BACKGROUND The outcome of microbiological diagnostics may depend on the sampling technique. It was the aim of the present study to compare two widely used sampling techniques for subgingival bacteria using quantitative real-time polymerase chain reaction. METHODS Twenty patients with chronic periodontitis were randomized into two groups. In group A, samples were taken first with a paper point and then with a curet at the same site (single-rooted teeth with probing depth >5 mm) before scaling and root planing and after 10 weeks. The sampling sequence was reversed in group B. The analysis enabled the quantification of Actinobacillus actinomycetemcomitans, Fusobacterium nucleatum, Porphyromonas gingivalis, Prevotella intermedia, Treponema denticola, and Tannerella forsythensis and total bacterial counts (TBCs). Statistical analysis included t test, kappa, and Spearman correlation. RESULTS Higher TBC was harvested with curets than by paper points (P = 0.008). The plaque composition with regard to total target pathogens was similar for both sampling techniques. A strong positive correlation was found between curet and paper point samples for TBC and single target bacteria. CONCLUSIONS Overall, there was a relatively good agreement for the results of paper point and curet sampling. Thus, both techniques seem to be suitable for microbiological diagnostics.