OBJECTIVE The antibacterial activity of two different formulations of a chlorhexidine/thymol varnish should be elucidated in vitro and in vivo. METHODS The agar diffusion assay with Cervitec(®) and CervitecPlus(®) and three reference strains each of streptococci, lactobacilli, actinomyces and periodontal pathogens was performed. In a split-mouth study, 40 volunteers applied the test (CervitecPlus(®), solvent water and ethanol) and control (Cervitec(®), solvent ethyl acetate) varnish at buccal recessions of premolar teeth at baseline as well as after two, four and seven days. Supra- and subgingival plaques were collected 2 weeks before baseline and at the screening appointments. Supragingival plaque was analysed for mutans streptococci and lactobacilli and subgingival samples for Aggregatibacter actinomycetemcomitans, Fusobacterium nucleatum, Porphyromonas gingivalis and Porphyromonas intermedia. Friedman/Wilcoxon tests and U-test were used for statistical analysis (P < 0.05). RESULTS Most reference strains were susceptible with inhibition zones (mm) as follows: Cervitec(®)/CervitecPlus(®) streptococci 27 ± 1.7/21.3 ± 2.5, lactobacilli 26 ± 9.2/23.7 ± 4.9, actinomyces 36.3 ± 6.6/27.3 ± 1.5, periodontal pathogens 18.7 ± 7.6/18 ± 1.7. Both varnishes reduced significantly the counts of mutans streptococci and lactobacilli in the patients. However, no significant differences were found between test and control sides at any time. The total counts of periodontal pathogens were low. A tendency to higher counts of A. actinomycetemcomitans at the control side could be shown; the test side did not harbour significantly higher counts. CONCLUSIONS Both varnishes may influence the plaque formation and reduce mutans streptococci in supragingival plaque.