<jats:title>Abstract</jats:title> <jats:p>The neomycin microbiological cylinder plate assay using an ionic extraction method was validated for 2 swine feeds with neomycin base concentrations of 140.0 and 280.0 g/ton, 2 cattle feeds with neomycin base concentrations of 140.0 and 280.0 g/ton, and 2 premixes with neomycin base concentrations of 1.4 and 7.0 g/lb. The accuracy, precision, and reproducibility were evaluated by assaying the 6 feed samples in duplicate on 6 different days by an Upjohn laboratory (Laboratory 1) and 2 outside laboratories (Laboratories 2 and 3) for a total of 72 data points per laboratory. Three plates were used for each sample assayed. The recovery over all complete (type C) feed samples and (type B) premixes tested, ranged from 82 to 109% for Laboratory 1,85 to 114% for Laboratory 2, and 78 to 134% for Laboratory 3. Mean percent recoveries for all samples tested were 98.0,98.9, and 103.7% for Laboratories 1,2, and 3, respectively. Day-to-day performance yielded 95% confidence limits of ±11.9, ±16.9, and ±25% for Laboratories 1,2, and 3, respectively. Of 215 assay values, 206 (96%) fell within 20% of the actual targeted recovery (100%). Plate-to-plate variation of each of the 3 laboratories also was assessed. Statistical estimates for the use of 1,5, and an infinite number of plates showed that when 2 plates were used by Laboratories 1,2, and 3, than 99,99, and 95%, respectively, of the assay results fell within the acceptable assay compliance limits (70–125% recovery) for complete feeds and premixes. Slight improvement was found only for Laboratory 3, when more than 2 plates were used.</jats:p>