The obvious desirability of finding a means for the more adequate concentration of any of the filterable viruses, led us to attempt to adapt the ultrafiltration method, used by Seibert in the study of tuberculin, to the concentration of poliomyelitis virus. That several of the well studied filterable viruses can be retained by filters or ultrafilters has been known for some time. But the work has been done largely for the purpose of determining the size of the viruses. Krueger and Schultz and Clifton, Schultz and Gebhardt, for example, have made careful studies of the filterability of poliomyelitis virus through graded collodion membranes. Their results indicate that the size of this virus is certainly less than 50 millimicrons in diameter and probably below 25 millimicrons. We have previously employed distillation in vacuo and fractional protein precipitation (Clark, Schindler and Roberts and Clark, Roberts and Preston) for the successful purification and concentration of poliomyelitis virus. Both of these methods, however, necessitate the subsequent elimination of the large amounts of salts by dialysis, before the material can be used for animal inoculation. The method here presented avoids that difficulty and we hope that it may give a more adequate basis for the study of the properties of poliomyelitis and other viruses. Dried soluble cotton was dissolved in concentrations of 8, 10, and 12% in glacial acetic acid and allowed to age, as a rule, at least a month, before being used to form the membranes over the alundum shells of medium porosity. Sterile distilled water was employed for washing the membranes, subsequently testing for the presence of acid with Congo Red. At least a liter of water was run through the membrane after the test for acid became negative. The permeability of the membranes was determined by their capacity to hold back Congo Red, a solution of crystallized tgg albumin and a solution of hemoglobin. The Congo Red was tested for by the addition of acid to the ultrafiltrate, and the presence or absence of tgg albumin was determined by using a specific precipitating serum of high titre. The increasing density of the red color in the “concentrate,” when a small amount of laked blood had been added to the material to be filtered, and the colorless character of the ultrafiltrate served as an indicator of value. Conclusions. A method for the ultrafiltration and concentration of poliomyelitis virus is described. This should be applicable to the further study of the properties of poliomyelitis and other filterable viruses. The method does not result in a direct proportional volumetric increase in the potency of the ultraresidue. A concentration of an inhibiting agent is suggested.