Raw milk, reconstituted skim milk, skim milk, sweet whey, and acid whey were membrane processed on different units from several manufacturers using various membranes with pore sizes ranging from nanofiltration through microfiltration. The bulk fluids were inoculated with either Staphylococcus aureus 196E, Salmonella typhimurium var. Hillfarm, and/or Pseudomonas fragi 4973. In addition, indigenous microorganisms were present. The permeate and retentate streams were monitored for bacterial numbers. Percent total solids of the permeate streams was determined. Temperatures and pressures were controlled. In no cases were the bacteria completely retained while concomitantly allowing permeated solids to equal the solids in the original bulk fluid. Findings indicated different membranes of same molecular weight cut-off exhibited dissimilar bacterial retention characteristics. Unit design/configuration appeared to play a role in retention of bacteria. Spiral wound microfiltration and ultrafiltration membranes reduced bacterial loads in the permeate by 98.9 to 99.99% while allowing 5 to 6% of the solids in the bulk fluid to pass through the membrane. The bulk fluid does not appear to affect the bacterial retention. The different wheys, milks, and reconstituted skim milk showed similar reductions in bacterial numbers when microfiltered or ultrafiltered through the same type of membrane. All three test microbes demonstrated similar declines during membrane processing. It appeared that bacterial morphology and size did not affect the bacterial retention characteristics. Results indicated that low-temperature membrane processing will not eliminate all microorganisms in the permeate nor did all milk components pass through the membrane into the permeate.