much as 63 per cent. was once observed. 
(iv) By heating liquid to about 50°C. and allowing it to cool again, caused 
increase: eg., X for the once-dialysed liquid became, after two such heatings, 
=0°6571 ; whilst for the twice-dialysed, after five such heatings, the author obtained 
a value as high as 1'8364, z.c., nine times original value (0:2018). 
These results gave some very good ‘ parallel’ ‘ parabolic’ curves. 
(v) The influence of the size of the oscillation of disc was very marked with 
more viscous solutions: e.g., A=1:3889 for amplitude of 502 scale divisions ; 
and = 06020 for one of 42 divisions. These gave a good series of radial curves—z.e., 
the more viscous the liquid the steeper the curves. 
(vi) A violent shaking or accidental shaking, or making disc swing through 
large arcs, caused large decreases in values of A. 
Further, the ‘ zero-position’ (with the more viscous solutions) of the image of 
the cross-wires on the scale suffered large displacements to right or left of initial 
position accordingly as the disc was set in motion by bringing up the N. or 
the S. pole of a magnet to the directing magnet supported above the disc: e.7., 
on one occasion a displacement of as much as 658°5 scale divisions was obtained by 
holding second magnet for a few seconds near the directing magnet—this was 
with a small disc, when solution was very viscous. 
An explanation of these apparently very irregular results is to be found in 
Professor Quincke’s ‘ Schaum-zellen’ theory of a colloid, viz., that such a ‘ pseudo- 
solution’ really consists of two solutions, A and B (like milk or an emulsion of oil 
and water), the one, A, being rich in colloidal substance, heavier and more 
viscous than the other, B, and which possess at their common surfaces of separa- 
tion a surface-tension, which will vary with the common surface, z.e., with the 
formation of the small Quincke Schaumwinde, particles, spheres, &c. 
The values found by the transpiration method also showed a gradual increase 
during the course of these experiments, but far less strongly emphasised and not 
subjected to such irregularities. The flow of the colloidal solution through the 
solution in the one direction in the Ostwald apparatus does not favour the forma- 
tion of the foam-cells, walls, &c. 
5. Magnetic Double Refraction of Colloidal Iron Hydrate, 
By A. D. Dennine, I.Sc., Ph.D. 
The magnetic double refraction of iron salts has often been looked for, but up 
to the present only observed by J. Kerr (‘ Brush Grating Experiments,’ Brit, Assoc. 
Report, 1901), and by Qu. Majorana (Rendic. Acc. det Lincei, 1902) and Schmaus 
(Ann. d. Physik, 1903), with old specimens of ‘ Bravais’s Iron.’ These experi- 
menters found very transient results and give no absolute values for the same. 
Having found very remarkable results for the viscosity of the colloidal 
Fe(OH), from Merck, it was deemed desirable to examine if solution would show 
properties of magnetic double refraction. But although the author made many 
(more than eighty) different experiments under various conditions—z.c., after repeated 
dialysis, frequent heating and cooling, by varying concentrations and various 
magnetic field strengths in differently shaped polarisation tubes—he has only been 
able to find traces of such properties. 
Only on one occasion have really measurable results with this solution from 
Merck been obtained. This was after slowly evaporating a portion until it con- 
tained approximately 20 per cent. Fe. Then, by placing a few drops between thin 
glass strips on the poles of the electro-magnet, a displacement of the interference 
