When filtrates of lysed cultures (bacteriophage) are subjected to prolonged dialysis under osmotic pressure against water, the presence of the lytic agent can be detected outside the membrane only during the first few days. The residue remaining inside the membrane contains the bulk of the original lytic agent, and yet it is no longer capable of diffusing into the outer solution.
The interruption of diffusion is shown not to be due to any alteration in the permeability of the membrane. Moreover, the residue fails to diffuse through a fresh membrane of similar permeability, while the dialyzed portion of the phage passes quantitatively through a new membrane. When ultrafiltration under pressure was substituted for dialysis, the residue on the filter could be washed repeatedly with water without giving off into the filtrate any more active agent. However, if broth was substituted for water, a renewed diffusion of the active agent resulted.
These results are interpreted as indicating that the colloidal particles present in the lytic filtrates (and apparently endowed with properties of bacteriophage) do not represent autonomous units of the active agent, but merely serve as a vehicle on which the agent is adsorbed. The vary in size within limits wide enough to permit fractionation by means of ultrafiltration. When the coarser particles retained by the ultrafilter are washed with broth, some of the active agent is detached from its coarse vehicle particles. The agent, now more highly dispersed, is capable of passing the filter which held it back previously.
Preparation of a simple ultrafilter used in these experiments is given in detail.