The procedures currently used for isolating legionellae from environmental samples recommend filtration through a 0.2-,im-pore-size polycarbonate filter. In this study we evaluated the performance of 23 other filters composed of various materials and having various pore sizes. We prefer the 0.2-,im-pore-size Gelman Supor filter because of its high level of recovery, faster filtration rate, and ease of handling.

Eight laboratories compared counts of Escherichia coli from naturally or artificially contaminated ground beef, other meats and poultry, vegetables, fish and shellfish, cheese, and diverse sources such as swabs, by the Anderson-Baird-Parker direct plate (DP) and a hydrophobic grid-membrane filter (HGMF) method. For five of the eight laboratories overall counts by HGMF were significantly low (51-83%) compared with those by DP. Counts by HGMF tended to be lower for naturally contaminated samples; several possible causes were investigated. In a subsidiary study, analyst variation in counting HGMF ranged from 0.8-7.3%, with little evidence of effects from counting positive versus negative grid cells or from the fullness of growth or staining intensity.