4 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
When the experiment is repeated with diluted blood- 
serum no such solid is visibly separated, but if the solution — 
is first mixed with an equal volume of saturated ammonium 
sulphate solution and filtered, the filtrate gives threads and 
membranes like those obtained from white of egg, with the 
important difference, however, that they rapidly dissolve 
up again in the mother-liquid—protein molecules have in 
this case been simply aggregated into visible masses of 
solid, but no coagulation has taken place. They are | 
presumably formed even in the absence of ammonium 
sulphate, but then dissolve up again so rapidly that they 
cannot be seen at all. 
Microscopic observation of small air-bubbles in dilute 
ege albumin imprisoned between slide and cover-slip, and 
there repeatedly flattened out by pressure on the cover-slip, 
showed that similar solid structures were formed by this 
procedure, and strongly suggested that they were formed 
in some special connection with the air-surface of the 
bubbles. 
This idea was confirmed by the next experiment. 
When a dilute solution of egg albumin is allowed to flow 
at an appropriate pace down the inside wall of a burette as 
a thin film of liquid lining its whole interior, a wrinkled 
membrane develops on its surface in the lower part of the 
burette, and this is slowly washed down and heaped up as 
a cake of solid protein on the surface of the i collecting 
at the bottom. | 
The next step was the discovery that the air surface 
of the protein solution, even when at rest and under 
conditions which entirely preclude evaporation of water, 
rapidly becomes rigid, and must therefore spontaneously 
coat itself with a thin membrane of solid protein. In this 
vessel containing egg-white diluted 100 times with water 
and filtered, are suspended two magnetised bits of watch- 
