antimicrobial containing agar plates versus a dilution test method as reference using different breakpoints. Vieira et al. compared the prevalence of tetracycline and sulphonamide MDV3100 CYP17 inhibitor resistance between both methods using similar breakpoint concentrations for tetracycline, whereas the agar plates contained a breakpoint concentration one dilution step up for sulphonamide. sterblad et al. compared the prevalence of resistance to ampicillin, trimethoprim and tetracycline using breakpoint concentrations one step dilution up for ampicillin, and one step dilution down for trimethoprim and tetracycline compared to the reference method. Despite the use of different breakpoint concentrations compared to the reference method, the rate of resistance detection did not differ statistically between the test methods for any of the antibiotics tested. These results can be explained by the bimodal distribution of MIC values of E. coli isolates around the used breakpoint concentrations for ampicillin, tetracycline, cefotaxime and trimethoprim and sulphamethoxazole. For this reason the use of a higher breakpoint concentration of these antimicrobials in an agar plate did not give different resistance results compared to broth microdilution using EUCAST epidemiological cut-off values. For ciprofloxacin, on the other hand, MIC values of the isolates within both test methods due to the fact that development of resistance to ciprofloxacin increases gradually by single-step mutations in the chromosome. Thus, a one or two-step higher dilution in breakpoint concentration resulted in a lower prevalence of resistance to ciprofloxacin found by replica plating. Besides test method, also the application of pooled samples was investigated in this study. Estimating the proportion of resistant isolates within a herd based on pooled faecal samples is convenient and least expensive to perform compared to individual samples. This study showed that the use of pooled samples is justified as the odds of finding resistant isolates within pooled samples did not significantly differ compared to individual samples. This is in accordance with studies among finisher pigs and feedlot cattle, although Wagner et al. only found similar resistance patterns for individual and pooled samples when prevalence of resistance to an antimicrobial was.2%. Whereas the study of Dunlop et al. was based on a replica plating test method, Wagner et al. used broth microdilution, but could not detect any resistant isolate within pooled samples of 10 samples for antimicrobials with a low prevalence. In this study the prevalence of isolates resistant to cefotaxime was also low and resistant isolates were not detected among individual or pooled samples when tested by broth microdilution. Surprisingly, isolates resistant to cefotaxime in individual and pooled samples were detected by replica plating even though the breakpoint concentration was one-step higher in dilution compared to broth microdilution. The fact that resistant isolates were found by replica plating is probably due to the higher number of isolates tested within each sample which increased the chance of picking at least one resistant isolate. This result emphasizes the advantage of testing multiple isolates per faecal sample for the determination of resistance, especially when resistance is rare. Simulation studies are used to investigate the impact of number of isolates and number of samples tested per herd on the precision of the estimated prevalence of resistance. These studies are usually based on prevalence of resistance obtained from one herd or flock of one farm only.