Membrane filters with pore-size ratings of 0.22 microns and 0.45 microns were tested for their ability to recover Pseudomonas diminuta ATCC 19146 (P. diminuta), the organism typically used in bacterial retention testing of sterilizing-grade membrane filters. For each of the two pore-size ratings, filters of two membrane filter polymer materials, hydrophilic PVDF (Millipore Durapore) and mixed esters of cellulose, were tested, resulting in an evaluation of four potential recovery filters. The 0.45 microns mixed esters of cellulose filter is the currently accepted membrane for this purpose. The data show no difference in the ability of the four filters to recover freshly cultured P. diminuta. Moreover, the membrane-filter method was shown to provide a very high bacterial-recovery efficiency, equivalent to that of the spread-plate method. Thus, 0.22 micron filters, despite their ability to retain higher levels of bacteria, proved not to have an advantage over 0.45 micron membranes in terms of bacterial recovery. This result, combined with (1) the knowledge that more open membranes have been shown experimentally to more efficiently recover stressed organisms; (2) the potential to produce stressed cells in an actual bacterial retention test; and (3) the long history of the successful use of 0.45 microns mixed esters of cellulose for bacterial recovery, support the continued use of the 0.45 micron filter in this application.