SURFACE-FILMS 5 
spring, one on the surface, the other deeply submerged— 
observe that, when a magnet is brought near, the latter is 
freely mobile but the former is rigidly fixed. A simpler 
way of observing this surface rigidity is to compare drops 
of albumin solution and of water on a glass plate when both 
are lightly dusted with sulphur—the sulphur slides down 
the slope of the water drop and leaves the vertex clear, but 
on the albumin solution it is instantly fixed just where it 
falls. 
The mystery was by these experiments to some extent 
solved. In this surface-rigidity of egg-albumin I had 
re-discovered a little-known fact described long previously 
by Plateau, and it became clear that the solid separated 
out when its solutions were shaken consisted simply of 
rolled-up and contorted membranes which had originally 
formed spontaneously at the water-air surfaces, in some 
cases coagulated, in others apparently unchanged. 
Applying the burette method to a wide range of other 
solutions, mostly colloid (including amongst them solution 
of the glucoside saponin which Plateau had described as 
having rigid air surfaces), solid ‘* mechanical surface 
aggregates,” 
as they may be termed, were obtained from 
every protein solution tested, and also from numerous 
soaps, aniline dyes, quinine (the free alkaloid), saponin, 
methyl orange in neutral solution, phenol-phthalein in 
neutral or acid solution, and from suspensions of gamboge, 
mastic, resin, and sulphur. From bile-salt and sodium 
oleate only viscid gums were obtained. Colloidal. silica, 
quinine salts, and phenol-phthalein in alkaline solution gave 
no aggregates. 
Testing the mobility of the air surface of each of the 
above solutions it was found that all except bile-salts and 
sodium oleate developed rapidly, or only on standing, 
either a true rigidity or a marked viscosity in the Plateau 
