Campylobacter can be difficult to recover from complex samples due to overgrowth by background bacteria. A 0.45- or 0.65-μm-pore-size filter overlaid on agar plates can be used as a means to separate Campylobacter from confounding non-Campylobacter cells, facilitating detection on solid plating media. It is unclear what percentage of cells in a Campylobacter suspension passes through a filter and results in visible colonies. The objective of this study was to compare the number of Campylobacter cells detected by the filter method with those detected by direct plating and determine if the filter method can be used to estimate cellular density of an unknown Campylobacter in suspension. Overnight liquid cultures of six subtypes of Campylobacter jejuni and six of Campylobacter coli, all originally detected in chicken samples, were used for this study. Motility of isolates was tested by using a stab into soft agar, incubating plates, and measuring colony size. Each subtype was applied to Campy-Cefex agar directly and through a 0.45- or 0.65-μm-pore-size filter. Filters were removed, plates were incubated, and colonies were counted; three replications were conducted. Mean recovery by direct plating was 8.3 log CFU/mL. Regardless of pore size, the overall mean number of Campylobacter detected by using the filter method was significantly less than that using direct plating (P < 0.05). The mean difference between direct plating and plating though a 0.65-μm-pore-size filter for motile Campylobacter was log 2.4 CFU/mL, with a 95% confidence interval of ±0.2 log CFU/mL.